From weiyu at cs.cmu.edu  Fri Jan 13 11:54:00 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Fri, 13 Jan 2017 11:54:00 -0500
Subject: [AI Seminar] AI Lunch -- Swaprava Nath -- January 17
Message-ID: <CABzq7eq3m3nWE4106R9gdnE8bOWvp=s9w5eSQQnDjovYCeQjow@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you Next Tuesday, January 17, at noon in NSH 3305
for the *first* AI lunch this semester. To learn more about the seminar and
lunch, please visit
the AI Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Swaprava Nath <http://www.cs.cmu.edu/~swapravn/> will give a
talk titled ?Preference Elicitation For Participatory Budgeting?.

*Abstract:* Participatory budgeting enables the allocation of public funds
by collecting and aggregating individual preferences; it has already had a
sizable real-world impact. But making the most of this new paradigm
requires a rethinking of some of the basics of computational social choice,
including the very way in which individuals express their preferences. We
analytically compare four preference elicitation methods -- knapsack votes,
rankings by value or value for money, and threshold approval votes --
through the lens of implicit utilitarian voting, and find that threshold
approval votes are qualitatively superior. This conclusion is supported by
experiments using data from real participatory budgeting elections.

This is a joint work with Gerdus Benade, Ariel D. Procaccia, and Nisarg
Shah.
Forthcoming in AAAI 2017.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170113/889ea552/attachment.html>

From weiyu at cs.cmu.edu  Mon Jan 16 13:34:44 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 16 Jan 2017 13:34:44 -0500
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 70, Issue 1
In-Reply-To: <mailman.65.1484413216.27330.ai-seminar-announce@cs.cmu.edu>
References: <mailman.65.1484413216.27330.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7eqKsgQQX_MmK-d9jUhP9LGXL+OoHmSzG8HOOB6_shROeQ@mail.gmail.com>

This is a reminder that the talk will be tomorrow noon, Jan 17.

On Sat, Jan 14, 2017 at 12:00 PM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Lunch -- Swaprava Nath -- January 17 (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Fri, 13 Jan 2017 11:54:00 -0500
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Subject: [AI Seminar] AI Lunch -- Swaprava Nath -- January 17
> Message-ID:
>         <CABzq7eq3m3nWE4106R9gdnE8bOWvp=s9w5eSQQnDjovYCeQjow at mail.
> gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you Next Tuesday, January 17, at noon in NSH 3305
> for the *first* AI lunch this semester. To learn more about the seminar and
> lunch, please visit
> the AI Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Swaprava Nath <http://www.cs.cmu.edu/~swapravn/> will give a
> talk titled ?Preference Elicitation For Participatory Budgeting?.
>
> *Abstract:* Participatory budgeting enables the allocation of public funds
> by collecting and aggregating individual preferences; it has already had a
> sizable real-world impact. But making the most of this new paradigm
> requires a rethinking of some of the basics of computational social choice,
> including the very way in which individuals express their preferences. We
> analytically compare four preference elicitation methods -- knapsack votes,
> rankings by value or value for money, and threshold approval votes --
> through the lens of implicit utilitarian voting, and find that threshold
> approval votes are qualitatively superior. This conclusion is supported by
> experiments using data from real participatory budgeting elections.
>
> This is a joint work with Gerdus Benade, Ariel D. Procaccia, and Nisarg
> Shah.
> Forthcoming in AAAI 2017.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20170113/889ea552/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 70, Issue 1
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170116/80368ee4/attachment.html>

From weiyu at cs.cmu.edu  Fri Jan 20 15:27:56 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Fri, 20 Jan 2017 15:27:56 -0500
Subject: [AI Seminar] AI Lunch -- Hoda Heidari -- January 24
Message-ID: <CABzq7epF2kMt9QhhDxL7=ucen2oXb7fab-ee74=n4pAROjb1PQ@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you Next Tuesday, January 24, at noon in NSH 3305
for AI lunch. To learn more about the seminar and lunch, please visit
the AI Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Hoda Heidari <http://www.seas.upenn.edu/~hoda/> from
 University of Pennsylvania will give a talk titled ?Pricing a Low-regret
Seller?.

*Abstract:* As the number of ad exchanges has grown, publishers have turned
to low regret learning algorithms to decide which exchange offers the best
price for their inventory. This in turn opens the following question for
the exchange: how to set prices to attract as many sellers as possible in
order to maximize revenue. In this work, we formulate this precisely as a
learning problem, and present algorithms showing that simply knowing the
counterparty is using a low regret algorithm is enough for the exchange to
have its own low regret learning algorithm for the optimal price.

(Joint work with Mohammad Mahdian, Umar Syed, Sergei Vassilvistkii, and
Sadra Yazdanbod. Appeared in ICML'16)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170120/c4b09bcf/attachment.html>

From weiyu at cs.cmu.edu  Tue Jan 24 09:40:19 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Tue, 24 Jan 2017 09:40:19 -0500
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 70, Issue 3
In-Reply-To: <mailman.91.1485018019.21794.ai-seminar-announce@cs.cmu.edu>
References: <mailman.91.1485018019.21794.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7eqrAPfykoCA2DnKXt+4h3_=-mKqp8sM4KVYYRNZ0+QbOA@mail.gmail.com>

This is a reminder that the talk will be noon TODAY, Jan 24.

On Sat, Jan 21, 2017 at 12:00 PM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Lunch -- Hoda Heidari -- January 24 (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Fri, 20 Jan 2017 15:27:56 -0500
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Subject: [AI Seminar] AI Lunch -- Hoda Heidari -- January 24
> Message-ID:
>         <CABzq7epF2kMt9QhhDxL7=ucen2oXb7fab-ee74=n4pAROjb1PQ@
> mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you Next Tuesday, January 24, at noon in NSH 3305
> for AI lunch. To learn more about the seminar and lunch, please visit
> the AI Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Hoda Heidari <http://www.seas.upenn.edu/~hoda/> from
>  University of Pennsylvania will give a talk titled ?Pricing a Low-regret
> Seller?.
>
> *Abstract:* As the number of ad exchanges has grown, publishers have turned
> to low regret learning algorithms to decide which exchange offers the best
> price for their inventory. This in turn opens the following question for
> the exchange: how to set prices to attract as many sellers as possible in
> order to maximize revenue. In this work, we formulate this precisely as a
> learning problem, and present algorithms showing that simply knowing the
> counterparty is using a low regret algorithm is enough for the exchange to
> have its own low regret learning algorithm for the optimal price.
>
> (Joint work with Mohammad Mahdian, Umar Syed, Sergei Vassilvistkii, and
> Sadra Yazdanbod. Appeared in ICML'16)
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20170120/c4b09bcf/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 70, Issue 3
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170124/686b8add/attachment.html>

From weiyu at cs.cmu.edu  Fri Jan 27 13:43:20 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Fri, 27 Jan 2017 13:43:20 -0500
Subject: [AI Seminar] AI Lunch -- Po-Wei Wang -- January 31
Message-ID: <CABzq7erCO_7iHYp0zr5EaNqt2sf7t+UkWPH6UEqWpOmaSO4RiQ@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you Next Tuesday, January 31, at noon in NSH 3305
for AI lunch. To learn more about the seminar and lunch, please visit
the AI Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Po-Wei Wang <http://www.powei.tw/> will give a talk titled
?Polynomial optimization methods for matrix factorization?.

*Abstract:* Matrix factorization is a core technique in many machine
learning problems, yet also presents a nonconvex and often
difficult-to-optimize problem. In this paper we present an approach based
upon polynomial optimization techniques that both improves the convergence
time of matrix factorization algorithms and helps them escape from local
optima. Our method is based on the realization that given a joint search
direction in a matrix factorization task, we can solve the ``subspace
search'' problem (the task of jointly finding the steps to take in each
direction) by solving a bivariate quartic polynomial optimization problem.
We derive two methods for solving this problem based upon sum of squares
moment relaxations and the Durand-Kerner method, then apply these
techniques on matrix factorization to derive a direct coordinate descent
approach and a method for speeding up existing approaches. On three
benchmark datasets we show the method substantially improves convergence
speed over state-of-the-art approaches, while also attaining lower
objective value.

This is a joint work with Chun-Liang Li and J. Zico Kolter. Forthcoming in
AAAI 2017.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170127/27133949/attachment.html>

From weiyu at cs.cmu.edu  Mon Jan 30 13:09:54 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 30 Jan 2017 13:09:54 -0500
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 70, Issue 5
In-Reply-To: <mailman.93.1485622820.9768.ai-seminar-announce@cs.cmu.edu>
References: <mailman.93.1485622820.9768.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7eojpbSrfS9P290z8yYssBvPetxP9AKig=3_+Pn07V2c1A@mail.gmail.com>

A gentle reminder that the talk will be tomorrow noon.

On Sat, Jan 28, 2017 at 12:00 PM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Lunch -- Po-Wei Wang -- January 31 (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Fri, 27 Jan 2017 13:43:20 -0500
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Subject: [AI Seminar] AI Lunch -- Po-Wei Wang -- January 31
> Message-ID:
>         <CABzq7erCO_7iHYp0zr5EaNqt2sf7t+UkWPH6UEqWpOmaSO4RiQ at mail.
> gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you Next Tuesday, January 31, at noon in NSH 3305
> for AI lunch. To learn more about the seminar and lunch, please visit
> the AI Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Po-Wei Wang <http://www.powei.tw/> will give a talk titled
> ?Polynomial optimization methods for matrix factorization?.
>
> *Abstract:* Matrix factorization is a core technique in many machine
> learning problems, yet also presents a nonconvex and often
> difficult-to-optimize problem. In this paper we present an approach based
> upon polynomial optimization techniques that both improves the convergence
> time of matrix factorization algorithms and helps them escape from local
> optima. Our method is based on the realization that given a joint search
> direction in a matrix factorization task, we can solve the ``subspace
> search'' problem (the task of jointly finding the steps to take in each
> direction) by solving a bivariate quartic polynomial optimization problem.
> We derive two methods for solving this problem based upon sum of squares
> moment relaxations and the Durand-Kerner method, then apply these
> techniques on matrix factorization to derive a direct coordinate descent
> approach and a method for speeding up existing approaches. On three
> benchmark datasets we show the method substantially improves convergence
> speed over state-of-the-art approaches, while also attaining lower
> objective value.
>
> This is a joint work with Chun-Liang Li and J. Zico Kolter. Forthcoming in
> AAAI 2017.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20170127/27133949/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 70, Issue 5
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170130/cda6a59e/attachment.html>

From weiyu at cs.cmu.edu  Fri Feb  3 05:41:29 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Fri, 3 Feb 2017 05:41:29 -0500
Subject: [AI Seminar] AI Lunch -- David Kurokawa -- February 7
Message-ID: <CABzq7eoxKD2DJhTC1vC717dj_5xKeTa02=1bYepnY8pvzaqf2A@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you Next Tuesday, February 7, at noon in NSH 3305
for AI lunch. To learn more about the seminar and lunch, please visit
the AI Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, David Kurokawa <http://www.cs.cmu.edu/~dkurokaw/> will give a
talk titled ?Fairness Notions in the Indivisible Good Setting: Comparisons
and their Approximations?.

*Abstract:*

Fair division of indivisible goods is the study of allocating a set of
discrete goods among several interested parties. Often in such settings a
desired allocation is hoped to satisfy some notion of fairness.  In this
talk we investigate several such notions studied in the literature: maximin
share guarantee (MMS), pairwise maximin share guarantee (PMMS), and
envy-freeness up to any good (EFX).

We begin by first defining MMS and exploring the pros and cons of it as the
benchmark for fairness in the setting. We then define PMMS and EFX and
demonstrate their potential as answers to these shortcomings of MMS. We
further demonstrate a hierarchical nature between these and relevant
notions from the literature --- namely envy-freeness (EF) and envy-freeness
up to one good (EF1) as well as give approximation existence results. We
close by examining the age-old method of drafting (such as in American
sports leagues) and show that there exist far fairer approaches to this
problem.

This is joint work with Ariel Procaccia and Junxing Wang.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170203/86433b97/attachment.html>

From weiyu at cs.cmu.edu  Mon Feb  6 13:55:21 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 6 Feb 2017 13:55:21 -0500
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 71, Issue 1
In-Reply-To: <mailman.81.1486141218.17375.ai-seminar-announce@cs.cmu.edu>
References: <mailman.81.1486141218.17375.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7epjjr3uQY-E2kJ-R4W7jyGqNtMyh-p_eWBNfSpSsHgjTw@mail.gmail.com>

Just gentle reminder that the talk will be at tomorrow noon.

On Fri, Feb 3, 2017 at 12:00 PM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Lunch -- David Kurokawa -- February 7 (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Fri, 3 Feb 2017 05:41:29 -0500
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Subject: [AI Seminar] AI Lunch -- David Kurokawa -- February 7
> Message-ID:
>         <CABzq7eoxKD2DJhTC1vC717dj_5xKeTa02=1bYepnY8pvzaqf2A@
> mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you Next Tuesday, February 7, at noon in NSH 3305
> for AI lunch. To learn more about the seminar and lunch, please visit
> the AI Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, David Kurokawa <http://www.cs.cmu.edu/~dkurokaw/> will give a
> talk titled ?Fairness Notions in the Indivisible Good Setting: Comparisons
> and their Approximations?.
>
> *Abstract:*
>
> Fair division of indivisible goods is the study of allocating a set of
> discrete goods among several interested parties. Often in such settings a
> desired allocation is hoped to satisfy some notion of fairness.  In this
> talk we investigate several such notions studied in the literature: maximin
> share guarantee (MMS), pairwise maximin share guarantee (PMMS), and
> envy-freeness up to any good (EFX).
>
> We begin by first defining MMS and exploring the pros and cons of it as the
> benchmark for fairness in the setting. We then define PMMS and EFX and
> demonstrate their potential as answers to these shortcomings of MMS. We
> further demonstrate a hierarchical nature between these and relevant
> notions from the literature --- namely envy-freeness (EF) and envy-freeness
> up to one good (EF1) as well as give approximation existence results. We
> close by examining the age-old method of drafting (such as in American
> sports leagues) and show that there exist far fairer approaches to this
> problem.
>
> This is joint work with Ariel Procaccia and Junxing Wang.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20170203/86433b97/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 71, Issue 1
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170206/bd038cc2/attachment.html>

From weiyu at cs.cmu.edu  Mon Feb 13 15:57:55 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 13 Feb 2017 15:57:55 -0500
Subject: [AI Seminar] No AI Lunch this week
Message-ID: <CABzq7eq=d3H4hKy--+yBFT+jkRdf73zOJ45EJ9Meh78EVoxWnA@mail.gmail.com>

There is no AI lunch this week.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170213/a83a8620/attachment.html>

From weiyu at cs.cmu.edu  Fri Feb 17 11:39:21 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Fri, 17 Feb 2017 11:39:21 -0500
Subject: [AI Seminar] AI Lunch -- Dirk Bergemann (Yale) -- February 21
Message-ID: <CABzq7eq0XQWwe3D4Fm5vkJ8upTCeQbgGK7D-1UpHakkB+YH78Q@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you Next Tuesday, February 21, at noon in NSH
3305 for AI lunch. To learn more about the seminar and lunch, please visit
the AI Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Professor Dirk Bergemann
<https://campuspress.yale.edu/dirkbergemann/> from Yale will give a talk
titled ?Information Design: A Unified Perspective?.

*Abstract:* Fixing a game with uncertain payoffs, information design
identifies the information structure and equilibrium that maximizes the
payoff of an information designer. We show how this perspective unifies
existing work, including that on communication in games (Myerson (1991)),
Bayesian persuasion (Kamenica and Genzkow (2011)) and some of our own
recent work. Information design has a literal interpretation, under which
there is a real information designer who can commit to the choice of the
best information structure (from her perspective) for a set of participants
in a game. We emphasize a metaphorical interpretation, under which the
information design problem is used by the analyst to characterize play in
the game under many different information structures.


*Bio:* Dirk Bergemann is Douglass and Marion Campbell Professor of
Economics at Yale University. He has secondary appointments as Professor of
Computer Science at the School of Engineering and Professor of Finance at
the School of Management. He received his Vordiplom in Economics at J.W.
Goethe University in Frankfurt in 1989 and his Ph.D. in Economics from the
University of Pennsylvania in 1994. Dirk Bergemann is the Chair of the
Department of Economics since 2013. He joined Yale in 1995 as an assistant
professor, having previously served as a faculty member at Princeton
University. He has been affiliated with the Cowles Foundation for Research
in Economics at Yale since 1996 and a fellow of the Econometric Society
since 2007. His research is concerned with game theory, contract theory and
mechanism design. His research has been supported by grants from the
National Science Foundation, the Alfred P. Sloan Research Fellowship and
the German National Science Foundation. Dirk Bergemann is currently a
Co-Editor of Econometrica.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170217/3361118c/attachment.html>

From weiyu at cs.cmu.edu  Mon Feb 20 12:29:10 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 20 Feb 2017 12:29:10 -0500
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 71, Issue 4
In-Reply-To: <mailman.115.1487350821.6596.ai-seminar-announce@cs.cmu.edu>
References: <mailman.115.1487350821.6596.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7er2Uh+CONVia_2vBGNDGxvdy4JY8+XgrJ6YfhsqMqqnFQ@mail.gmail.com>

This is a gentle reminder that the talk will be tomorrow (Tuesday) noon.

On Fri, Feb 17, 2017 at 12:00 PM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Lunch -- Dirk Bergemann (Yale) -- February 21 (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Fri, 17 Feb 2017 11:39:21 -0500
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Subject: [AI Seminar] AI Lunch -- Dirk Bergemann (Yale) -- February 21
> Message-ID:
>         <CABzq7eq0XQWwe3D4Fm5vkJ8upTCeQbgGK7D-1UpHakkB+YH78Q at mail.
> gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you Next Tuesday, February 21, at noon in NSH
> 3305 for AI lunch. To learn more about the seminar and lunch, please visit
> the AI Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Professor Dirk Bergemann
> <https://campuspress.yale.edu/dirkbergemann/> from Yale will give a talk
> titled ?Information Design: A Unified Perspective?.
>
> *Abstract:* Fixing a game with uncertain payoffs, information design
> identifies the information structure and equilibrium that maximizes the
> payoff of an information designer. We show how this perspective unifies
> existing work, including that on communication in games (Myerson (1991)),
> Bayesian persuasion (Kamenica and Genzkow (2011)) and some of our own
> recent work. Information design has a literal interpretation, under which
> there is a real information designer who can commit to the choice of the
> best information structure (from her perspective) for a set of participants
> in a game. We emphasize a metaphorical interpretation, under which the
> information design problem is used by the analyst to characterize play in
> the game under many different information structures.
>
>
> *Bio:* Dirk Bergemann is Douglass and Marion Campbell Professor of
> Economics at Yale University. He has secondary appointments as Professor of
> Computer Science at the School of Engineering and Professor of Finance at
> the School of Management. He received his Vordiplom in Economics at J.W.
> Goethe University in Frankfurt in 1989 and his Ph.D. in Economics from the
> University of Pennsylvania in 1994. Dirk Bergemann is the Chair of the
> Department of Economics since 2013. He joined Yale in 1995 as an assistant
> professor, having previously served as a faculty member at Princeton
> University. He has been affiliated with the Cowles Foundation for Research
> in Economics at Yale since 1996 and a fellow of the Econometric Society
> since 2007. His research is concerned with game theory, contract theory and
> mechanism design. His research has been supported by grants from the
> National Science Foundation, the Alfred P. Sloan Research Fellowship and
> the German National Science Foundation. Dirk Bergemann is currently a
> Co-Editor of Econometrica.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20170217/3361118c/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 71, Issue 4
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170220/a66bee0f/attachment.html>

From weiyu at cs.cmu.edu  Tue Feb 21 11:56:25 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Tue, 21 Feb 2017 11:56:25 -0500
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 71, Issue 4
In-Reply-To: <mailman.115.1487350821.6596.ai-seminar-announce@cs.cmu.edu>
References: <mailman.115.1487350821.6596.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7er0faRXXL-N1uv-fyNJ+-YQTs2DBZVZzbUV5PYrJ+Sbeg@mail.gmail.com>

The talk happens now!

On Fri, Feb 17, 2017 at 12:00 PM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Lunch -- Dirk Bergemann (Yale) -- February 21 (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Fri, 17 Feb 2017 11:39:21 -0500
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Subject: [AI Seminar] AI Lunch -- Dirk Bergemann (Yale) -- February 21
> Message-ID:
>         <CABzq7eq0XQWwe3D4Fm5vkJ8upTCeQbgGK7D-1UpHakkB+YH78Q at mail.
> gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you Next Tuesday, February 21, at noon in NSH
> 3305 for AI lunch. To learn more about the seminar and lunch, please visit
> the AI Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Professor Dirk Bergemann
> <https://campuspress.yale.edu/dirkbergemann/> from Yale will give a talk
> titled ?Information Design: A Unified Perspective?.
>
> *Abstract:* Fixing a game with uncertain payoffs, information design
> identifies the information structure and equilibrium that maximizes the
> payoff of an information designer. We show how this perspective unifies
> existing work, including that on communication in games (Myerson (1991)),
> Bayesian persuasion (Kamenica and Genzkow (2011)) and some of our own
> recent work. Information design has a literal interpretation, under which
> there is a real information designer who can commit to the choice of the
> best information structure (from her perspective) for a set of participants
> in a game. We emphasize a metaphorical interpretation, under which the
> information design problem is used by the analyst to characterize play in
> the game under many different information structures.
>
>
> *Bio:* Dirk Bergemann is Douglass and Marion Campbell Professor of
> Economics at Yale University. He has secondary appointments as Professor of
> Computer Science at the School of Engineering and Professor of Finance at
> the School of Management. He received his Vordiplom in Economics at J.W.
> Goethe University in Frankfurt in 1989 and his Ph.D. in Economics from the
> University of Pennsylvania in 1994. Dirk Bergemann is the Chair of the
> Department of Economics since 2013. He joined Yale in 1995 as an assistant
> professor, having previously served as a faculty member at Princeton
> University. He has been affiliated with the Cowles Foundation for Research
> in Economics at Yale since 1996 and a fellow of the Econometric Society
> since 2007. His research is concerned with game theory, contract theory and
> mechanism design. His research has been supported by grants from the
> National Science Foundation, the Alfred P. Sloan Research Fellowship and
> the German National Science Foundation. Dirk Bergemann is currently a
> Co-Editor of Econometrica.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20170217/3361118c/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 71, Issue 4
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170221/e2b83f2d/attachment-0001.html>

From weiyu at cs.cmu.edu  Mon Feb 27 13:35:36 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 27 Feb 2017 13:35:36 -0500
Subject: [AI Seminar] No AI Lunch this week
Message-ID: <CABzq7eqRVTBkkZEu9xwpbaXtTMhx=E8gBQko+3gpqciAXeZPDQ@mail.gmail.com>

There is no AI lunch this week.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170227/181f6a8a/attachment.html>

From mmv at cs.cmu.edu  Fri Mar  3 08:26:28 2017
From: mmv at cs.cmu.edu (Manuela Veloso)
Date: Fri, 3 Mar 2017 08:26:28 -0500
Subject: [AI Seminar] Fwd: ML Seminar - Francesca Rossi - March 3,
	2017@ 10am in NSH 3305
In-Reply-To: <f9d78a44-a90e-3023-4707-7ecbf117fce8@cs.cmu.edu>
References: <f9d78a44-a90e-3023-4707-7ecbf117fce8@cs.cmu.edu>
Message-ID: <8aa0ea7a-8668-f176-ca31-25c36dc576a6@cs.cmu.edu>

Friendly reminder!

      Manuela




-------- Forwarded Message --------
Subject: 	ML Seminar - Francesca Rossi - March 3, 2017@ 10am in NSH 3305
Date: 	Wed, 1 Mar 2017 11:40:35 -0500
From: 	Sharon Cavlovich <sharonw+ at cs.cmu.edu>
To: 	ml-seminar at cs.cmu.edu



Please join us for a special ML seminar on March 3rd!


If you would like to meet with Francesca during her visit, please go to 
https://docs.google.com/spreadsheets/d/1CgzuUooUG3hICyVUyWO6CtNI-fZmrWo-WqVhYbQjUZA/edit?usp=sharing 
to sign up for meetings.

Thanks,
Sharon


March 3, 2017
10:00am
NSH 3305


Francesca Rossi, Professor -  University of Padova


Title: Ethical AI


Abstract:

Intelligent systems are going to be pervasive in our everyday lives. 
They will take care of elderly people and educate kids, they will drive 
for us, and they will help doctors understand how to cure a disease. 
However, in order to get all their beneficial effects in these and other 
scenarios, we need to trust that they can discriminate between good and 
bad decisions. In other words, that they function according to ethical 
principles, professional codes, social norms, and moral values that are 
aligned to those of humans. Therefore it is important to understand how 
to embed ethical principles into intelligent machines, as well as to 
assess the ethical capabilities of an AI system.

In this talk I will review the main approaches and research efforts 
related to AI ethics, with special attention to the possible use of 
preference-based frameworks to model and reason with ethical principles 
in the context of individual or group decision making.

I will also discuss the main initiatives around the ethical development 
and deployment of AI systems, describing their goals and potential in 
building beneficial AI.

Affiliation: IBM Research and University of Padova

Short bio:
Francesca Rossi is a distinguished research scientist at the IBM T.J. 
Watson Research Centre, and an professor of computer science at 
the University of Padova, Italy, currently on leave.
Her research interests focus on artificial intelligence, specifically 
they include constraint reasoning, preferences, multi-agent systems, 
computational social choice, and collective decision making. She is also 
interested in ethical issues in the development and behaviour of AI 
systems, in particular for decision support systems for group decision 
making. She has published over 170 scientific articles in journals and 
conference proceedings, and as book chapters. She has co-authored a 
book. She has edited 17 volumes, between conference proceedings, 
collections of contributions, special issues of journals, as well as 
the Handbook of Constraint Programming.  She has more than 100 co-authors.

She is a AAAI and a EurAI fellow, and a Radcliffe fellow 2015. She has 
been president of IJCAI and an executive councillor of AAAI. She is 
Associate Editor in Chief of JAIR and a member of the editorial board 
of Constraints, Artificial Intelligence, AMAI, and KAIS. She co-chairs 
the AAAI committee on AI and ethics and she is a member of the 
scientific advisory board of the Future of Life Institute. She is in the 
executive committee of the IEEE global initiative on ethical 
considerations on the development of autonomous and intelligent systems 
and she belongs to the World Economic Forum Council on AI and robotics.
She has given several media interviews about the future of AI and AI 
ethics (including to the Wall Street Journal, the Washington Post, 
Motherboard, Science, The Economist, CNBC, Eurovision, Corriere della 
Sera, and Repubblica) and she has delivered three TEDx talks on these 
topics.

-- 
Sharon Cavlovich
Senior Administrative Assistant to the Department Head | Machine Learning Department

Carnegie Mellon University
5000 Forbes Avenue | Gates Hillman Complex 8007
Pittsburgh, PA 15213
412.268.5196 (office) | 412.268.2205 (fax)

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170303/3f80f1fb/attachment-0001.html>

From weiyu at cs.cmu.edu  Fri Mar  3 09:49:12 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Fri, 3 Mar 2017 09:49:12 -0500
Subject: [AI Seminar] AI Lunch -- Han Zhao -- March 7
Message-ID: <CABzq7erGyLk8kZDftzUkpC0ZEDQbHmb3qd4+0=fDPnzR_5_Jwg@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you Next Tuesday, March 7, at noon in NSH 3305
for AI lunch. To learn more about the seminar and lunch, please visit
the AI Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Han Zhao <https://www.cs.cmu.edu/~hzhao1/> will give a talk
titled ?Sum-Product Networks: A New Probabilistic Inference Machine?.

*Abstract:* Sum-product networks (SPNs) are new deep inference machines
that admit exact probabilistic inference in linear time in the size of the
network. In this talk I will establish some theoretical connections between
SPNs and traditional graphical models like Bayesian Networks (BNs).
Specifically, I will show that every SPN can be converted into a BN in
linear time and space in terms of the network size. The key insight is to
use Algebraic Decision Diagrams (ADDs) to compactly represent the local
conditional probability distributions at each node in the resulting BN by
exploiting context-specific independence (CSI). The generated BN has a
simple directed bipartite graphical structure. I will also discuss some
implications of the proof and establish a connection between the depth of
an SPN and a lower bound of the tree-width of its corresponding BN. I will
conclude the talk by discussing some algorithms for learning the parameters
of SPNs based on maximum-likelihood principle and Bayesian approaches.

This is joint work with Geoff Gordon, Mazen Melibari and Pascal Poupart.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170303/b5eb6671/attachment.html>

From weiyu at cs.cmu.edu  Mon Mar  6 13:21:43 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 6 Mar 2017 13:21:43 -0500
Subject: [AI Seminar] AI Lunch -- Han Zhao -- March 7
In-Reply-To: <CABzq7erGyLk8kZDftzUkpC0ZEDQbHmb3qd4+0=fDPnzR_5_Jwg@mail.gmail.com>
References: <CABzq7erGyLk8kZDftzUkpC0ZEDQbHmb3qd4+0=fDPnzR_5_Jwg@mail.gmail.com>
Message-ID: <CABzq7epu1+vhYJeNahPKPuH1FW-A+ag9aXyJ8iYHR2wkbdhFAg@mail.gmail.com>

A gentle reminder that the talk will be tomorrow (Tuesday) noon.

On Fri, Mar 3, 2017 at 9:49 AM, Adams Wei Yu <weiyu at cs.cmu.edu> wrote:

> Dear faculty and students,
>
> We look forward to seeing you Next Tuesday, March 7, at noon in NSH 3305
> for AI lunch. To learn more about the seminar and lunch, please visit
> the AI Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Han Zhao <https://www.cs.cmu.edu/~hzhao1/> will give a talk
> titled ?Sum-Product Networks: A New Probabilistic Inference Machine?.
>
> *Abstract:* Sum-product networks (SPNs) are new deep inference machines
> that admit exact probabilistic inference in linear time in the size of the
> network. In this talk I will establish some theoretical connections between
> SPNs and traditional graphical models like Bayesian Networks (BNs).
> Specifically, I will show that every SPN can be converted into a BN in
> linear time and space in terms of the network size. The key insight is to
> use Algebraic Decision Diagrams (ADDs) to compactly represent the local
> conditional probability distributions at each node in the resulting BN by
> exploiting context-specific independence (CSI). The generated BN has a
> simple directed bipartite graphical structure. I will also discuss some
> implications of the proof and establish a connection between the depth of
> an SPN and a lower bound of the tree-width of its corresponding BN. I will
> conclude the talk by discussing some algorithms for learning the parameters
> of SPNs based on maximum-likelihood principle and Bayesian approaches.
>
> This is joint work with Geoff Gordon, Mazen Melibari and Pascal Poupart.
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170306/4fd3c74c/attachment.html>

From weiyu at cs.cmu.edu  Fri Mar 10 11:48:35 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Fri, 10 Mar 2017 11:48:35 -0500
Subject: [AI Seminar] AI Lunch -- Aaditya Ramdas(UC Berkeley) -- March 14
Message-ID: <CABzq7eoGfHenP+ZtumPdgHd7O-WGuXMBz4OKo3o86=j+5QW4sg@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you Next Tuesday, March 14, at noon in NSH 3305
for AI lunch. To learn more about the seminar and lunch, please visit
the AI Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Aaditya Ramdas <http://people.eecs.berkeley.edu/~aramdas/> from
UC Berkeley will give a talk titled *Multi A(rmed)/B(andit) Testing with
online FDR control*.

*Abstract*: We propose a new framework as an alternative to existing setups
for controlling false alarms across multiple A/B tests; it combines ideas
from pure exploration for best-arm identification in multi-armed bandits
(MAB), with online false discovery rate (FDR) control. This framework has
various applications, including pharmaceutical companies testing a control
pill against a few treatment options, to internet companies testing their
current default webpage (control) versus many alternatives (treatment). Our
setup involves running a possibly infinite sequence of best-arm MAB
instances, and controlling the overall FDR of the process in a fully online
manner. Our main contributions are: (i) to propose reasonable definitions
for a null hypothesis; (ii) to demonstrate how one can derive an
always-valid sequential p-value for such a null hypothesis which allows
users to continuously monitor and stop any running MAB instance at any
time; and (iii) to embed MAB instances within online FDR algorithms in a
way that allows setting MAB confidence-levels based on FDR rejection
thresholds. In addition, we adapt existing theory from both the MAB and
online FDR literature to ensure that our framework comes with strong
sample-optimality guarantees, as well as control of the power and (a
modified) FDR at any time. We run extensive simulations to verify our
claims and report results on real data collected from the New Yorker
Cartoon Caption contest.

Joint work with Fan Yang, Kevin Jamieson, Martin Wainwright.

*Bio*: Aaditya Ramdas is a postdoctoral researcher in Statistics and EECS
at UC Berkeley, advised by Michael Jordan and Martin Wainwright. He
finished his PhD in Statistics and Machine Learning at CMU, advised by
Larry Wasserman and Aarti Singh. A lot of his research focuses on modern
aspects of reproducibility in science and technology -- involving
statistical testing and false discovery rate control in static and dynamic
settings.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170310/f4e9ab47/attachment.html>

From weiyu at cs.cmu.edu  Mon Mar 13 12:20:44 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 13 Mar 2017 12:20:44 -0400
Subject: [AI Seminar] AI Lunch -- Aaditya Ramdas(UC Berkeley) -- March 14
In-Reply-To: <CABzq7eoGfHenP+ZtumPdgHd7O-WGuXMBz4OKo3o86=j+5QW4sg@mail.gmail.com>
References: <CABzq7eoGfHenP+ZtumPdgHd7O-WGuXMBz4OKo3o86=j+5QW4sg@mail.gmail.com>
Message-ID: <CABzq7epMBxXgWkzPeNCF+Uryg-5Q1_9g-T2h56yfRHWowB7DFw@mail.gmail.com>

A gentle reminder that the talk will be tomorrow (Tuesday) noon.

On Fri, Mar 10, 2017 at 11:48 AM, Adams Wei Yu <weiyu at cs.cmu.edu> wrote:

> Dear faculty and students,
>
> We look forward to seeing you Next Tuesday, March 14, at noon in NSH 3305
> for AI lunch. To learn more about the seminar and lunch, please visit
> the AI Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Aaditya Ramdas <http://people.eecs.berkeley.edu/~aramdas/> from
> UC Berkeley will give a talk titled *Multi A(rmed)/B(andit) Testing with
> online FDR control*.
>
> *Abstract*: We propose a new framework as an alternative to existing
> setups for controlling false alarms across multiple A/B tests; it combines
> ideas from pure exploration for best-arm identification in multi-armed
> bandits (MAB), with online false discovery rate (FDR) control. This
> framework has various applications, including pharmaceutical companies
> testing a control pill against a few treatment options, to internet
> companies testing their current default webpage (control) versus many
> alternatives (treatment). Our setup involves running a possibly infinite
> sequence of best-arm MAB instances, and controlling the overall FDR of the
> process in a fully online manner. Our main contributions are: (i) to
> propose reasonable definitions for a null hypothesis; (ii) to demonstrate
> how one can derive an always-valid sequential p-value for such a null
> hypothesis which allows users to continuously monitor and stop any running
> MAB instance at any time; and (iii) to embed MAB instances within online
> FDR algorithms in a way that allows setting MAB confidence-levels based on
> FDR rejection thresholds. In addition, we adapt existing theory from both
> the MAB and online FDR literature to ensure that our framework comes with
> strong sample-optimality guarantees, as well as control of the power and (a
> modified) FDR at any time. We run extensive simulations to verify our
> claims and report results on real data collected from the New Yorker
> Cartoon Caption contest.
>
> Joint work with Fan Yang, Kevin Jamieson, Martin Wainwright.
>
> *Bio*: Aaditya Ramdas is a postdoctoral researcher in Statistics and EECS
> at UC Berkeley, advised by Michael Jordan and Martin Wainwright. He
> finished his PhD in Statistics and Machine Learning at CMU, advised by
> Larry Wasserman and Aarti Singh. A lot of his research focuses on modern
> aspects of reproducibility in science and technology -- involving
> statistical testing and false discovery rate control in static and dynamic
> settings.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170313/5d7d0b76/attachment.html>

From weiyu at cs.cmu.edu  Fri Mar 17 23:34:07 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Fri, 17 Mar 2017 23:34:07 -0400
Subject: [AI Seminar] AI Lunch -- Wen Sun -- March 21 (Unusual Room: NSH
	1507)
Message-ID: <CABzq7eq+SV=czupGEGXK_XVYmeNwvRgH9WnNessGpAn9BeT1nw@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you Next Tuesday, March 21, at noon in *NSH 1507*
for AI lunch. To learn more about the seminar and lunch, please visit the AI
 Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Wen Sun <http://www.cs.cmu.edu/~wensun/> will give a talk
titled *Differentiable Imitation Learning and Sequential Prediction*.

*Abstract*: Recently, researchers have demonstrated state-of-the-art
performance on sequential decision making problems (e.g., robotics control,
sequential prediction) with deep neural networks and Reinforcement Learning
(RL). However, for some of these problems, oracles that can demonstrate
good performance are available during training. In this work, we propose
AggreVaTeD, a policy gradient extension of the Imitation Learning (IL)
approach of Ross & Bagnell (2014) that can leverage oracles to achieve
faster and more accurate solutions with less training data than with a
less-informed RL approaches. Specifically, we provide a comprehensive
theoretical study of IL that demonstrates we can expect up to exponentially
lower sample complexity for learning with AggreVaTeD than with RL
algorithms. Finally, we present two stochastic gradient procedures that
learn neural network policies for several problems including a sequential
prediction task as well as various high dimensional robotics control
problems. Our results and theory indicate that the proposed approach can
achieve superior performance with respect to the oracle when the
demonstrator is sub-optimal.

This a joint work with Arun Venkatraman, Geoff Gordon, Byron Boots and Drew
Bagnell.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170317/371f8226/attachment.html>

From weiyu at cs.cmu.edu  Mon Mar 20 08:14:00 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 20 Mar 2017 08:14:00 -0400
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 72, Issue 6
In-Reply-To: <mailman.79.1489852818.7627.ai-seminar-announce@cs.cmu.edu>
References: <mailman.79.1489852818.7627.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7epiOR6RdmchG=MMs+XxPHcY=cv=gsecRAaC1wMyTpUGyQ@mail.gmail.com>

A gentle reminder that the talk is tomorrow (Tuesday) noon in NSH 1507!

On Sat, Mar 18, 2017 at 12:00 PM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Lunch -- Wen Sun -- March 21 (Unusual Room: NSH       1507)
>       (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Fri, 17 Mar 2017 23:34:07 -0400
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Subject: [AI Seminar] AI Lunch -- Wen Sun -- March 21 (Unusual Room:
>         NSH     1507)
> Message-ID:
>         <CABzq7eq+SV=czupGEGXK_XVYmeNwvRgH9WnNessGpAn9BeT1nw@
> mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you Next Tuesday, March 21, at noon in *NSH 1507*
> for AI lunch. To learn more about the seminar and lunch, please visit the
> AI
>  Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Wen Sun <http://www.cs.cmu.edu/~wensun/> will give a talk
> titled *Differentiable Imitation Learning and Sequential Prediction*.
>
> *Abstract*: Recently, researchers have demonstrated state-of-the-art
> performance on sequential decision making problems (e.g., robotics control,
> sequential prediction) with deep neural networks and Reinforcement Learning
> (RL). However, for some of these problems, oracles that can demonstrate
> good performance are available during training. In this work, we propose
> AggreVaTeD, a policy gradient extension of the Imitation Learning (IL)
> approach of Ross & Bagnell (2014) that can leverage oracles to achieve
> faster and more accurate solutions with less training data than with a
> less-informed RL approaches. Specifically, we provide a comprehensive
> theoretical study of IL that demonstrates we can expect up to exponentially
> lower sample complexity for learning with AggreVaTeD than with RL
> algorithms. Finally, we present two stochastic gradient procedures that
> learn neural network policies for several problems including a sequential
> prediction task as well as various high dimensional robotics control
> problems. Our results and theory indicate that the proposed approach can
> achieve superior performance with respect to the oracle when the
> demonstrator is sub-optimal.
>
> This a joint work with Arun Venkatraman, Geoff Gordon, Byron Boots and Drew
> Bagnell.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20170317/371f8226/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 72, Issue 6
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170320/87208ef9/attachment.html>

From weiyu at cs.cmu.edu  Sun Mar 26 14:04:09 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Sun, 26 Mar 2017 14:04:09 -0400
Subject: [AI Seminar] AI Lunch -- Manzil Zaheer -- March 28
Message-ID: <CABzq7eoKr=tUOybWFToZ81vt32rJv1zqpZoRWLV=g_5ZKJbJ6w@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you Next Tuesday, March 28, at noon in NSH 3305
for AI lunch. To learn more about the seminar and lunch, please visit the AI
 Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Manzil Zaheer <http://manzil.ml/> will give a talk titled
*Exponential Stochastic Cellular Automata For Massively Parallel Inference*.

*Abstract*:

Often statistical models and inference procedures thereof are directly not
good fit for the modern computational resources. To elaborate, current
computational resources are racks of fast, cheap, and heavily multicored
machines yet with a limited memory bandwidth whereas inference strategies
can be inherently sequentially like Gibbs sampling or memory access
intensive like expectation-maximization or other variational inference.

In this talk, we discuss an embarrassingly parallel, memory efficient
inference algorithm for latent variable models in which the complete data
likelihood is in the exponential family. The algorithm is a stochastic
cellular automaton and converges to a valid maximum a posteriori fixed
point. We explore further tricks to improve performance by reducing
pressure on memory bandwidth by use of better data structures.
We apply the algorithm to Gaussian mixture model (GMM) and latent Dirichlet
allocation (LDA) and empirically find that our algorithm is order of
magnitudes faster than state-of-the-art approaches. A simple C++/MPI
implementation on a 16-node cluster can sample more than a billion tokens
per second in case of LDA and a million images in case of GMM.

This is a joint work with Alex Smola, Jean-Baptiste Tristan, Michael Wick,
and Satwik Kottur.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170326/7075dabc/attachment.html>

From weiyu at cs.cmu.edu  Mon Mar 27 13:31:05 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 27 Mar 2017 13:31:05 -0400
Subject: [AI Seminar] AI Lunch -- Manzil Zaheer -- March 28
In-Reply-To: <CABzq7eoKr=tUOybWFToZ81vt32rJv1zqpZoRWLV=g_5ZKJbJ6w@mail.gmail.com>
References: <CABzq7eoKr=tUOybWFToZ81vt32rJv1zqpZoRWLV=g_5ZKJbJ6w@mail.gmail.com>
Message-ID: <CABzq7epjDv8EHGW6HOLX+roR55LCbo1a9dBE2Rv+09KTfJSvYw@mail.gmail.com>

Just a gentle reminder that the talk would be tomorrow (Tuesday) noon.

On Sun, Mar 26, 2017 at 2:04 PM, Adams Wei Yu <weiyu at cs.cmu.edu> wrote:

> Dear faculty and students,
>
> We look forward to seeing you Next Tuesday, March 28, at noon in NSH 3305
> for AI lunch. To learn more about the seminar and lunch, please visit the
> AI Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Manzil Zaheer <http://manzil.ml/> will give a talk titled
> *Exponential Stochastic Cellular Automata For Massively Parallel Inference*.
>
> *Abstract*:
>
> Often statistical models and inference procedures thereof are directly not
> good fit for the modern computational resources. To elaborate, current
> computational resources are racks of fast, cheap, and heavily multicored
> machines yet with a limited memory bandwidth whereas inference strategies
> can be inherently sequentially like Gibbs sampling or memory access
> intensive like expectation-maximization or other variational inference.
>
> In this talk, we discuss an embarrassingly parallel, memory efficient
> inference algorithm for latent variable models in which the complete data
> likelihood is in the exponential family. The algorithm is a stochastic
> cellular automaton and converges to a valid maximum a posteriori fixed
> point. We explore further tricks to improve performance by reducing
> pressure on memory bandwidth by use of better data structures.
> We apply the algorithm to Gaussian mixture model (GMM) and latent
> Dirichlet allocation (LDA) and empirically find that our algorithm is order
> of magnitudes faster than state-of-the-art approaches. A simple C++/MPI
> implementation on a 16-node cluster can sample more than a billion tokens
> per second in case of LDA and a million images in case of GMM.
>
> This is a joint work with Alex Smola, Jean-Baptiste Tristan, Michael Wick,
> and Satwik Kottur.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170327/93b4e30c/attachment.html>

From weiyu at cs.cmu.edu  Sun Apr  2 14:04:15 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Sun, 2 Apr 2017 14:04:15 -0400
Subject: [AI Seminar] AI Lunch -- Hsiao-Yu Tung -- April 4
Message-ID: <CABzq7eo=jxb88YJuDecJU_==kGnf9nRFfRQB0nhY9mnCAn_2vg@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you Next Tuesday, April 4, at noon in NSH 3305
for AI lunch. To learn more about the seminar and lunch, please visit the AI
 Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Hsiao-Yu Tung <https://sfish0101.bitbucket.io/> will give the
following talk:

Title: Adversarial Inversion: Self-supervision with Adversarial Priors.

Abstract:

We as humans form explanations of visual observations in terms of familiar
concepts and memories that are used to interpret and complete information
of the image pixels. Computer Vision researchers have developed excellent
methods that learn a direct  mapping  from  images  to  desired outputs
using human annotations or synthetically generated data. Despite their
 success,  such  supervised  models  very  much  depend on the amount of
annotated data available, a gap we seek to address.

In this talks, we introduce adversarial inversion, a weakly supervised
neural network model that combines self-supervision with adversarial
constraints. Given visual input, our model first generates a set of
desirable intermediate latent variables, which we call ?imaginations?,
e.g.,  3D pose and camera viewpoint, such that these imagination matches
what we observe. Adversarial inversion can be trained with or without
paired supervision of standard supervised models, as it does not require
paired annotations.  It can instead exploit a large number of unlabelled
 images. We  empirically  show  adversarial  inversion outperforms previous
state-of-the-art supervised models on 3D human pose estimation and 3D scene
depth estimation. Further, we show interesting results on biased image
editing.


Joint work with Adam Harley, William Seto and Katerina Fragkiadaki.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170402/fd69c97a/attachment.html>

From weiyu at cs.cmu.edu  Tue Apr  4 10:40:24 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Tue, 4 Apr 2017 10:40:24 -0400
Subject: [AI Seminar] AI Lunch -- Hsiao-Yu Tung -- April 4
In-Reply-To: <CABzq7eo=jxb88YJuDecJU_==kGnf9nRFfRQB0nhY9mnCAn_2vg@mail.gmail.com>
References: <CABzq7eo=jxb88YJuDecJU_==kGnf9nRFfRQB0nhY9mnCAn_2vg@mail.gmail.com>
Message-ID: <CABzq7er+zcYm7w2vh+nuQFCB9k10QoSpunJAe5Gb-fQz6tc4Kg@mail.gmail.com>

A reminder that the talk is today (Tuesday) noon.

On Sun, Apr 2, 2017 at 2:04 PM, Adams Wei Yu <weiyu at cs.cmu.edu> wrote:

> Dear faculty and students,
>
> We look forward to seeing you Next Tuesday, April 4, at noon in NSH 3305
> for AI lunch. To learn more about the seminar and lunch, please visit the
> AI Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Hsiao-Yu Tung <https://sfish0101.bitbucket.io/> will give the
> following talk:
>
> Title: Adversarial Inversion: Self-supervision with Adversarial Priors.
>
> Abstract:
>
> We as humans form explanations of visual observations in terms of familiar
> concepts and memories that are used to interpret and complete information
> of the image pixels. Computer Vision researchers have developed excellent
> methods that learn a direct  mapping  from  images  to  desired outputs
> using human annotations or synthetically generated data. Despite their
>  success,  such  supervised  models  very  much  depend on the amount of
> annotated data available, a gap we seek to address.
>
> In this talks, we introduce adversarial inversion, a weakly supervised
> neural network model that combines self-supervision with adversarial
> constraints. Given visual input, our model first generates a set of
> desirable intermediate latent variables, which we call ?imaginations?,
> e.g.,  3D pose and camera viewpoint, such that these imagination matches
> what we observe. Adversarial inversion can be trained with or without
> paired supervision of standard supervised models, as it does not require
> paired annotations.  It can instead exploit a large number of unlabelled
>  images. We  empirically  show  adversarial  inversion outperforms previous
> state-of-the-art supervised models on 3D human pose estimation and 3D scene
> depth estimation. Further, we show interesting results on biased image
> editing.
>
>
> Joint work with Adam Harley, William Seto and Katerina Fragkiadaki.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170404/81dcd43a/attachment.html>

From arielpro at cs.cmu.edu  Wed Apr  5 18:56:24 2017
From: arielpro at cs.cmu.edu (Ariel Procaccia)
Date: Wed, 5 Apr 2017 18:56:24 -0400
Subject: [AI Seminar] Meetings with Iyad Rahwan on April 18
Message-ID: <CADqQPTOLbjj_NYX5JT3v5AU2te9Ok6hnB-2WqFx_z+zi5Kt9_g@mail.gmail.com>

Hi all,

Iyad Rahwan of the MIT Media Lab will give the AI seminar talk on April 18
(title and abstract below). He'll be available for meetings on the day of
the talk. If you'd like to meet him, please email me your constraints.

Cheers,
Ariel

=======

TITLE: The Psychological Dilemmas of Autonomous Vehicles

ABSTRACT:

Autonomous vehicles promise to revolutionize transportation and
substantially improve safety. But they also made salient an  ethical
question: how should an algorithm decide relative risk in situations of
unavoidable harm? I present a series of psychological  studies that explore
how people think about such scenarios, and discuss  the regulatory
challenges these studies reveal. I also describe a crowdsourcing effort
that engaged millions of people worldwide with the  ethical questions
facing car makers and regulators.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170405/2691ea03/attachment.html>

From weiyu at cs.cmu.edu  Fri Apr  7 12:00:17 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Fri, 7 Apr 2017 12:00:17 -0400
Subject: [AI Seminar] AI Lunch -- Jiaji Zhou -- April 11
Message-ID: <CABzq7epyCMdGWmRt9r0+z++9Obb7eamM-Wb1YCHJ_Q+O1Y9EqQ@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you Next Tuesday, April 11, at noon in NSH 3305
for AI lunch. To learn more about the seminar and lunch, please visit the AI
 Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Jiaji Zhou <http://www.cs.cmu.edu/~jiajiz/> will give the
following talk:

Title: Exploiting Task Mechanics for Contact-rich Robotic Manipulation:
Mechanics Model Learning, Uncertainty Reduction and Control.

Abstract:

Effective robotic manipulation requires an understanding of the underlying
physical processes. In this talk, I will start with first principle guided
mechanics model learning for planar sliding. The model is further extended
for simulating generic quasi-static planar contact problems including
pushing and grasping. The second part of the talk is on tree-based grasping
policy synthesis for uncertainty reduction given encoder feedback. If time
permits, I will end the talk with some recent work on control synthesis and
trajectory planning for pushing.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170407/d89ebf5d/attachment.html>

From cga at cs.cmu.edu  Mon Apr 10 11:36:04 2017
From: cga at cs.cmu.edu (Chris Atkeson)
Date: Mon, 10 Apr 2017 11:36:04 -0400 (EDT)
Subject: [AI Seminar] Talk on Data Driven Dialog Management using
	reinforcement learning
Message-ID: <Pine.LNX.4.63.1704101135340.24838@SCORE.GRAPHICS.cs.cmu.edu>

Tuesday April 11, 3pm, NSH 3305
If you would like to meet with the speaker, please send free times on Tuesday 
to cga at cmu.edu. Alborz is here all day Tuesday. This is a good
way to find out about Amazon's Alexa project.

Title:
Data Driven Dialog Management

Speaker:
Alborz Geramifard
Manager, Machine Learning, Amazon
http://alborz-geramifard.com/

Abstract:
Speech-based AI assistants such as Alexa and Google Now are becoming 
increasingly popular as a convenient way for people to interact with machines. 
However, users find interactions with their assistants more natural if 
conducted in a conversational manner, with multiple requests made and responses 
provided in a given dialog session. Creating robust dialog policies for 
conversational bots is challenging. This talk presents a data driven approach 
for dialog management through reinforcement learning. We first introduce a 
framework for building conversational bots and describe MovieBot as an 
implementation of the framework that was launched as an Alexa skill. We then 
describe approaches to creating the reward function based on sentiment analysis 
on text, using various techniques including Long Short Term Memory networks 
(LSTMs). The talk will end by discussing potential directions, and how all 
pieces of the puzzle can fit together.

Bio:
Alborz Geramifard is currently a Machine Learning Manager at Amazon working on 
conversation AI for Alexa. Before joining Amazon, he was a postdoctoral 
associate at MIT's Laboratory for Information and Decision Systems. Alborz 
received his PhD from MIT working on representation learning and safe 
exploration in large scale sensitive sequential decision-making problems in 
2012. He finished his MSc at University of Alberta in 2008, where he worked on 
data efficient online reinforcement learning techniques. His research interests 
lie at machine learning with the focus on reinforcement learning, natural 
language understanding, planning, and brain and cognitive sciences. Alborz was 
the recipient of the NSERC postgraduate scholarships 2010-2012 program.

Host:
Chris Atkeson cga at cmu.edu

From weiyu at cs.cmu.edu  Mon Apr 10 12:20:58 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 10 Apr 2017 12:20:58 -0400
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 73, Issue 4
In-Reply-To: <mailman.117.1491667223.20755.ai-seminar-announce@cs.cmu.edu>
References: <mailman.117.1491667223.20755.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7eoZe0kiHJEp1nEb_ZpX43XENbcpkxELJOqAF6bcLaAFcA@mail.gmail.com>

A gentle reminder that the talk would be tomorrow (Tuesday).


On Sat, Apr 8, 2017 at 12:00 PM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Lunch -- Jiaji Zhou -- April 11 (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Fri, 7 Apr 2017 12:00:17 -0400
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Subject: [AI Seminar] AI Lunch -- Jiaji Zhou -- April 11
> Message-ID:
>         <CABzq7epyCMdGWmRt9r0+z++9Obb7eamM-Wb1YCHJ_Q+O1Y9EqQ@
> mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you Next Tuesday, April 11, at noon in NSH 3305
> for AI lunch. To learn more about the seminar and lunch, please visit the
> AI
>  Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Jiaji Zhou <http://www.cs.cmu.edu/~jiajiz/> will give the
> following talk:
>
> Title: Exploiting Task Mechanics for Contact-rich Robotic Manipulation:
> Mechanics Model Learning, Uncertainty Reduction and Control.
>
> Abstract:
>
> Effective robotic manipulation requires an understanding of the underlying
> physical processes. In this talk, I will start with first principle guided
> mechanics model learning for planar sliding. The model is further extended
> for simulating generic quasi-static planar contact problems including
> pushing and grasping. The second part of the talk is on tree-based grasping
> policy synthesis for uncertainty reduction given encoder feedback. If time
> permits, I will end the talk with some recent work on control synthesis and
> trajectory planning for pushing.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20170407/d89ebf5d/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 73, Issue 4
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170410/1bc2b310/attachment.html>

From weiyu at cs.cmu.edu  Fri Apr 14 14:12:50 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Fri, 14 Apr 2017 14:12:50 -0400
Subject: [AI Seminar] AI Lunch -- Iyad Rahwan (MIT) -- April 18
Message-ID: <CABzq7eozfLXp_=CJrZX1ieyiv19R2rTeE+tRZBfdJ0mhBsGDcw@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you Next Tuesday, April 18, at noon in NSH 3305
for AI lunch. To learn more about the seminar and lunch, please visit the AI
 Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Iyad Rahwan <http://www.mit.edu/~irahwan/> from MIT Media
Lab will give the following talk:


Title: The Psychological Dilemmas of Autonomous Vehicles.

Abstract:
Autonomous vehicles promise to revolutionize transportation and
substantially improve safety. But they also made salient an ethical
question: how should an algorithm decide relative risk in situations of
unavoidable harm? I present a series of psychological studies that explore
how people think about such scenarios, and discuss the regulatory
challenges these studies reveal. I also describe a crowdsourcing effort
that engaged millions of people worldwide with the ethical questions facing
car makers and regulators.

Bio:
Iyad Rahwan is the AT&T Career Development Professor and an Associate
Professor of Media Arts & Sciences at the MIT Media Lab, where he leads the
Scalable Cooperation group. He holds a PhD from the University of
Melbourne, Australia, and is an affiliate faculty at the MIT Institute of
Data, Systems and Society (IDSS). Rahwan's work lies at the intersection of
the computer and social sciences, with a focus on collective intelligence,
large-scale cooperation, and the social aspects of Artificial Intelligence.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170414/1c13c7d7/attachment.html>

From weiyu at cs.cmu.edu  Mon Apr 17 14:41:45 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 17 Apr 2017 14:41:45 -0400
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 73, Issue 6
In-Reply-To: <mailman.93.1492272019.11717.ai-seminar-announce@cs.cmu.edu>
References: <mailman.93.1492272019.11717.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7epWvo_-w0rzgJJ8nxAwQUTT570f0Znu1B-ofJ2dn0xpWw@mail.gmail.com>

A gentle reminder that the talk would be tomorrow(Tuesday).

On Sat, Apr 15, 2017 at 12:00 PM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Lunch -- Iyad Rahwan (MIT) -- April 18 (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Fri, 14 Apr 2017 14:12:50 -0400
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Subject: [AI Seminar] AI Lunch -- Iyad Rahwan (MIT) -- April 18
> Message-ID:
>         <CABzq7eozfLXp_=CJrZX1ieyiv19R2rTeE+tRZBfdJ0mhBsGDcw at mail.
> gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you Next Tuesday, April 18, at noon in NSH 3305
> for AI lunch. To learn more about the seminar and lunch, please visit the
> AI
>  Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Iyad Rahwan <http://www.mit.edu/~irahwan/> from MIT Media
> Lab will give the following talk:
>
>
> Title: The Psychological Dilemmas of Autonomous Vehicles.
>
> Abstract:
> Autonomous vehicles promise to revolutionize transportation and
> substantially improve safety. But they also made salient an ethical
> question: how should an algorithm decide relative risk in situations of
> unavoidable harm? I present a series of psychological studies that explore
> how people think about such scenarios, and discuss the regulatory
> challenges these studies reveal. I also describe a crowdsourcing effort
> that engaged millions of people worldwide with the ethical questions facing
> car makers and regulators.
>
> Bio:
> Iyad Rahwan is the AT&T Career Development Professor and an Associate
> Professor of Media Arts & Sciences at the MIT Media Lab, where he leads the
> Scalable Cooperation group. He holds a PhD from the University of
> Melbourne, Australia, and is an affiliate faculty at the MIT Institute of
> Data, Systems and Society (IDSS). Rahwan's work lies at the intersection of
> the computer and social sciences, with a focus on collective intelligence,
> large-scale cooperation, and the social aspects of Artificial Intelligence.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20170414/1c13c7d7/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 73, Issue 6
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170417/b88b8093/attachment.html>

From weiyu at cs.cmu.edu  Fri Apr 21 11:42:10 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Fri, 21 Apr 2017 11:42:10 -0400
Subject: [AI Seminar] AI Lunch -- Noam Brown -- April 25
Message-ID: <CABzq7erRgRKOF7r0XyoJYpjD_Lsw1PfUgyqLuii1kBxfD-LMGw@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you Next Tuesday, April 25, at noon in NSH 3305
for AI lunch. To learn more about the seminar and lunch, please visit the AI
 Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Noam Brown will give a talk about Libratus
<https://en.wikipedia.org/wiki/Libratus>:

Title: Libratus: Beating Top Humans in No-Limit Poker

Abstract:
Poker has been a challenge problem in AI and game theory for decades. As a
game of imperfect information, poker involves obstacles not present in
games like chess or go. No program has been able to beat top professionals
in large poker games, until now. In January 2017, our AI Libratus
decisively defeated a team of the top professional players in heads-up
no-limit Texas Hold'em. Libratus features a number of innovations which
form a new approach to AI for imperfect-information games. The algorithms
are domain-independent and are widely applicable to any strategic
interaction involving hidden information.

Bio:
Noam Brown is a PhD student in computer science at Carnegie Mellon
University advised by Professor Tuomas Sandholm. His research combines
reinforcement learning and game theory to develop AIs capable of strategic
reasoning in imperfect-information interactions. He has applied this
research to creating Libratus, the first AI to defeat top professional
poker players in no-limit Texas Hold'em.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170421/8f8c71d9/attachment.html>

From weiyu at cs.cmu.edu  Mon Apr 24 11:35:18 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 24 Apr 2017 11:35:18 -0400
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 73, Issue 8
In-Reply-To: <mailman.93.1492790419.17871.ai-seminar-announce@cs.cmu.edu>
References: <mailman.93.1492790419.17871.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7erqJ4P37G23qQ3xQfj4kM-xSq3Pgi-Nk9QCJ3x=WC3fRQ@mail.gmail.com>

This is a gentle reminder that the talk about Libratus will be tomorrow
(Tuesday) noon.

On Fri, Apr 21, 2017 at 12:00 PM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Lunch -- Noam Brown -- April 25 (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Fri, 21 Apr 2017 11:42:10 -0400
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Subject: [AI Seminar] AI Lunch -- Noam Brown -- April 25
> Message-ID:
>         <CABzq7erRgRKOF7r0XyoJYpjD_Lsw1PfUgyqLuii1kBxfD-LMGw@
> mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you Next Tuesday, April 25, at noon in NSH 3305
> for AI lunch. To learn more about the seminar and lunch, please visit the
> AI
>  Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Noam Brown will give a talk about Libratus
> <https://en.wikipedia.org/wiki/Libratus>:
>
> Title: Libratus: Beating Top Humans in No-Limit Poker
>
> Abstract:
> Poker has been a challenge problem in AI and game theory for decades. As a
> game of imperfect information, poker involves obstacles not present in
> games like chess or go. No program has been able to beat top professionals
> in large poker games, until now. In January 2017, our AI Libratus
> decisively defeated a team of the top professional players in heads-up
> no-limit Texas Hold'em. Libratus features a number of innovations which
> form a new approach to AI for imperfect-information games. The algorithms
> are domain-independent and are widely applicable to any strategic
> interaction involving hidden information.
>
> Bio:
> Noam Brown is a PhD student in computer science at Carnegie Mellon
> University advised by Professor Tuomas Sandholm. His research combines
> reinforcement learning and game theory to develop AIs capable of strategic
> reasoning in imperfect-information interactions. He has applied this
> research to creating Libratus, the first AI to defeat top professional
> poker players in no-limit Texas Hold'em.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20170421/8f8c71d9/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 73, Issue 8
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170424/91296eb7/attachment.html>

From weiyu at cs.cmu.edu  Fri Apr 28 18:16:40 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Fri, 28 Apr 2017 18:16:40 -0400
Subject: [AI Seminar] AI Lunch -- Travis Dick -- May 2
Message-ID: <CABzq7eoVfOcpdZUBJ=Ryeepx+FZFv0h+8vgs2PEkXoWJb0rjvw@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you Next Tuesday, May 2, at noon in NSH 3305 for
AI lunch. To learn more about the seminar and lunch, please visit the AI
Lunch webpage.

On Tuesday, Travis Dick <http://www.cs.cmu.edu/~tdick/> will give a talk :

Title: Data Driven Resource Allocation for Distributed Learning

Abstract:
The goal of distributed machine learning is to build useful models from
more data than can be processed by a single machine. In this talk I will
present a new data-dependent approach for partitioning large datasets onto
multiple machines motivated by the fact that similar data points often
belong to the same or similar classes, and more generally, classification
rules of high accuracy tend to be "locally simple but globally complex"
(Vapnik and Bottou, 1993). We present an in-depth analysis of our approach,
provide new algorithms with provable worst-case guarantees, analysis
proving existing scalable heuristics perform well in natural non worst-case
conditions, and techniques for extending the partitioning of a small sample
to the entire dataset. We overcome novel technical challenges to satisfy
important conditions for accurate distributed learning, including fault
tolerance and balancedness. We empirically compare our approach with
baselines based on random partitioning, balanced partition trees, and
locality sensitive hashing, showing that we achieve significantly higher
accuracy on both synthetic and real world image and advertising datasets.
We also demonstrate that our technique strongly scales with the available
computing power.

This is joint work with Mu Li, Krishna Pillutla, Colin White, Nina Balcan,
and Alex Smola. In Partial Fulfillment of the Speaking Requirement.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170428/199e2441/attachment.html>

From weiyu at cs.cmu.edu  Mon May  1 13:43:18 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 1 May 2017 13:43:18 -0400
Subject: [AI Seminar] AI Lunch -- Travis Dick -- May 2
In-Reply-To: <CABzq7eoVfOcpdZUBJ=Ryeepx+FZFv0h+8vgs2PEkXoWJb0rjvw@mail.gmail.com>
References: <CABzq7eoVfOcpdZUBJ=Ryeepx+FZFv0h+8vgs2PEkXoWJb0rjvw@mail.gmail.com>
Message-ID: <CABzq7eouzZPWVrPpfGZeK-GsxiruWQvFp6NKi9We_aDWDZ-39g@mail.gmail.com>

A gentle reminder that the talk will be tomorrow (Tuesday).

On Fri, Apr 28, 2017 at 6:16 PM, Adams Wei Yu <weiyu at cs.cmu.edu> wrote:

> Dear faculty and students,
>
> We look forward to seeing you Next Tuesday, May 2, at noon in NSH 3305 for
> AI lunch. To learn more about the seminar and lunch, please visit the AI
> Lunch webpage.
>
> On Tuesday, Travis Dick <http://www.cs.cmu.edu/~tdick/> will give a talk
> :
>
> Title: Data Driven Resource Allocation for Distributed Learning
>
> Abstract:
> The goal of distributed machine learning is to build useful models from
> more data than can be processed by a single machine. In this talk I will
> present a new data-dependent approach for partitioning large datasets onto
> multiple machines motivated by the fact that similar data points often
> belong to the same or similar classes, and more generally, classification
> rules of high accuracy tend to be "locally simple but globally complex"
> (Vapnik and Bottou, 1993). We present an in-depth analysis of our approach,
> provide new algorithms with provable worst-case guarantees, analysis
> proving existing scalable heuristics perform well in natural non worst-case
> conditions, and techniques for extending the partitioning of a small sample
> to the entire dataset. We overcome novel technical challenges to satisfy
> important conditions for accurate distributed learning, including fault
> tolerance and balancedness. We empirically compare our approach with
> baselines based on random partitioning, balanced partition trees, and
> locality sensitive hashing, showing that we achieve significantly higher
> accuracy on both synthetic and real world image and advertising datasets.
> We also demonstrate that our technique strongly scales with the available
> computing power.
>
> This is joint work with Mu Li, Krishna Pillutla, Colin White, Nina Balcan,
> and Alex Smola. In Partial Fulfillment of the Speaking Requirement.
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170501/69ec01ee/attachment.html>

From weiyu at cs.cmu.edu  Sat May  6 15:45:16 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Sat, 6 May 2017 15:45:16 -0400
Subject: [AI Seminar] AI Lunch -- Andrew Mao (MSR) -- May 9
Message-ID: <CABzq7ere0ypXVcFP3vQ8gV3eAd1k5PnaKkaaOGWr_Vq5znop_A@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you Next Tuesday, May 9, at noon in NSH 3305 for
AI lunch. To learn more about the seminar and lunch, please visit the AI
Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Andrew Mao <http://www.andrewmao.net/> will give a talk :

Title: Studying Teamwork and Cooperation in the Virtual Lab

Abstract:
For decades, physical behavioral labs have been a primary, yet limited,
method for controlled experimental studies of human behavior. Now,
software-based "virtual labs" on the Internet allow for studies of
increasing complexity, size, and scope. In this talk, I highlight the
potential of virtual lab experiments for studying social interaction and
coordination. First, I present a study of cooperation in a social dilemma
over a month of real time, using crowdsourcing participants to overcome the
time constraints of behavioral labs. Our study of about 100 participants
over 20 consecutive weekdays finds that a group of resilient altruists
sustain a high level of cooperation across the entire population. We also
explore collective intelligence and digital teamwork in "crisis mapping",
where digital volunteers organize to assess and pinpoint damage in the
aftermath of humanitarian crises. By simulating a crisis mapping scenario
to study self-organization in teams of varying size, and find a tradeoff
between individual effort in small groups and collective coordination in
larger teams. Together, our work motivates the potential of controlled,
highly instrumented studies of social interaction; the importance of
behavioral experiments on longer timescales; and how open-source software
both can speed up the iteration and improve the reproducibility of
experimental work.

This talk is based on joint work with Lili Dworkin, Winter Mason, Siddharth
Suri, and Duncan Watts.

Bio:
Andrew Mao is currently a postdoc at Microsoft Research in NYC, where his
research focuses on experimental studies of collective behavior using
Internet participants by combining approaches from social and computer
science. Andrew is especially interested in expanding the boundaries of
experimental methods, and his work has appeared in interdisciplinary
journals including Nature Communications and PLoS ONE as well as computer
science conferences such as AAAI, EC, and HCOMP. He has also designed
TurkServer (http://turkserver.readthedocs.io/), an open-source platform for
real-time, interactive, web-based behavioral experiments, to share these
methods with other researchers. He received his PhD from Harvard University
in 2015.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170506/2a266473/attachment.html>

From weiyu at cs.cmu.edu  Mon May  8 15:58:22 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 8 May 2017 15:58:22 -0400
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 74, Issue 2
In-Reply-To: <mailman.21.1494172811.19079.ai-seminar-announce@cs.cmu.edu>
References: <mailman.21.1494172811.19079.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7erwaG4xFZCp3jy93_8udrEbZPYOh9qMfqD-17JbHyqi5A@mail.gmail.com>

Just a reminder that the talk will be tomorrow (Tuesday) noon.

On Sun, May 7, 2017 at 12:00 PM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Lunch -- Andrew Mao (MSR) -- May 9 (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Sat, 6 May 2017 15:45:16 -0400
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Subject: [AI Seminar] AI Lunch -- Andrew Mao (MSR) -- May 9
> Message-ID:
>         <CABzq7ere0ypXVcFP3vQ8gV3eAd1k5PnaKkaaOGWr_Vq5znop_A at mail.
> gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you Next Tuesday, May 9, at noon in NSH 3305 for
> AI lunch. To learn more about the seminar and lunch, please visit the AI
> Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Andrew Mao <http://www.andrewmao.net/> will give a talk :
>
> Title: Studying Teamwork and Cooperation in the Virtual Lab
>
> Abstract:
> For decades, physical behavioral labs have been a primary, yet limited,
> method for controlled experimental studies of human behavior. Now,
> software-based "virtual labs" on the Internet allow for studies of
> increasing complexity, size, and scope. In this talk, I highlight the
> potential of virtual lab experiments for studying social interaction and
> coordination. First, I present a study of cooperation in a social dilemma
> over a month of real time, using crowdsourcing participants to overcome the
> time constraints of behavioral labs. Our study of about 100 participants
> over 20 consecutive weekdays finds that a group of resilient altruists
> sustain a high level of cooperation across the entire population. We also
> explore collective intelligence and digital teamwork in "crisis mapping",
> where digital volunteers organize to assess and pinpoint damage in the
> aftermath of humanitarian crises. By simulating a crisis mapping scenario
> to study self-organization in teams of varying size, and find a tradeoff
> between individual effort in small groups and collective coordination in
> larger teams. Together, our work motivates the potential of controlled,
> highly instrumented studies of social interaction; the importance of
> behavioral experiments on longer timescales; and how open-source software
> both can speed up the iteration and improve the reproducibility of
> experimental work.
>
> This talk is based on joint work with Lili Dworkin, Winter Mason, Siddharth
> Suri, and Duncan Watts.
>
> Bio:
> Andrew Mao is currently a postdoc at Microsoft Research in NYC, where his
> research focuses on experimental studies of collective behavior using
> Internet participants by combining approaches from social and computer
> science. Andrew is especially interested in expanding the boundaries of
> experimental methods, and his work has appeared in interdisciplinary
> journals including Nature Communications and PLoS ONE as well as computer
> science conferences such as AAAI, EC, and HCOMP. He has also designed
> TurkServer (http://turkserver.readthedocs.io/), an open-source platform
> for
> real-time, interactive, web-based behavioral experiments, to share these
> methods with other researchers. He received his PhD from Harvard University
> in 2015.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20170506/2a266473/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 74, Issue 2
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170508/c6fae6a5/attachment.html>

From weiyu at cs.cmu.edu  Fri May 12 21:40:44 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Fri, 12 May 2017 21:40:44 -0400
Subject: [AI Seminar] AI Lunch -- Ramanathan V. Guha (Schema.org) -- May 16
Message-ID: <CABzq7epc+DDP14q98cASSn-yU3Oof8k4yxiCLXLGQRw0iAjY0w@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you Next Tuesday, May 16, at noon in NSH 3305 for
AI lunch. To learn more about the seminar and lunch, please visit the AI
Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Ramanathan V. Guha <http://www.guha.com/cv.html> will give a
talk :

Title: Communicating Semantics

Abstract:
Messages often refer to entities such as people, places and events. Correct
identification of the intended reference is an essential part of
communication. Lack of shared unique names often complicates entity
reference. Humans do not and cannot have a shared unique names for
everything. Yet, we communicate in our daily lives about things that do not
have a unique name (like John McCarthy) or lack a name (like his first
car). Our long term goal is to enable programs to achieve communication
just as effectively.

Humans overcome the lack of shared unique names by using descriptions. We
use some shared language and shared domain knowledge to construct uniquely
identifying descriptions for such entities. We call this `Reference by
Description' and argue that it could form the basis for a programatic use
of descriptions.

We introduce a mathematical model of `Reference by Description'. We use an
information theoretic approach to address questions such as: What is the
minimum that needs to be shared for two communicating parties to understand
each other? When can we bootstrap from no shared language? What is the
computational cost of using descriptions instead of unique names?

In addition to their inherent interestingness, these questions also have
practical applications that range from database/application integration to
privacy preserving information sharing.

Bio:
Guha is the creator of widely used web standards such as RSS, RDF and
Schema.org and products such as Google Custom Search. He has made
substantial contributions to Netscape???s browser, Google???s Adwords,
Google Now, etc. He was a co-founder of Epinions.com and Alpiri. Until
recently, he was a Google Fellow. He has a Ph.D. in computer science from
Stanford University and B.Tech in mechanical engineering from IIT Chennai.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170512/2dad4abb/attachment.html>

From weiyu at cs.cmu.edu  Mon May 15 13:06:28 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 15 May 2017 13:06:28 -0400
Subject: [AI Seminar] AI Lunch -- Ramanathan V. Guha (Schema.org) -- May
	16
In-Reply-To: <CABzq7epc+DDP14q98cASSn-yU3Oof8k4yxiCLXLGQRw0iAjY0w@mail.gmail.com>
References: <CABzq7epc+DDP14q98cASSn-yU3Oof8k4yxiCLXLGQRw0iAjY0w@mail.gmail.com>
Message-ID: <CABzq7epAR-RRYCc4obiO29kK1ZiHPjwCos+FHMW8i3t6QWeJXQ@mail.gmail.com>

Just a reminder that the talk is tomorrow(Tuesday).

On Fri, May 12, 2017 at 9:40 PM, Adams Wei Yu <weiyu at cs.cmu.edu> wrote:

> Dear faculty and students,
>
> We look forward to seeing you Next Tuesday, May 16, at noon in NSH 3305
> for AI lunch. To learn more about the seminar and lunch, please visit the
> AI Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Ramanathan V. Guha <http://www.guha.com/cv.html> will give a
> talk :
>
> Title: Communicating Semantics
>
> Abstract:
> Messages often refer to entities such as people, places and events.
> Correct identification of the intended reference is an essential part of
> communication. Lack of shared unique names often complicates entity
> reference. Humans do not and cannot have a shared unique names for
> everything. Yet, we communicate in our daily lives about things that do not
> have a unique name (like John McCarthy) or lack a name (like his first
> car). Our long term goal is to enable programs to achieve communication
> just as effectively.
>
> Humans overcome the lack of shared unique names by using descriptions. We
> use some shared language and shared domain knowledge to construct uniquely
> identifying descriptions for such entities. We call this `Reference by
> Description' and argue that it could form the basis for a programatic use
> of descriptions.
>
> We introduce a mathematical model of `Reference by Description'. We use an
> information theoretic approach to address questions such as: What is the
> minimum that needs to be shared for two communicating parties to understand
> each other? When can we bootstrap from no shared language? What is the
> computational cost of using descriptions instead of unique names?
>
> In addition to their inherent interestingness, these questions also have
> practical applications that range from database/application integration to
> privacy preserving information sharing.
>
> Bio:
> Guha is the creator of widely used web standards such as RSS, RDF and
> Schema.org and products such as Google Custom Search. He has made
> substantial contributions to Netscape???s browser, Google???s Adwords,
> Google Now, etc. He was a co-founder of Epinions.com and Alpiri. Until
> recently, he was a Google Fellow. He has a Ph.D. in computer science from
> Stanford University and B.Tech in mechanical engineering from IIT Chennai.
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170515/051c7a87/attachment.html>

From weiyu at cs.cmu.edu  Mon May 22 11:45:56 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 22 May 2017 08:45:56 -0700
Subject: [AI Seminar] Call for Talks
Message-ID: <CABzq7erfGMiphyLTpb_=M_VY=QXQ_F4yyH-MN4mvr99B554Z+Q@mail.gmail.com>

Dear all,

There will be no more talks this semester. The AI lunch will resume at the
beginning of fall. Please sign up for talks if you want to give one!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170522/50517685/attachment.html>

From weiyu at cs.cmu.edu  Sun Aug 27 18:40:54 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Sun, 27 Aug 2017 15:40:54 -0700
Subject: [AI Seminar] AI Lunch -- Jason Hartline (Northwestern) -- August 29
Message-ID: <CABzq7ercL-HWv2KAi_t7O9rMeJYLzJisiAQhU1g1u0=2WOFs7Q@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you this Tuesday, August 29, at noon in NSH 1507
(subject to change) for AI lunch. To learn more about the seminar and
lunch, please visit the AI Lunch webpage <http://www.cs.cmu.edu/~aiseminar/>
.

On Tuesday, Jason Hartline <https://sites.northwestern.edu/hartline/> from
Northwestern will visit CMU. If you want to meet him, please sign up in
this spreadsheet
<https://docs.google.com/spreadsheets/d/1iAtFpEjGcXSip8EFbMvJi6LARtyKbjmhn0HAqh7aLgw/edit#gid=0>.
His talk is the following:

Title: Peer Grading and Mechanism Design

Abstract:

The first part of the talk will overview a peer grading system that is
under development at Northwestern U. In courses that use the system it has
(a) reduced the grading load of course staff by over 75%, (b) expanded and
improved the students??? interaction with the course material, (c) and
improved turn-around time of feedback on student work (students receive
comments on their work after three days, rather than two weeks). As a
research platform, this system enables a dialogue between theory and
practice for algorithms, machine learning, and mechanism design.

Of particular focus for the talk is a connection between incentivizing
students to give accurate peer reviews and all-pay position auctions.
All-pay auctions are a common model for competitions based on effort. In an
all-pay position auction, bidders compete for positions, where higher
positions are more valuable, and all bidders pay their bids (i.e., their
effort is spent). The equilibrium outcome of an all-pay auction depends on
the relative values of the positions. The second part of the talk will
describe recent work on reoptimizing all-pay position auctions from their
equilibrium bid data.

The second part of the talk is joint work with Shuchi Chawla and Denis
Nekipelov.

Bio: Prof. Hartline received his Ph.D. in 2003 from the University of
Washington under the supervision of Anna Karlin. He was a postdoctoral
fellow at Carnegie Mellon University under the supervision of Avrim Blum;
and subsequently a researcher at Microsoft Research in Silicon Valley. He
joined Northwestern University in 2008 where he is an associate professor
of computer science. He was on sabbatical at Harvard University in the
Economics Department during the 2014 calendar year and visiting Microsoft
Research, New England for the Spring of 2015.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170827/3676cf62/attachment.html>

From weiyu at cs.cmu.edu  Mon Aug 28 16:50:35 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 28 Aug 2017 13:50:35 -0700
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 75, Issue 1
In-Reply-To: <mailman.55.1503936017.18662.ai-seminar-announce@cs.cmu.edu>
References: <mailman.55.1503936017.18662.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7erj5e6R-76zqdzQjkF8o+UdEhohUQdUtPW19eHz2Qw08w@mail.gmail.com>

A gentle reminder that the talk will be tomorrow noon, in NSH 1507.

On Mon, Aug 28, 2017 at 9:00 AM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Lunch -- Jason Hartline (Northwestern) -- August 29
>       (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Sun, 27 Aug 2017 15:40:54 -0700
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Cc: hartline at eecs.northwestern.edu
> Subject: [AI Seminar] AI Lunch -- Jason Hartline (Northwestern) --
>         August 29
> Message-ID:
>         <CABzq7ercL-HWv2KAi_t7O9rMeJYLzJisiAQhU1g1u0=2WOFs
> 7Q at mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you this Tuesday, August 29, at noon in NSH 1507
> (subject to change) for AI lunch. To learn more about the seminar and
> lunch, please visit the AI Lunch webpage <http://www.cs.cmu.edu/~
> aiseminar/>
> .
>
> On Tuesday, Jason Hartline <https://sites.northwestern.edu/hartline/> from
> Northwestern will visit CMU. If you want to meet him, please sign up in
> this spreadsheet
> <https://docs.google.com/spreadsheets/d/1iAtFpEjGcXSip8EFbMvJi6LARtyKb
> jmhn0HAqh7aLgw/edit#gid=0>.
> His talk is the following:
>
> Title: Peer Grading and Mechanism Design
>
> Abstract:
>
> The first part of the talk will overview a peer grading system that is
> under development at Northwestern U. In courses that use the system it has
> (a) reduced the grading load of course staff by over 75%, (b) expanded and
> improved the students??? interaction with the course material, (c) and
> improved turn-around time of feedback on student work (students receive
> comments on their work after three days, rather than two weeks). As a
> research platform, this system enables a dialogue between theory and
> practice for algorithms, machine learning, and mechanism design.
>
> Of particular focus for the talk is a connection between incentivizing
> students to give accurate peer reviews and all-pay position auctions.
> All-pay auctions are a common model for competitions based on effort. In an
> all-pay position auction, bidders compete for positions, where higher
> positions are more valuable, and all bidders pay their bids (i.e., their
> effort is spent). The equilibrium outcome of an all-pay auction depends on
> the relative values of the positions. The second part of the talk will
> describe recent work on reoptimizing all-pay position auctions from their
> equilibrium bid data.
>
> The second part of the talk is joint work with Shuchi Chawla and Denis
> Nekipelov.
>
> Bio: Prof. Hartline received his Ph.D. in 2003 from the University of
> Washington under the supervision of Anna Karlin. He was a postdoctoral
> fellow at Carnegie Mellon University under the supervision of Avrim Blum;
> and subsequently a researcher at Microsoft Research in Silicon Valley. He
> joined Northwestern University in 2008 where he is an associate professor
> of computer science. He was on sabbatical at Harvard University in the
> Economics Department during the 2014 calendar year and visiting Microsoft
> Research, New England for the Spring of 2015.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20170827/3676cf62/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 75, Issue 1
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170828/82cffa84/attachment.html>

From weiyu at cs.cmu.edu  Thu Aug 31 02:43:11 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Wed, 30 Aug 2017 23:43:11 -0700
Subject: [AI Seminar] Call for Talks
Message-ID: <CABzq7erM9ftaZCjW5+H=unvHmLsrCTEqpVSdYG=r6KjaX9UsEQ@mail.gmail.com>

Dear all,

Welcome and Welcome Back to CMU! The tentative schedule for the AI seminar
this fall is here <http://www.cs.cmu.edu/~aiseminar/>. More than half of
the slots are already filled, but we are still looking for more talks.

If you are interested in presenting, please do not hesitate to contact
Adams (weiyu at cs.cmu.edu) and/or Zico (zkolter at cs.cmu.edu) for
the arrangement. You are also welcome to nominate internal and/or external
speakers to us!

Best!
Adams
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170830/57c3ee31/attachment.html>

From weiyu at cs.cmu.edu  Sun Sep  3 21:30:57 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Sun, 3 Sep 2017 18:30:57 -0700
Subject: [AI Seminar] No AI Seminar on Sep 5
Message-ID: <CABzq7eo4o+AkD49zZikvMZxrt0Pdb2mV6YCWppryK9x7Q4FaLw@mail.gmail.com>

There will be no AI Seminar on Sep 5. The next talk will be on Sep 12.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170903/7f5b18c4/attachment.html>

From weiyu at cs.cmu.edu  Sat Sep  9 08:09:53 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Sat, 9 Sep 2017 05:09:53 -0700
Subject: [AI Seminar] AI Seminar -- Fei Fang -- September 12
Message-ID: <CABzq7epr7eHGbPwHS9ftwoyTPtVmZmmxZxXosOKZg1gTkbcLog@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you next Tuesday, September 12, at noon in NSH
3305 for AI Seminar. To learn more about the seminar series, please visit
the AI Seminar webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Professor Fei Fang <https://feifang.info/> will give the
following talk:

Title: Data-Aware Game Theory and Mechanism Design for Security,
Sustainability, and Mobility

Abstract:

There is a rise of interest in developing artificial intelligence-based
tools to address societal challenges in security, sustainability, and
mobility domains, e.g., protecting critical infrastructure and cyber
network, protecting wildlife, fishery, and forest, and improving the
efficiency of ridesharing systems. Motivated by these challenges, we have
proposed game theory and machine learning based models and algorithms that
provide descriptive, predictive and prescriptive analysis for problems with
strategic interactions between intelligent agents (such as law enforcement
agencies and their adversaries, or the drivers and riders) and data
available to some of the agents. The algorithms have led to several
applications deployed in the field for protecting the Staten Island Ferry
in New York City, for poaching threat tracking and prediction in Africa,
and for tiger conservation in Southeast Asia.


Bio:

Fei Fang is an Assistant Professor at the Institute for Software Research
at Carnegie Mellon University. Before joining CMU, she was a Postdoctoral
Fellow at the Center for Research on Computation and Society (CRCS) at
Harvard University, advised by Prof. David Parkes and Prof. Barbara Grosz.
She received her Ph.D. from the Department of Computer Science at the
University of Southern California in June 2016, advised by Prof. Milind
Tambe. She received her bachelor degree from the Department of Electronic
Engineering, Tsinghua University in July 2011. Her research lies in the
field of artificial intelligence and multi-agent systems, focusing on
computational game theory and mechanism design with applications to
security, sustainability, and mobility domains. Her dissertation is
selected as the runner-up for IFAAMAS-16 Victor Lesser Distinguished
Dissertation Award. Her work has won the Innovative Application Award at
Innovative Applications of Artificial Intelligence (IAAI 16), the
Outstanding Paper Award in Computational Sustainability Track at the
International Joint Conferences on Artificial Intelligence (IJCAI 15). Her
work on "Protecting Moving Targets with Mobile Resources" has been deployed
by the US Coast Guard for protecting the Staten Island Ferry in New York
City since April 2013. Her work on designing patrol strategies to combat
illegal poaching has lead to the deployment of PAWS application in a
conservation area in Southeast Asia for protecting tigers.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170909/03cf8b27/attachment.html>

From weiyu at cs.cmu.edu  Mon Sep 11 07:06:14 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 11 Sep 2017 04:06:14 -0700
Subject: [AI Seminar] AI Seminar -- Fei Fang -- September 12
In-Reply-To: <CABzq7epr7eHGbPwHS9ftwoyTPtVmZmmxZxXosOKZg1gTkbcLog@mail.gmail.com>
References: <CABzq7epr7eHGbPwHS9ftwoyTPtVmZmmxZxXosOKZg1gTkbcLog@mail.gmail.com>
Message-ID: <CABzq7eo+ps04mJiVNL0_tAannekDNkwb64UebAyQN7vCDOFKZA@mail.gmail.com>

A gentle reminder that the talk will be tomorrow (Tuesday) noon.

On Sat, Sep 9, 2017 at 5:09 AM, Adams Wei Yu <weiyu at cs.cmu.edu> wrote:

> Dear faculty and students,
>
> We look forward to seeing you next Tuesday, September 12, at noon in NSH
> 3305 for AI Seminar. To learn more about the seminar series, please visit
> the AI Seminar webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Professor Fei Fang <https://feifang.info/> will give the
> following talk:
>
> Title: Data-Aware Game Theory and Mechanism Design for Security,
> Sustainability, and Mobility
>
> Abstract:
>
> There is a rise of interest in developing artificial intelligence-based
> tools to address societal challenges in security, sustainability, and
> mobility domains, e.g., protecting critical infrastructure and cyber
> network, protecting wildlife, fishery, and forest, and improving the
> efficiency of ridesharing systems. Motivated by these challenges, we have
> proposed game theory and machine learning based models and algorithms that
> provide descriptive, predictive and prescriptive analysis for problems with
> strategic interactions between intelligent agents (such as law enforcement
> agencies and their adversaries, or the drivers and riders) and data
> available to some of the agents. The algorithms have led to several
> applications deployed in the field for protecting the Staten Island Ferry
> in New York City, for poaching threat tracking and prediction in Africa,
> and for tiger conservation in Southeast Asia.
>
>
> Bio:
>
> Fei Fang is an Assistant Professor at the Institute for Software Research
> at Carnegie Mellon University. Before joining CMU, she was a Postdoctoral
> Fellow at the Center for Research on Computation and Society (CRCS) at
> Harvard University, advised by Prof. David Parkes and Prof. Barbara Grosz.
> She received her Ph.D. from the Department of Computer Science at the
> University of Southern California in June 2016, advised by Prof. Milind
> Tambe. She received her bachelor degree from the Department of Electronic
> Engineering, Tsinghua University in July 2011. Her research lies in the
> field of artificial intelligence and multi-agent systems, focusing on
> computational game theory and mechanism design with applications to
> security, sustainability, and mobility domains. Her dissertation is
> selected as the runner-up for IFAAMAS-16 Victor Lesser Distinguished
> Dissertation Award. Her work has won the Innovative Application Award at
> Innovative Applications of Artificial Intelligence (IAAI 16), the
> Outstanding Paper Award in Computational Sustainability Track at the
> International Joint Conferences on Artificial Intelligence (IJCAI 15). Her
> work on "Protecting Moving Targets with Mobile Resources" has been deployed
> by the US Coast Guard for protecting the Staten Island Ferry in New York
> City since April 2013. Her work on designing patrol strategies to combat
> illegal poaching has lead to the deployment of PAWS application in a
> conservation area in Southeast Asia for protecting tigers.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170911/ef6fd021/attachment.html>

From weiyu at cs.cmu.edu  Sat Sep 16 05:32:21 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Sat, 16 Sep 2017 02:32:21 -0700
Subject: [AI Seminar] AI Seminar sponsored by Apple -- Bhuwan Dhingra --
	September 19
Message-ID: <CABzq7eqJ28iuidzxr-4ZtZgLdqExDu1Azpjpxzpgc0+E6YVbbw@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you next Tuesday, September 19, at noon in NSH
3305 for AI Seminar sponsored by Apple. To learn more about the seminar
series, please visit the AI Seminar webpage
<http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Bhuwan Dhingra <http://www.cs.cmu.edu/~bdhingra/> will give the
following talk:

Title: Neural architectures for reading and reasoning over documents

Abstract: Reading and understanding natural language text is important for
AI applications which need to extract information from unstructured
sources. Models designed for this task must deal with complex linguistic
phenomena such as paraphrasing, co-reference, logical entailment, syntactic
and semantic dependencies, and so on. In this talk I will show how
architectural biases, motivated from such phenomena, can be built into
neural network models to boost machine reading performance.

The first half of the talk will focus on the Gated-Attention (GA) Reader
model for learning fine-grained alignments between natural language queries
and documents. The output of this model is a query-focused representation
of the tokens in the document, which is used to extract the answer to the
query. The second half of the talk will focus on extensions which utilize
prior knowledge in the form of linguistic annotations to model long term
dependencies in the document. Modeling long-term dependencies is the first
step towards the more ambitious goal of reasoning over distinct parts of a
document. Finally, I will discuss some of the key directions for future
research, in terms of both improving the models and utilizing them for
concrete applications.

This is joint work with Zhilin Yang, Hanxiao Liu, Russ Salakhutdinov and
William Cohen.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170916/58b70271/attachment.html>

From adamsyuwei at gmail.com  Mon Sep 18 07:10:04 2017
From: adamsyuwei at gmail.com (Adams Wei Yu)
Date: Mon, 18 Sep 2017 04:10:04 -0700
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 76, Issue 4
In-Reply-To: <mailman.71.1505577617.12381.ai-seminar-announce@cs.cmu.edu>
References: <mailman.71.1505577617.12381.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7eqK0ZWoE5C9MXgszX3jyosp81-EYdP81PUtAvZ81S4iSw@mail.gmail.com>

A gentle reminder that the talk would be tomorrow (Tuesday) noon.

On Sat, Sep 16, 2017 at 9:00 AM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Seminar sponsored by Apple -- Bhuwan Dhingra --       September
>       19 (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Sat, 16 Sep 2017 02:32:21 -0700
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Cc: Bhuwan Dhingra <bdhingra at andrew.cmu.edu>
> Subject: [AI Seminar] AI Seminar sponsored by Apple -- Bhuwan Dhingra
>         --      September 19
> Message-ID:
>         <CABzq7eqJ28iuidzxr-4ZtZgLdqExDu1Azpjpxzpgc0+
> E6YVbbw at mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you next Tuesday, September 19, at noon in NSH
> 3305 for AI Seminar sponsored by Apple. To learn more about the seminar
> series, please visit the AI Seminar webpage
> <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Bhuwan Dhingra <http://www.cs.cmu.edu/~bdhingra/> will give
> the
> following talk:
>
> Title: Neural architectures for reading and reasoning over documents
>
> Abstract: Reading and understanding natural language text is important for
> AI applications which need to extract information from unstructured
> sources. Models designed for this task must deal with complex linguistic
> phenomena such as paraphrasing, co-reference, logical entailment, syntactic
> and semantic dependencies, and so on. In this talk I will show how
> architectural biases, motivated from such phenomena, can be built into
> neural network models to boost machine reading performance.
>
> The first half of the talk will focus on the Gated-Attention (GA) Reader
> model for learning fine-grained alignments between natural language queries
> and documents. The output of this model is a query-focused representation
> of the tokens in the document, which is used to extract the answer to the
> query. The second half of the talk will focus on extensions which utilize
> prior knowledge in the form of linguistic annotations to model long term
> dependencies in the document. Modeling long-term dependencies is the first
> step towards the more ambitious goal of reasoning over distinct parts of a
> document. Finally, I will discuss some of the key directions for future
> research, in terms of both improving the models and utilizing them for
> concrete applications.
>
> This is joint work with Zhilin Yang, Hanxiao Liu, Russ Salakhutdinov and
> William Cohen.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20170916/58b70271/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 76, Issue 4
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170918/8906f766/attachment.html>

From weiyu at cs.cmu.edu  Sun Sep 24 05:12:51 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Sun, 24 Sep 2017 02:12:51 -0700
Subject: [AI Seminar] AI Seminar sponsored by Apple -- Jianbo Ye --
	September 26
Message-ID: <CABzq7epRBO=7bbZgQa+87Gh26z63BYbXE9HjjC30cOrxSSDpOw@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you next Tuesday, September 26, at noon in NSH
1507 (Unusual place) for AI Seminar sponsored by Apple. To learn more about
the seminar series, please visit the AI Seminar webpage
<http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Jianbo Ye <http://personal.psu.edu/jxy198/> from PSU will give
the following talk:

Title: Optimal Transport for Machine Learning: The State-of-the-art
Numerical Tools

Abstract: Representation of datasets, classification and measurement of
similarities/disparities between complex data or objects such as images or
collection of histograms are ubiquitous problems in machine learning.
Optimal transport based distances are used more and more frequently to
address these questions. Despite its attractiveness, the calculations
related to OT are quite non-trivial, posing great computational challenges
to machine learning practitioners. In this talk, I will cover three major
approaches including entropic regularization, Bregman ADMM and Gibbs
sampling for approximately solving OT and variational Wasserstein problems
in machine learning. Part of the talk is based on my joint work with Prof.
James Z. Wang and Prof. Jia Li.

Bio: Jianbo Ye is now a Ph.D. candidate at College of Information Science
and Technology, The Pennsylvania State University. He works on machine
learning, optimization methods and computational statistics with an
emphasis on their connections to real-world. His thesis has been focused on
developing scalable and robust numerical algorithms that apply optimal
transport theory and Wasserstein geometry to machine learning models. He
received the B.Sc. degree in Mathematics from University of Science and
Technology of China (USTC). He has worked as a research intern at Intel
(2013) and Adobe (2017).
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170924/58b15120/attachment.html>

From weiyu at cs.cmu.edu  Mon Sep 25 07:17:54 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 25 Sep 2017 04:17:54 -0700
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 76, Issue 6
In-Reply-To: <mailman.31.1506268816.10599.ai-seminar-announce@cs.cmu.edu>
References: <mailman.31.1506268816.10599.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7eqepcCyZJSFKp7HhjcqeisxFYWfTgO==ppTGiAsz0CfTQ@mail.gmail.com>

A gentle reminder that the talk will happen tomorrow (Tuesday) noon in NSH
1507.

On Sun, Sep 24, 2017 at 9:00 AM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Seminar sponsored by Apple -- Jianbo Ye --    September 26
>       (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Sun, 24 Sep 2017 02:12:51 -0700
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Cc: jxy198 at ist.psu.edu
> Subject: [AI Seminar] AI Seminar sponsored by Apple -- Jianbo Ye --
>         September 26
> Message-ID:
>         <CABzq7epRBO=7bbZgQa+87Gh26z63BYbXE9HjjC30cOrxSSDpO
> w at mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you next Tuesday, September 26, at noon in NSH
> 1507 (Unusual place) for AI Seminar sponsored by Apple. To learn more about
> the seminar series, please visit the AI Seminar webpage
> <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Jianbo Ye <http://personal.psu.edu/jxy198/> from PSU will give
> the following talk:
>
> Title: Optimal Transport for Machine Learning: The State-of-the-art
> Numerical Tools
>
> Abstract: Representation of datasets, classification and measurement of
> similarities/disparities between complex data or objects such as images or
> collection of histograms are ubiquitous problems in machine learning.
> Optimal transport based distances are used more and more frequently to
> address these questions. Despite its attractiveness, the calculations
> related to OT are quite non-trivial, posing great computational challenges
> to machine learning practitioners. In this talk, I will cover three major
> approaches including entropic regularization, Bregman ADMM and Gibbs
> sampling for approximately solving OT and variational Wasserstein problems
> in machine learning. Part of the talk is based on my joint work with Prof.
> James Z. Wang and Prof. Jia Li.
>
> Bio: Jianbo Ye is now a Ph.D. candidate at College of Information Science
> and Technology, The Pennsylvania State University. He works on machine
> learning, optimization methods and computational statistics with an
> emphasis on their connections to real-world. His thesis has been focused on
> developing scalable and robust numerical algorithms that apply optimal
> transport theory and Wasserstein geometry to machine learning models. He
> received the B.Sc. degree in Mathematics from University of Science and
> Technology of China (USTC). He has worked as a research intern at Intel
> (2013) and Adobe (2017).
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20170924/58b15120/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 76, Issue 6
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170925/a066582d/attachment.html>

From weiyu at cs.cmu.edu  Sat Sep 30 06:07:09 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Sat, 30 Sep 2017 03:07:09 -0700
Subject: [AI Seminar] AI Seminar sponsored by Apple -- Nihar B. Shah --
	October 03
Message-ID: <CABzq7eqU0OA2Ekk+qf7CmzHmvgZ2mOPSMzPNGzFuinre2=RxRA@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you next Tuesday, October 03, at noon in NSH 1507
(unusual place) for AI Seminar sponsored by Apple. To learn more about the
seminar series, please visit the AI Seminar webpage
<http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Professor Nihar B. Shah <http://www.cs.cmu.edu/~nihars/> will
give the following talk:

Title: Learning from People

Abstract: Learning from people represents a new and expanding frontier for
data science. Two critical challenges in this domain are of developing
algorithms for robust learning and designing incentive mechanisms for
eliciting high-quality data. In this talk, I describe progress on these
challenges in the context of two canonical settings, namely those of
ranking and classification. In addressing the first challenge, I introduce
a class of "permutation-based" models that are considerably richer than
classical models, and present algorithms for estimation that are both
rate-optimal and significantly more robust than prior state-of-the-art
methods. I also discuss how these estimators automatically adapt and are
simultaneously also rate-optimal over the classical models, thereby
enjoying a surprising a win-win in the bias-variance tradeoff. As for the
second challenge, I present a class of "multiplicative" incentive
mechanisms, and show that they are the unique mechanisms that can guarantee
honest responses. Extensive experiments on a popular crowdsourcing platform
reveal that the theoretical guarantees of robustness and efficiency indeed
translate to practice, yielding several-fold improvements over prior art.

Bio: Nihar B. Shah is an Assistant Professor in the Machine Learning and
Computer Science departments at CMU. He is a recipient of the 2017 David J.
Sakrison memorial prize from EECS Berkeley for a "truly outstanding and
innovative PhD thesis", the Microsoft Research PhD Fellowship 2014-16, the
Berkeley Fellowship 2011-13, the IEEE Data Storage Best Paper and Best
Student Paper Awards for the years 2011/2012, and the SVC Aiya Medal 2010.
His research interests include statistics, machine learning, and game
theory, with a current focus on applications to learning from people.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20170930/9e1d9b90/attachment.html>

From weiyu at cs.cmu.edu  Mon Oct  2 07:36:20 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 2 Oct 2017 04:36:20 -0700
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 76, Issue 8
In-Reply-To: <mailman.75.1506787221.5629.ai-seminar-announce@cs.cmu.edu>
References: <mailman.75.1506787221.5629.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7eqj=Pyz3-1nXyiazYifXkqDqa+BfeDg_rsgMbT5rNmVKw@mail.gmail.com>

A gentle reminder that the talk will happen tomorrow (Tuesday) noon in NSH
1507.

On Sat, Sep 30, 2017 at 9:00 AM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Seminar sponsored by Apple -- Nihar B. Shah --        October 03
>       (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Sat, 30 Sep 2017 03:07:09 -0700
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Subject: [AI Seminar] AI Seminar sponsored by Apple -- Nihar B. Shah
>         --      October 03
> Message-ID:
>         <CABzq7eqU0OA2Ekk+qf7CmzHmvgZ2mOPSMzPNGzFuinre2=
> RxRA at mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you next Tuesday, October 03, at noon in NSH 1507
> (unusual place) for AI Seminar sponsored by Apple. To learn more about the
> seminar series, please visit the AI Seminar webpage
> <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Professor Nihar B. Shah <http://www.cs.cmu.edu/~nihars/> will
> give the following talk:
>
> Title: Learning from People
>
> Abstract: Learning from people represents a new and expanding frontier for
> data science. Two critical challenges in this domain are of developing
> algorithms for robust learning and designing incentive mechanisms for
> eliciting high-quality data. In this talk, I describe progress on these
> challenges in the context of two canonical settings, namely those of
> ranking and classification. In addressing the first challenge, I introduce
> a class of "permutation-based" models that are considerably richer than
> classical models, and present algorithms for estimation that are both
> rate-optimal and significantly more robust than prior state-of-the-art
> methods. I also discuss how these estimators automatically adapt and are
> simultaneously also rate-optimal over the classical models, thereby
> enjoying a surprising a win-win in the bias-variance tradeoff. As for the
> second challenge, I present a class of "multiplicative" incentive
> mechanisms, and show that they are the unique mechanisms that can guarantee
> honest responses. Extensive experiments on a popular crowdsourcing platform
> reveal that the theoretical guarantees of robustness and efficiency indeed
> translate to practice, yielding several-fold improvements over prior art.
>
> Bio: Nihar B. Shah is an Assistant Professor in the Machine Learning and
> Computer Science departments at CMU. He is a recipient of the 2017 David J.
> Sakrison memorial prize from EECS Berkeley for a "truly outstanding and
> innovative PhD thesis", the Microsoft Research PhD Fellowship 2014-16, the
> Berkeley Fellowship 2011-13, the IEEE Data Storage Best Paper and Best
> Student Paper Awards for the years 2011/2012, and the SVC Aiya Medal 2010.
> His research interests include statistics, machine learning, and game
> theory, with a current focus on applications to learning from people.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20170930/9e1d9b90/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 76, Issue 8
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171002/de91e5f9/attachment.html>

From weiyu at cs.cmu.edu  Sat Oct  7 19:18:31 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Sat, 7 Oct 2017 16:18:31 -0700
Subject: [AI Seminar] AI Seminar sponsored by Apple -- Chun-Liang Li --
	October 10
Message-ID: <CABzq7erc-_U2G4PyYaLdB0QaZsgyJBP2r8BaDKjPzRvGiKYZYw@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you next Tuesday, October 10, at noon in NSH 1507
(unusual place) for AI Seminar sponsored by Apple. To learn more about the
seminar series, please visit the AI Seminar webpage
<http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Chun-Liang Li <http://www.cs.cmu.edu/~chunlial/> will give the
following talk:

Title: MMD GAN: Towards Deeper Understanding of Moment Matching Network

Abstract:

Generative moment matching network (GMMN) is a deep generative model that
differs from Generative Adversarial Network (GAN) by replacing the
discriminator in GAN with a two-sample test based on kernel maximum mean
discrepancy (MMD). Although some theoretical guarantees of MMD have been
studied, the empirical performance of GMMN is still not as competitive as
that of GAN on challenging and large benchmark datasets. The computational
efficiency of GMMN is also less desirable in comparison with GAN, partially
due to its requirement for a rather large batch size during the training.
In this paper, we propose to improve both the model expressiveness of GMMN
and its computational efficiency by introducing adversarial kernel learning
techniques, as the replacement of a fixed Gaussian kernel in the original
GMMN. The new approach combines the key ideas in both GMMN and GAN, hence
we name it MMD-GAN. The new distance measure in MMD-GAN is a meaningful
loss that enjoys the advantage of weak topology and can be optimized via
gradient descent with relatively small batch sizes. In our evaluation on
multiple benchmark datasets, including MNIST, CIFAR- 10, CelebA and LSUN,
the performance of MMD-GAN significantly outperforms GMMN, and is
competitive with other representative GAN works.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171007/11fd61d9/attachment.html>

From weiyu at cs.cmu.edu  Mon Oct  9 08:08:46 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 9 Oct 2017 05:08:46 -0700
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 77, Issue 2
In-Reply-To: <mailman.31.1507478414.31422.ai-seminar-announce@cs.cmu.edu>
References: <mailman.31.1507478414.31422.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7eqA7m60-d09KyZjW5Q3rTnc-b8M3bRZwZoTpuZnQMK4Hg@mail.gmail.com>

A gentle reminder that the talk will happen tomorrow (Tuesday) noon at NSH
1507.

On Sun, Oct 8, 2017 at 9:00 AM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Seminar sponsored by Apple -- Chun-Liang Li --        October 10
>       (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Sat, 7 Oct 2017 16:18:31 -0700
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Subject: [AI Seminar] AI Seminar sponsored by Apple -- Chun-Liang Li
>         --      October 10
> Message-ID:
>         <CABzq7erc-_U2G4PyYaLdB0QaZsgyJBP2r8BaDKjP
> zRvGiKYZYw at mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you next Tuesday, October 10, at noon in NSH 1507
> (unusual place) for AI Seminar sponsored by Apple. To learn more about the
> seminar series, please visit the AI Seminar webpage
> <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Chun-Liang Li <http://www.cs.cmu.edu/~chunlial/> will give the
> following talk:
>
> Title: MMD GAN: Towards Deeper Understanding of Moment Matching Network
>
> Abstract:
>
> Generative moment matching network (GMMN) is a deep generative model that
> differs from Generative Adversarial Network (GAN) by replacing the
> discriminator in GAN with a two-sample test based on kernel maximum mean
> discrepancy (MMD). Although some theoretical guarantees of MMD have been
> studied, the empirical performance of GMMN is still not as competitive as
> that of GAN on challenging and large benchmark datasets. The computational
> efficiency of GMMN is also less desirable in comparison with GAN, partially
> due to its requirement for a rather large batch size during the training.
> In this paper, we propose to improve both the model expressiveness of GMMN
> and its computational efficiency by introducing adversarial kernel learning
> techniques, as the replacement of a fixed Gaussian kernel in the original
> GMMN. The new approach combines the key ideas in both GMMN and GAN, hence
> we name it MMD-GAN. The new distance measure in MMD-GAN is a meaningful
> loss that enjoys the advantage of weak topology and can be optimized via
> gradient descent with relatively small batch sizes. In our evaluation on
> multiple benchmark datasets, including MNIST, CIFAR- 10, CelebA and LSUN,
> the performance of MMD-GAN significantly outperforms GMMN, and is
> competitive with other representative GAN works.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20171007/11fd61d9/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 77, Issue 2
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171009/677a882e/attachment.html>

From weiyu at cs.cmu.edu  Mon Oct  9 23:20:40 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 9 Oct 2017 20:20:40 -0700
Subject: [AI Seminar] Fwd: This Week at Theory Lunch: Ellen Vitercik
In-Reply-To: <268FB28A-E9DA-4F20-93BC-F3879116DAEB@cs.cmu.edu>
References: <268FB28A-E9DA-4F20-93BC-F3879116DAEB@cs.cmu.edu>
Message-ID: <CABzq7erK7SVvR7s49zt_JchYveUufsHBLGXVSNoTWX-3kq+sKA@mail.gmail.com>

FYI, the theory lunch this week (Wednesday) might be of interest to you.
Please see the message below.


---------- Forwarded message ----------
From: Nicolas Resch <nresch at cs.cmu.edu>
Date: Mon, Oct 9, 2017 at 8:50 AM
Subject: This Week at Theory Lunch: Ellen Vitercik
To: theory-announce at cs.cmu.edu


Hi all,

Please join us Wednesday at noon in GHC 6115 for a talk by Ellen Vitercik,
where lunch will be provided. A video recording of the talk will be
available on the CMU Youtube Theory channel
<https://www.youtube.com/channel/UCWFp4UWNiOv71j0sPbdNiqw>.

*Location/Time:* Wednesday, October 11, 12-1pm
*Speaker: *Ellen Vitercik
*Title:* Differentially private algorithm and auction configuration
*Abstract: *

Algorithm configuration is an important aspect of modern data science and
algorithm design. Algorithms regularly depend on tunable parameters which
have a substantial effect on run-time and solution quality. Researchers
have developed machine learning techniques for automated parameter tuning
which use a training set of problem instances to determine a configuration
with high expected performance over future instances. This line of work has
inspired breakthroughs in diverse fields including combinatorial auctions,
scientific computing, vehicle routing, and SAT. The resulting configuration
depends on the training set, implying that without proper precautions, it
may leak sensitive information about problem instances contained therein.

In this work, we address this problem by showing that a natural adaptation
of the exponential mechanism provides strong privacy and utility guarantees
for a wide range of algorithm configuration problems. We uncover a
structural property shared among many problems which ensures this
algorithm?s success: these problems reduce to maximizing piecewise
L-Lipschitz functions. We call upon techniques from high-dimensional
geometry to efficiently implement this algorithm when the algorithm
parameter space is multi-dimensional. We apply our differentially private
algorithm to many configuration problems where privacy preservation is
essential, such as integer quadratic programming and greedy algorithm
configuration. We also show that our differentially private algorithm can
be used to design auctions and pricing mechanisms based on consumer data,
thus further exhibiting the algorithm?s broad applicability.

See you there!
Ellis and Nic
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171009/aab29fde/attachment.html>

From weiyu at cs.cmu.edu  Sat Oct 14 08:32:34 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Sat, 14 Oct 2017 05:32:34 -0700
Subject: [AI Seminar] AI Seminar sponsored by Apple -- Xiaolong Wang --
	October 17
Message-ID: <CABzq7epRGZ8vK1qBjUomXKynJDyD=WS2caQeNeaYLEiDBrB86w@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you next Tuesday, October 17, at noon in NSH 3305
for AI Seminar sponsored by Apple. To learn more about the seminar series,
please visit the AI Seminar webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Xiaolong Wang <http://www.cs.cmu.edu/~xiaolonw/> will give the
following talk:

Title:  Learning Visual Representations for Object Detection

Abstract:

Object detection is in the center of applications in computer vision. The
current pipeline for training object detectors include ConvNet pre-training
and fine-tuning. In this talk, I am going to cover our works on
self-supervised/unsupervised ConvNet pre-training as well as optimization
strategies on fine-tuning.

For ConvNet pre-training, instead of using millions of labeled images, we
explored to learn visual representations using supervisions from the data
itself without any human labels, i.e., self-supervised learning.
Specifically, we proposed to exploit different self-supervised approaches
to learn representations invariant to (i) inter-instance variations (two
objects in the same class should have similar features) and (ii)
intra-instance variations (viewpoint, pose, deformations, illumination).
Instead of combining two approaches with multi-task learning, we organized
the data with multiple variations in a graph and applied simple transitive
rules to generate pairs of images with richer visual invariance for
training. This approach brings the object detection accuracies on MSCOCO
dataset less than 1% away from methods using large amount of labeled data
(e.g., ImageNet).

For object detection fine-tuning, we proposed to train object detectors
invariant to occlusions and deformations. The common solution is to use a
data-driven strategy -- collect large-scale datasets which have object
instances under different conditions. However, like categories, occlusions
and object deformations also follow a long-tail. Some occlusions and
deformations are so rare that they hardly happen; yet we want to learn a
model invariant to such occurrences. In this talk, we propose to learn an
adversarial network that generates examples with occlusions and
deformations. The goal of the adversary is to generate examples that are
difficult for the object detector to classify. In our framework both the
original detector and adversary are learned in a joint manner. We show
significant improvements on different datasets (VOC, COCO) with different
network architectures (AlexNet, VGG16, ResNet101).
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171014/763f531f/attachment.html>

From weiyu at cs.cmu.edu  Mon Oct 16 08:05:15 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 16 Oct 2017 05:05:15 -0700
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 77, Issue 4
In-Reply-To: <mailman.57.1507996819.18614.ai-seminar-announce@cs.cmu.edu>
References: <mailman.57.1507996819.18614.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7epSi=ZZoVw1neLi6EJzp9D-Pf2=gEf3i82f1y3_8q8XBg@mail.gmail.com>

A gentle reminder that the talk will happen tomorrow (Tuesday) noon at NSH
3305.

On Sat, Oct 14, 2017 at 9:00 AM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Seminar sponsored by Apple -- Xiaolong Wang --        October 17
>       (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Sat, 14 Oct 2017 05:32:34 -0700
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Subject: [AI Seminar] AI Seminar sponsored by Apple -- Xiaolong Wang
>         --      October 17
> Message-ID:
>         <CABzq7epRGZ8vK1qBjUomXKynJDyD=WS2caQeNeaYLEiDBrB86w at mail.
> gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you next Tuesday, October 17, at noon in NSH 3305
> for AI Seminar sponsored by Apple. To learn more about the seminar series,
> please visit the AI Seminar webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Xiaolong Wang <http://www.cs.cmu.edu/~xiaolonw/> will give the
> following talk:
>
> Title:  Learning Visual Representations for Object Detection
>
> Abstract:
>
> Object detection is in the center of applications in computer vision. The
> current pipeline for training object detectors include ConvNet pre-training
> and fine-tuning. In this talk, I am going to cover our works on
> self-supervised/unsupervised ConvNet pre-training as well as optimization
> strategies on fine-tuning.
>
> For ConvNet pre-training, instead of using millions of labeled images, we
> explored to learn visual representations using supervisions from the data
> itself without any human labels, i.e., self-supervised learning.
> Specifically, we proposed to exploit different self-supervised approaches
> to learn representations invariant to (i) inter-instance variations (two
> objects in the same class should have similar features) and (ii)
> intra-instance variations (viewpoint, pose, deformations, illumination).
> Instead of combining two approaches with multi-task learning, we organized
> the data with multiple variations in a graph and applied simple transitive
> rules to generate pairs of images with richer visual invariance for
> training. This approach brings the object detection accuracies on MSCOCO
> dataset less than 1% away from methods using large amount of labeled data
> (e.g., ImageNet).
>
> For object detection fine-tuning, we proposed to train object detectors
> invariant to occlusions and deformations. The common solution is to use a
> data-driven strategy -- collect large-scale datasets which have object
> instances under different conditions. However, like categories, occlusions
> and object deformations also follow a long-tail. Some occlusions and
> deformations are so rare that they hardly happen; yet we want to learn a
> model invariant to such occurrences. In this talk, we propose to learn an
> adversarial network that generates examples with occlusions and
> deformations. The goal of the adversary is to generate examples that are
> difficult for the object detector to classify. In our framework both the
> original detector and adversary are learned in a joint manner. We show
> significant improvements on different datasets (VOC, COCO) with different
> network architectures (AlexNet, VGG16, ResNet101).
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20171014/763f531f/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 77, Issue 4
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171016/dd9578f1/attachment.html>

From weiyu at cs.cmu.edu  Sat Oct 21 21:51:49 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Sat, 21 Oct 2017 18:51:49 -0700
Subject: [AI Seminar] AI Seminar sponsored by Apple -- Zhiting Hu -- October
	24
Message-ID: <CABzq7eq8jD_JaouoBdkBxEz3qfoYJ-vO1HAJJO6S2cJ+Xmfnww@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you next Tuesday, October 24, at noon in NSH 1507
(unusual place) for AI Seminar sponsored by Apple. To learn more about the
seminar series, please visit the AI Seminar webpage
<http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Zhiting Hu <http://www.cs.cmu.edu/~zhitingh/> will give the
following talk:

Title: On Unifying Deep Generative Models

Abstract:

Deep generative models have achieved impressive success in recent years.
Generative Adversarial Networks (GANs) and Variational Autoencoders (VAEs),
as powerful frameworks for deep generative model learning, have largely
been considered as two distinct paradigms and received extensive
independent study respectively. This paper establishes formal connections
between deep generative modeling approaches through a new formulation of
GANs and VAEs. We show that GANs and VAEs involve minimizing KL divergences
of respective posterior and inference distributions with opposite
directions, extending the two learning phases of classic wake-sleep
algorithm, respectively. The unified view provides a powerful tool to
analyze a diverse set of existing model variants, and enables to exchange
ideas across research lines in a principled way. For example, we transfer
the importance weighting method in VAE literatures for improved GAN
learning, and enhance VAEs with an adversarial mechanism for leveraging
generated samples. Quantitative experiments show generality and
effectiveness of the imported extensions.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171021/3c105d6c/attachment.html>

From weiyu at cs.cmu.edu  Mon Oct 23 08:15:23 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 23 Oct 2017 05:15:23 -0700
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 77, Issue 6
In-Reply-To: <mailman.65.1508688020.5021.ai-seminar-announce@cs.cmu.edu>
References: <mailman.65.1508688020.5021.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7ep7fm43nwUMsxC+AMwzLaw=TVMzAAPmAY--zgOFpDowMw@mail.gmail.com>

A gentle reminder that the talk will happen tomorrow (Tuesday) noon in NSH
1507.

On Sun, Oct 22, 2017 at 9:00 AM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Seminar sponsored by Apple -- Zhiting Hu -- October   24
>       (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Sat, 21 Oct 2017 18:51:49 -0700
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Cc: Zhiting Hu <zhitinghu at gmail.com>
> Subject: [AI Seminar] AI Seminar sponsored by Apple -- Zhiting Hu --
>         October 24
> Message-ID:
>         <CABzq7eq8jD_JaouoBdkBxEz3qfoYJ-vO1HAJJO6S2cJ+Xmfnww at mail.
> gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you next Tuesday, October 24, at noon in NSH 1507
> (unusual place) for AI Seminar sponsored by Apple. To learn more about the
> seminar series, please visit the AI Seminar webpage
> <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Zhiting Hu <http://www.cs.cmu.edu/~zhitingh/> will give the
> following talk:
>
> Title: On Unifying Deep Generative Models
>
> Abstract:
>
> Deep generative models have achieved impressive success in recent years.
> Generative Adversarial Networks (GANs) and Variational Autoencoders (VAEs),
> as powerful frameworks for deep generative model learning, have largely
> been considered as two distinct paradigms and received extensive
> independent study respectively. This paper establishes formal connections
> between deep generative modeling approaches through a new formulation of
> GANs and VAEs. We show that GANs and VAEs involve minimizing KL divergences
> of respective posterior and inference distributions with opposite
> directions, extending the two learning phases of classic wake-sleep
> algorithm, respectively. The unified view provides a powerful tool to
> analyze a diverse set of existing model variants, and enables to exchange
> ideas across research lines in a principled way. For example, we transfer
> the importance weighting method in VAE literatures for improved GAN
> learning, and enhance VAEs with an adversarial mechanism for leveraging
> generated samples. Quantitative experiments show generality and
> effectiveness of the imported extensions.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20171021/3c105d6c/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 77, Issue 6
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171023/d2c73ea5/attachment.html>

From weiyu at cs.cmu.edu  Sun Oct 29 02:13:17 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Sun, 29 Oct 2017 06:13:17 +0000
Subject: [AI Seminar] AI Seminar sponsored by Apple -- David Abel (Brown) --
	October 31
Message-ID: <CABzq7eq3rMieCriA4EqYJqtENkjdQNzGP+RCw-vWXJ-BFwqhKw@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you next Tuesday, October 31, at noon in NSH 3305
for AI Seminar sponsored by Apple. To learn more about the seminar series,
please visit the AI Seminar webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, David Abel <https://cs.brown.edu/~dabel/> from Brown
University will
give the following talk:

Title: Abstraction and Lifelong Reinforcement Learning

Abstract:

Lifelong Reinforcement Learning (RL) presents a diversity of challenges.
Agents must effectively transfer knowledge across tasks while
simultaneously addressing exploration, credit assignment, and
generalization. Abstraction can help overcome these hurdles by compressing
the state space or empowering the action space of a learning agent, thereby
reducing the computational and statistical burdens of learning. In this
talk, I summarize our new results on the effect of abstractions on lifelong
RL. First, we introduce a new class of value-preserving state abstractions
whose optimal form can be computed efficiently, improving over existing
NP-Hardness results. Second, we provide a generic sample bound for
computing high confidence state abstractions in the lifelong setting.
Third, we show experimentally that state abstractions only offer marginal
improvements to lifelong learning on their own, but when paired with action
abstraction, can enable efficient learning. Further, joint state-action
abstractions induce a closed operator on representations, thereby yielding
a simple recipe for constructing and analyzing hierarchies for RL.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171029/3870a37b/attachment.html>

From vitercik at cs.cmu.edu  Sun Oct 29 12:01:38 2017
From: vitercik at cs.cmu.edu (Ellen Vitercik)
Date: Sun, 29 Oct 2017 12:01:38 -0400
Subject: [AI Seminar] AI Seminar sponsored by Apple -- David Abel
 (Brown) -- October 31
In-Reply-To: <CABzq7eq3rMieCriA4EqYJqtENkjdQNzGP+RCw-vWXJ-BFwqhKw@mail.gmail.com>
References: <CABzq7eq3rMieCriA4EqYJqtENkjdQNzGP+RCw-vWXJ-BFwqhKw@mail.gmail.com>
Message-ID: <CALVim==hHuZ3umfwWa7aKKHRn6dPDLq+N6UPUny1_ftrzQsRcw@mail.gmail.com>

I've attached a schedule for David's visit. Please sign up for a slot if
you would like to meet with him.

https://docs.google.com/spreadsheets/d/1SPN4ZwO6w0b7hFHdg4cJuen6HtUojy-L4Oh59UpquwY/edit?usp=sharing

Best,
Ellen Vitercik

On Sun, Oct 29, 2017 at 2:13 AM, Adams Wei Yu <weiyu at cs.cmu.edu> wrote:

> Dear faculty and students,
>
> We look forward to seeing you next Tuesday, October 31, at noon in NSH
> 3305 for AI Seminar sponsored by Apple. To learn more about the seminar
> series, please visit the AI Seminar webpage
> <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, David Abel <https://cs.brown.edu/~dabel/> from Brown
> University will give the following talk:
>
> Title: Abstraction and Lifelong Reinforcement Learning
>
> Abstract:
>
> Lifelong Reinforcement Learning (RL) presents a diversity of challenges.
> Agents must effectively transfer knowledge across tasks while
> simultaneously addressing exploration, credit assignment, and
> generalization. Abstraction can help overcome these hurdles by compressing
> the state space or empowering the action space of a learning agent, thereby
> reducing the computational and statistical burdens of learning. In this
> talk, I summarize our new results on the effect of abstractions on lifelong
> RL. First, we introduce a new class of value-preserving state abstractions
> whose optimal form can be computed efficiently, improving over existing
> NP-Hardness results. Second, we provide a generic sample bound for
> computing high confidence state abstractions in the lifelong setting.
> Third, we show experimentally that state abstractions only offer marginal
> improvements to lifelong learning on their own, but when paired with action
> abstraction, can enable efficient learning. Further, joint state-action
> abstractions induce a closed operator on representations, thereby yielding
> a simple recipe for constructing and analyzing hierarchies for RL.
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171029/323bf69e/attachment.html>

From adamsyuwei at gmail.com  Mon Oct 30 11:11:14 2017
From: adamsyuwei at gmail.com (Adams Wei Yu)
Date: Mon, 30 Oct 2017 15:11:14 +0000
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 77, Issue 8
In-Reply-To: <mailman.35.1509292815.19493.ai-seminar-announce@cs.cmu.edu>
References: <mailman.35.1509292815.19493.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7epq2rh-EqPi2JU_WfApwmwzOwHYcA6ybivBDuJMH9=hCg@mail.gmail.com>

A gentle reminder that the talk will happen tomorrow (Tuesday) noon at NSH
3305.

On Sun, Oct 29, 2017 at 4:00 PM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Seminar sponsored by Apple -- David Abel (Brown) --
>       October 31 (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Sun, 29 Oct 2017 06:13:17 +0000
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Cc: David Abel <david_abel at brown.edu>
> Subject: [AI Seminar] AI Seminar sponsored by Apple -- David Abel
>         (Brown) --      October 31
> Message-ID:
>         <CABzq7eq3rMieCriA4EqYJqtENkjdQNzGP+RCw-vWXJ-BFwqhKw at mail.
> gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you next Tuesday, October 31, at noon in NSH 3305
> for AI Seminar sponsored by Apple. To learn more about the seminar series,
> please visit the AI Seminar webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, David Abel <https://cs.brown.edu/~dabel/> from Brown
> University will
> give the following talk:
>
> Title: Abstraction and Lifelong Reinforcement Learning
>
> Abstract:
>
> Lifelong Reinforcement Learning (RL) presents a diversity of challenges.
> Agents must effectively transfer knowledge across tasks while
> simultaneously addressing exploration, credit assignment, and
> generalization. Abstraction can help overcome these hurdles by compressing
> the state space or empowering the action space of a learning agent, thereby
> reducing the computational and statistical burdens of learning. In this
> talk, I summarize our new results on the effect of abstractions on lifelong
> RL. First, we introduce a new class of value-preserving state abstractions
> whose optimal form can be computed efficiently, improving over existing
> NP-Hardness results. Second, we provide a generic sample bound for
> computing high confidence state abstractions in the lifelong setting.
> Third, we show experimentally that state abstractions only offer marginal
> improvements to lifelong learning on their own, but when paired with action
> abstraction, can enable efficient learning. Further, joint state-action
> abstractions induce a closed operator on representations, thereby yielding
> a simple recipe for constructing and analyzing hierarchies for RL.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20171029/3870a37b/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 77, Issue 8
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171030/d1049b45/attachment.html>

From weiyu at cs.cmu.edu  Sat Nov  4 07:13:58 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Sat, 4 Nov 2017 05:13:58 -0600
Subject: [AI Seminar] AI Seminar sponsored by Apple -- Hanxiao Liu -- Nov 07
Message-ID: <CABzq7ep-Wcy=9x+GLOx91jNOnJW6YdEmu+6YAaGNxjAoONm51w@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you next Tuesday, Nov 07, at noon in NSH 1507
(unusual place) for AI Seminar sponsored by Apple. To learn more about the
seminar series, please visit the AI Seminar webpage
<http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Hanxiao Liu <http://www.cs.cmu.edu/~hanxiaol/> will give the
following talk:

Title: Hierarchical Representations for Efficient Architecture Search

Abstract:

We explore efficient neural architecture search methods and present a
simple yet powerful evolutionary algorithm that can discover new
architectures achieving state of the art results. Our approach combines a
novel hierarchical genetic representation scheme that imitates the
modularized design pattern commonly adopted by human experts, and an
expressive search space that supports complex topologies. Our algorithm
efficiently discovers architectures that outperform a large number of
manually designed models for image classification, obtaining top-1 error of
3.6% on CIFAR-10 and 20.3% when transferred to ImageNet, which is
competitive with the best existing neural architecture search approaches
and represents the new state of the art for evolutionary strategies on this
task. We also present results using random search, achieving 0.3% less
top-1 accuracy on CIFAR-10 and 0.1% less on ImageNet whilst reducing the
architecture search time from 36 hours down to 1 hour.

This is joint work with Karen Simonyan, Oriol Vinyals, Chrisantha Fernando
and Koray Kavukcuoglu at DeepMind.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171104/a5452b34/attachment.html>

From weiyu at cs.cmu.edu  Mon Nov  6 11:27:48 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 6 Nov 2017 09:27:48 -0700
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 78, Issue 1
In-Reply-To: <mailman.83.1509901220.16166.ai-seminar-announce@cs.cmu.edu>
References: <mailman.83.1509901220.16166.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7er4m0njMzUfGL8HsWeBLAhVFwwh=azX9pzgD3D_7jiZEA@mail.gmail.com>

A gentle reminder that the talk will happen tomorrow (Tuesday) noon in NSH
1507 (unusual place).

On Sun, Nov 5, 2017 at 10:00 AM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Seminar sponsored by Apple -- Hanxiao Liu -- Nov 07
>       (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Sat, 4 Nov 2017 05:13:58 -0600
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Subject: [AI Seminar] AI Seminar sponsored by Apple -- Hanxiao Liu --
>         Nov 07
> Message-ID:
>         <CABzq7ep-Wcy=9x+GLOx91jNOnJW6YdEmu+6YAaGNxjAoONm51w at mail.gmail.
> com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you next Tuesday, Nov 07, at noon in NSH 1507
> (unusual place) for AI Seminar sponsored by Apple. To learn more about the
> seminar series, please visit the AI Seminar webpage
> <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Hanxiao Liu <http://www.cs.cmu.edu/~hanxiaol/> will give the
> following talk:
>
> Title: Hierarchical Representations for Efficient Architecture Search
>
> Abstract:
>
> We explore efficient neural architecture search methods and present a
> simple yet powerful evolutionary algorithm that can discover new
> architectures achieving state of the art results. Our approach combines a
> novel hierarchical genetic representation scheme that imitates the
> modularized design pattern commonly adopted by human experts, and an
> expressive search space that supports complex topologies. Our algorithm
> efficiently discovers architectures that outperform a large number of
> manually designed models for image classification, obtaining top-1 error of
> 3.6% on CIFAR-10 and 20.3% when transferred to ImageNet, which is
> competitive with the best existing neural architecture search approaches
> and represents the new state of the art for evolutionary strategies on this
> task. We also present results using random search, achieving 0.3% less
> top-1 accuracy on CIFAR-10 and 0.1% less on ImageNet whilst reducing the
> architecture search time from 36 hours down to 1 hour.
>
> This is joint work with Karen Simonyan, Oriol Vinyals, Chrisantha Fernando
> and Koray Kavukcuoglu at DeepMind.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20171104/a5452b34/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 78, Issue 1
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171106/991ec720/attachment.html>

From weiyu at cs.cmu.edu  Sat Nov 11 18:03:08 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Sat, 11 Nov 2017 23:03:08 +0000
Subject: [AI Seminar] AI Seminar sponsored by Apple -- Nika Haghtalab -- Nov
	14
Message-ID: <CABzq7er_Fy6MBoBHA4s31OnxwOD2CQKva8vmFU0wpA8CQxDOYg@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you next Tuesday, Nov 14, at noon in NSH 3305 for
AI Seminar sponsored by Apple. To learn more about the seminar series,
please visit the AI Seminar webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Nika Haghtalab <https://www.cs.cmu.edu/~nhaghtal/> will give
the following talk:

Title: Algorithms for Generalized Topic Modeling

Abstract:

 Topic modeling is an area with significant recent work in the intersection
of algorithms and machine learning. In standard topic models, a topic (such
as sports, business, or politics) is viewed as a probability distribution
\vec a_i over words, and a document is generated by first selecting a
mixture \vec w over topics, and then generating words iid from the
associated mixture \vec w^T A. Given a large collection of such documents,
the goal is to recover the topic vectors and then to correctly classify new
documents according to their topic mixture.

In this work we consider a broad generalization of this framework in which
words are no longer assumed to be drawn iid and instead a topic is a
complex distribution over sequences of paragraphs. Since one could not hope
to even represent such a distribution in general (even if paragraphs are
given using some natural feature representation), we aim instead to
directly learn a document classifier. That is, we aim to learn a predictor
that given a new document, accurately predicts its topic mixture, without
learning the distributions explicitly. We present several natural
conditions under which one can do this efficiently and discuss issues such
as noise tolerance and sample complexity in this model. More generally, our
model can be viewed as a generalization of the multi-view or co-training
setting in machine learning.

This talk is based on joint work with Avrim Blum. To appear in AAAI 2018.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171111/8f67edef/attachment.html>

From weiyu at cs.cmu.edu  Mon Nov 13 17:22:50 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 13 Nov 2017 14:22:50 -0800
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 78, Issue 3
In-Reply-To: <mailman.31.1510506010.15114.ai-seminar-announce@cs.cmu.edu>
References: <mailman.31.1510506010.15114.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7eqbSWWoDqSJs-9iq4dKx1cGqgVhnJsq3=-o3Js=oP74zg@mail.gmail.com>

A gentle reminder that the talk will happen tomorrow (Tuesday) noon at NSH
3305.

On Sun, Nov 12, 2017 at 9:00 AM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Seminar sponsored by Apple -- Nika Haghtalab -- Nov   14
>       (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Sat, 11 Nov 2017 23:03:08 +0000
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Subject: [AI Seminar] AI Seminar sponsored by Apple -- Nika Haghtalab
>         -- Nov  14
> Message-ID:
>         <CABzq7er_Fy6MBoBHA4s31OnxwOD2CQKva8vmFU
> 0wpA8CQxDOYg at mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you next Tuesday, Nov 14, at noon in NSH 3305 for
> AI Seminar sponsored by Apple. To learn more about the seminar series,
> please visit the AI Seminar webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Nika Haghtalab <https://www.cs.cmu.edu/~nhaghtal/> will give
> the following talk:
>
> Title: Algorithms for Generalized Topic Modeling
>
> Abstract:
>
>  Topic modeling is an area with significant recent work in the intersection
> of algorithms and machine learning. In standard topic models, a topic (such
> as sports, business, or politics) is viewed as a probability distribution
> \vec a_i over words, and a document is generated by first selecting a
> mixture \vec w over topics, and then generating words iid from the
> associated mixture \vec w^T A. Given a large collection of such documents,
> the goal is to recover the topic vectors and then to correctly classify new
> documents according to their topic mixture.
>
> In this work we consider a broad generalization of this framework in which
> words are no longer assumed to be drawn iid and instead a topic is a
> complex distribution over sequences of paragraphs. Since one could not hope
> to even represent such a distribution in general (even if paragraphs are
> given using some natural feature representation), we aim instead to
> directly learn a document classifier. That is, we aim to learn a predictor
> that given a new document, accurately predicts its topic mixture, without
> learning the distributions explicitly. We present several natural
> conditions under which one can do this efficiently and discuss issues such
> as noise tolerance and sample complexity in this model. More generally, our
> model can be viewed as a generalization of the multi-view or co-training
> setting in machine learning.
>
> This talk is based on joint work with Avrim Blum. To appear in AAAI 2018.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20171111/8f67edef/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 78, Issue 3
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171113/ed2ccf48/attachment.html>

From weiyu at cs.cmu.edu  Sun Nov 19 08:46:06 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Sun, 19 Nov 2017 05:46:06 -0800
Subject: [AI Seminar] AI Seminar sponsored by Apple -- Vaishnavh Nagarajan
	-- Nov 21
Message-ID: <CABzq7ertCO5ccBL7_h_8d6zzV6MZCQOiu5idosiMg4Y7ytPs3Q@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you next Tuesday, Nov 21, at noon in NSH 3305 for
AI Seminar sponsored by Apple. To learn more about the seminar series,
please visit the AI Seminar webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Vaishnavh Nagarajan
<http://www.cs.cmu.edu/~vaishnan/home/index.html> will give the following
talk:

Title: Gradient Descent GANs are locally stable

Abstract:

Generative modeling, a core problem in unsupervised learning, aims at
understanding data by learning a model that can generate datapoints that
resemble the real-world distribution. Generative Adversarial Networks
(GANs) are an increasingly popular framework that solve this by optimizing
two deep networks, a "discriminator" and a "generator", in tandem.

However, this complex optimization procedure is still poorly understood.
More specifically, it was not known whether equilibrium points of this
system are "locally asymptotically stable" i.e., when initialized
sufficiently close to an equilibrium point, does the optimization procedure
converge to that point? In this work, we analyze the "gradient descent"
form of GAN optimization (i.e., the setting where we simultaneously take
small gradient steps in both generator and discriminator parameters). We
show that even though GAN optimization does not correspond to a
convex-concave game, even for simple parameterizations, under proper
conditions, its equilibrium points are still locally asymptotically stable.
On the other hand, we show that for the recently-proposed Wasserstein GAN
(WGAN), the optimization procedure might cycle around an equilibrium point
without ever converging to it. Finally, motivated by this stability
analysis, we propose an additional regularization term for GAN updates,
which can guarantee local stability for both the WGAN and for the
traditional GAN. Our regularizer also shows practical promise in speeding
up convergence and in addressing a well-known failure mode in GANs called
mode collapse.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171119/e9f7ac27/attachment.html>

From weiyu at cs.cmu.edu  Mon Nov 20 07:04:32 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 20 Nov 2017 04:04:32 -0800
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 78, Issue 5
In-Reply-To: <mailman.43.1511110815.15294.ai-seminar-announce@cs.cmu.edu>
References: <mailman.43.1511110815.15294.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7epMCVb0u-L4HkPxowWHcnXu9bStyRne+yWJtQgxwZSyBQ@mail.gmail.com>

A gentle reminder that the talk will happen tomorrow (Tuesday) noon.

On Sun, Nov 19, 2017 at 9:00 AM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Seminar sponsored by Apple -- Vaishnavh Nagarajan     -- Nov
>       21 (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Sun, 19 Nov 2017 05:46:06 -0800
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Cc: Vaishnavh Nagarajan <vaishnavh.nagarajan at gmail.com>
> Subject: [AI Seminar] AI Seminar sponsored by Apple -- Vaishnavh
>         Nagarajan       -- Nov 21
> Message-ID:
>         <CABzq7ertCO5ccBL7_h_8d6zzV6MZCQOiu5idosiMg4Y7ytPs3
> Q at mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you next Tuesday, Nov 21, at noon in NSH 3305 for
> AI Seminar sponsored by Apple. To learn more about the seminar series,
> please visit the AI Seminar webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Vaishnavh Nagarajan
> <http://www.cs.cmu.edu/~vaishnan/home/index.html> will give the following
> talk:
>
> Title: Gradient Descent GANs are locally stable
>
> Abstract:
>
> Generative modeling, a core problem in unsupervised learning, aims at
> understanding data by learning a model that can generate datapoints that
> resemble the real-world distribution. Generative Adversarial Networks
> (GANs) are an increasingly popular framework that solve this by optimizing
> two deep networks, a "discriminator" and a "generator", in tandem.
>
> However, this complex optimization procedure is still poorly understood.
> More specifically, it was not known whether equilibrium points of this
> system are "locally asymptotically stable" i.e., when initialized
> sufficiently close to an equilibrium point, does the optimization procedure
> converge to that point? In this work, we analyze the "gradient descent"
> form of GAN optimization (i.e., the setting where we simultaneously take
> small gradient steps in both generator and discriminator parameters). We
> show that even though GAN optimization does not correspond to a
> convex-concave game, even for simple parameterizations, under proper
> conditions, its equilibrium points are still locally asymptotically stable.
> On the other hand, we show that for the recently-proposed Wasserstein GAN
> (WGAN), the optimization procedure might cycle around an equilibrium point
> without ever converging to it. Finally, motivated by this stability
> analysis, we propose an additional regularization term for GAN updates,
> which can guarantee local stability for both the WGAN and for the
> traditional GAN. Our regularizer also shows practical promise in speeding
> up convergence and in addressing a well-known failure mode in GANs called
> mode collapse.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20171119/e9f7ac27/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 78, Issue 5
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171120/8b0f2fea/attachment.html>

From weiyu at cs.cmu.edu  Sat Nov 25 07:04:05 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Sat, 25 Nov 2017 04:04:05 -0800
Subject: [AI Seminar] AI Seminar sponsored by Apple -- Brandon Amos -- Nov 28
Message-ID: <CABzq7epSXb=Xg7AHavqpsL04uNjCwdJyeFEMoMiu41vRmfEteg@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you next Tuesday, Nov 28, at noon in NSH 3305 for
AI Seminar sponsored by Apple. To learn more about the seminar series,
please visit the AI Seminar webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Brandon Amos <http://bamos.github.io/> will give the following
talk:

Title: Modern Convex Optimization within Deep Learning

Abstract:

This talk discusses a new paradigm for deep learning that integrates the
solution of optimization problems "into the loop." We highlight two
challenges present in today's deep learning landscape that involve adding
structure to the input or latent space of a model. We will discuss how to
overcome some of these challenges with the use of learnable optimization
sub-problems that subsume standard architectures and layers. These
architectures obtain state-of-the-art empirical results in many domains
such as continuous action reinforcement learning and tasks that involve
learning hard constraints like the game Sudoku.

We will cover topics from these two papers:

1. Input Convex Neural Networks. Brandon Amos, Lei Xu, J. Zico Kolter. ICML
2017. https://arxiv.org/abs/1609.07152.

2. OptNet: Differentiable Optimization as a Layer in Neural Networks.
Brandon Amos, J. Zico Kolter. ICML 2017. https://arxiv.org/abs/1703.00443.

Joint work with J. Zico Kolter.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171125/9febc24f/attachment.html>

From weiyu at cs.cmu.edu  Mon Nov 27 08:10:30 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 27 Nov 2017 05:10:30 -0800
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 78, Issue 7
In-Reply-To: <mailman.57.1511629217.15580.ai-seminar-announce@cs.cmu.edu>
References: <mailman.57.1511629217.15580.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7epWaG2bDK4rWJC+-ibjtzyOSyfMGc-v2wntWSpx_ocbjQ@mail.gmail.com>

A gentle reminder that the talk will happen tomorrow (Tuesday) noon.

On Sat, Nov 25, 2017 at 9:00 AM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Seminar sponsored by Apple -- Brandon Amos -- Nov 28
>       (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Sat, 25 Nov 2017 04:04:05 -0800
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Subject: [AI Seminar] AI Seminar sponsored by Apple -- Brandon Amos --
>         Nov 28
> Message-ID:
>         <CABzq7epSXb=Xg7AHavqpsL04uNjCwdJyeFEMoMiu41vRmfEteg at mail.
> gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you next Tuesday, Nov 28, at noon in NSH 3305 for
> AI Seminar sponsored by Apple. To learn more about the seminar series,
> please visit the AI Seminar webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Brandon Amos <http://bamos.github.io/> will give the following
> talk:
>
> Title: Modern Convex Optimization within Deep Learning
>
> Abstract:
>
> This talk discusses a new paradigm for deep learning that integrates the
> solution of optimization problems "into the loop." We highlight two
> challenges present in today's deep learning landscape that involve adding
> structure to the input or latent space of a model. We will discuss how to
> overcome some of these challenges with the use of learnable optimization
> sub-problems that subsume standard architectures and layers. These
> architectures obtain state-of-the-art empirical results in many domains
> such as continuous action reinforcement learning and tasks that involve
> learning hard constraints like the game Sudoku.
>
> We will cover topics from these two papers:
>
> 1. Input Convex Neural Networks. Brandon Amos, Lei Xu, J. Zico Kolter. ICML
> 2017. https://arxiv.org/abs/1609.07152.
>
> 2. OptNet: Differentiable Optimization as a Layer in Neural Networks.
> Brandon Amos, J. Zico Kolter. ICML 2017. https://arxiv.org/abs/1703.00443.
>
> Joint work with J. Zico Kolter.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20171125/9febc24f/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 78, Issue 7
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171127/f1f9cf54/attachment.html>

From weiyu at cs.cmu.edu  Fri Dec  1 08:03:04 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Fri, 1 Dec 2017 05:03:04 -0800
Subject: [AI Seminar] AI Seminar sponsored by Apple -- Anson Kahng -- Dec 05
Message-ID: <CABzq7eoTk5wwF7hQUe-p6B9-Rvt2RJ9r-5wfS7PZ3RSgOujOVQ@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you next Tuesday, Dec 05, at noon in NSH 3305 for
AI Seminar sponsored by Apple. To learn more about the seminar series,
please visit the AI Seminar webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Anson Kahng <https://www.cs.cmu.edu/~akahng/> will give the
following talk:

Title: Impartial Rank Aggregation

Abstract:

We study rank aggregation algorithms that take as input the opinions of
players over their peers, represented as rankings, and output a social
ordering of the players (which reflects, e.g., relative contribution to a
project or fit for a job). To prevent strategic behavior, these algorithms
must be impartial, i.e., players should not be able to influence their own
position in the output ranking. We design several randomized algorithms
that are impartial and closely emulate given (non-impartial) rank
aggregation rules in a rigorous sense. Experimental results further support
the efficacy and practicability of our algorithms.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171201/12ed7eda/attachment.html>

From adamsyuwei at gmail.com  Mon Dec  4 09:32:42 2017
From: adamsyuwei at gmail.com (Adams Wei Yu)
Date: Mon, 4 Dec 2017 06:32:42 -0800
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 79, Issue 1
In-Reply-To: <mailman.97.1512147624.27460.ai-seminar-announce@cs.cmu.edu>
References: <mailman.97.1512147624.27460.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7eo_bVRkqNC0fwxxY=5ENMf4d8TJZT3Aj9rdHeVG4VT1Nw@mail.gmail.com>

A gentle reminder that the talk will happen tomorrow (Tuesday) noon.

On Fri, Dec 1, 2017 at 9:00 AM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Seminar sponsored by Apple -- Anson Kahng -- Dec 05
>       (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Fri, 1 Dec 2017 05:03:04 -0800
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Subject: [AI Seminar] AI Seminar sponsored by Apple -- Anson Kahng --
>         Dec 05
> Message-ID:
>         <CABzq7eoTk5wwF7hQUe-p6B9-Rvt2RJ9r-5wfS7PZ3RSgOujOVQ@
> mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you next Tuesday, Dec 05, at noon in NSH 3305 for
> AI Seminar sponsored by Apple. To learn more about the seminar series,
> please visit the AI Seminar webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Anson Kahng <https://www.cs.cmu.edu/~akahng/> will give the
> following talk:
>
> Title: Impartial Rank Aggregation
>
> Abstract:
>
> We study rank aggregation algorithms that take as input the opinions of
> players over their peers, represented as rankings, and output a social
> ordering of the players (which reflects, e.g., relative contribution to a
> project or fit for a job). To prevent strategic behavior, these algorithms
> must be impartial, i.e., players should not be able to influence their own
> position in the output ranking. We design several randomized algorithms
> that are impartial and closely emulate given (non-impartial) rank
> aggregation rules in a rigorous sense. Experimental results further support
> the efficacy and practicability of our algorithms.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20171201/12ed7eda/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 79, Issue 1
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171204/36c50806/attachment.html>

From weiyu at cs.cmu.edu  Sun Dec 10 03:06:40 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Sun, 10 Dec 2017 00:06:40 -0800
Subject: [AI Seminar] AI Seminar sponsored by Apple -- Veeranjaneyulu
	Sadhanala -- Dec 12
Message-ID: <CABzq7epyXB7yEg11M5RFnvt5PQtuhfCdopduPKMptYz620jeJA@mail.gmail.com>

Dear faculty and students,

We look forward to seeing you next Tuesday, Dec 12, at noon in NSH 3305 for
AI Seminar sponsored by Apple. To learn more about the seminar series,
please visit the AI Seminar webpage <http://www.cs.cmu.edu/~aiseminar/>.

On Tuesday, Veeranjaneyulu Sadhanala <http://www.cs.cmu.edu/~vsadhana/>
will give the following talk:

Title: Escaping saddle points in neural network training and other
non-convex optimization problems

Abstract:

In non-convex optimization problems, first-order based methods can get
stuck at saddle points which are not even local minima. The generalization
error at saddle points is typically large and hence it is important to move
away from them. We discuss recently developed algorithms to escape saddle
points. In particular, we discuss gradient descent perturbed with additive
isotropic noise and Newton method with cubic regularization. They converge
to \epsilon-second order stationary points (informally, local minima) in
O(polylog(d) / \epsilon^2) time and O(1/ \epsilon^1.5) iterations
respectively under some conditions on the structure of the objective
function.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171210/fa91fcc6/attachment.html>

From weiyu at cs.cmu.edu  Mon Dec 11 07:17:46 2017
From: weiyu at cs.cmu.edu (Adams Wei Yu)
Date: Mon, 11 Dec 2017 04:17:46 -0800
Subject: [AI Seminar] ai-seminar-announce Digest, Vol 79, Issue 3
In-Reply-To: <mailman.33.1512925213.2444.ai-seminar-announce@cs.cmu.edu>
References: <mailman.33.1512925213.2444.ai-seminar-announce@cs.cmu.edu>
Message-ID: <CABzq7ep0kbf76QNYRX1QrVtzYMTQe_0cm==r6zCkFpdv_XD6Dw@mail.gmail.com>

A gentle reminder that the talk will happen tomorrow (Tuesday) noon. This
is the last AI seminar in 2017.

On Sun, Dec 10, 2017 at 9:00 AM, <ai-seminar-announce-request at cs.cmu.edu>
wrote:

> Send ai-seminar-announce mailing list submissions to
>         ai-seminar-announce at cs.cmu.edu
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-
> seminar-announce
> or, via email, send a message with subject or body 'help' to
>         ai-seminar-announce-request at cs.cmu.edu
>
> You can reach the person managing the list at
>         ai-seminar-announce-owner at cs.cmu.edu
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of ai-seminar-announce digest..."
>
>
> Today's Topics:
>
>    1.  AI Seminar sponsored by Apple -- Veeranjaneyulu  Sadhanala --
>       Dec 12 (Adams Wei Yu)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Sun, 10 Dec 2017 00:06:40 -0800
> From: Adams Wei Yu <weiyu at cs.cmu.edu>
> To: ai-seminar-announce at cs.cmu.edu
> Cc: Veeranjaneyulu Sadhanala <veeranjaneyulus at gmail.com>
> Subject: [AI Seminar] AI Seminar sponsored by Apple -- Veeranjaneyulu
>         Sadhanala -- Dec 12
> Message-ID:
>         <CABzq7epyXB7yEg11M5RFnvt5PQtuhfCdopduPKMptYz620jeJA at mail.
> gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Dear faculty and students,
>
> We look forward to seeing you next Tuesday, Dec 12, at noon in NSH 3305 for
> AI Seminar sponsored by Apple. To learn more about the seminar series,
> please visit the AI Seminar webpage <http://www.cs.cmu.edu/~aiseminar/>.
>
> On Tuesday, Veeranjaneyulu Sadhanala <http://www.cs.cmu.edu/~vsadhana/>
> will give the following talk:
>
> Title: Escaping saddle points in neural network training and other
> non-convex optimization problems
>
> Abstract:
>
> In non-convex optimization problems, first-order based methods can get
> stuck at saddle points which are not even local minima. The generalization
> error at saddle points is typically large and hence it is important to move
> away from them. We discuss recently developed algorithms to escape saddle
> points. In particular, we discuss gradient descent perturbed with additive
> isotropic noise and Newton method with cubic regularization. They converge
> to \epsilon-second order stationary points (informally, local minima) in
> O(polylog(d) / \epsilon^2) time and O(1/ \epsilon^1.5) iterations
> respectively under some conditions on the structure of the objective
> function.
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-
> announce/attachments/20171210/fa91fcc6/attachment-0001.html>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> ai-seminar-announce mailing list
> ai-seminar-announce at cs.cmu.edu
> https://mailman.srv.cs.cmu.edu/mailman/listinfo/ai-seminar-announce
>
> ------------------------------
>
> End of ai-seminar-announce Digest, Vol 79, Issue 3
> **************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/ai-seminar-announce/attachments/20171211/a2aec82f/attachment.html>