From dhouston at cs.cmu.edu Wed Jan 11 11:52:54 2012 From: dhouston at cs.cmu.edu (Dana Houston) Date: Wed, 11 Jan 2012 11:52:54 -0500 Subject: [Intelligence Seminar] Jan. 16, 3:30pm:, Presentation by Andreas Krause Message-ID: <4F0DBE66.2010808@cs.cmu.edu> INTELLIGENCE SEMINAR JANUARY 16 (MONDAY) AT 3:30PM, IN GHC 4405 (PLEASE NOTE THE UNUSUAL DAY AND TIME) SPEAKER: ANDREAS KRAUSE (Swiss Federal Institute of Technology, Zurich) Host: Carlos Guestrin For meetings, contact Michelle Martin (michelle324 at cs.cmu.edu) SEQUENTIAL DECISION MAKING IN EXPERIMENTAL DESIGN SUSTAINABILITY VIA ADAPTIVE SUBMODULARITY Joint work with Daniel Golovin Solving sequential decision problems under partial observability is a fundamental but notoriously difficult challenge. I will introduce the new concept of adaptive submodularity, generalizing the classical notion of submodular set functions to adaptive policies. We prove that, if a problem satisfies this property, a simple adaptive greedy algorithm is guaranteed to be competitive with the optimal policy. The concept allows us to recover, generalize, and extend existing results in diverse applications, including sensor management, viral marketing, and active learning. I will focus on two case studies. In an application to Bayesian experimental design, we show how greedy optimization of a novel adaptive submodular criterion outperforms standard myopic techniques based on information gain and value of information. I will also discuss how adaptive submodularity can help to address problems in computational sustainability by presenting results on conservation planning for three rare species in the Pacific Northwest of the United States. BIO Andreas Krause received his Diplom in Computer Science and Mathematics from the Technical University of Munich (2004) and his Ph.D. in Computer Science from Carnegie Mellon University (2008). He joined the California Institute of Technology as an assistant professor of computer science in 2009, and is currently assistant professor in the Department of Computer Science at the Swiss Federal Institute of Technology, Zurich. His research is in adaptive systems that actively acquire information, reason, and make decisions in large, distributed, and uncertain domains, such as sensor networks and the web. Dr. Krause is a 2010 National Academy of Sciences' Kavli Frontiers Fellow. He received an NSF CAREER award, the Okawa Foundation Research Grant recognizing top young researchers in telecommunications, as well as awards at several premier conferences (AAAI, KDD, IPSN, ICML, UAI) and the ASCE Journal of Water Resources Planning and Management. -- Dana M. Houston Language Technologies Institute School of Computer Science Carnegie Mellon University 5405 Gates Hillman Complex 5000 Forbes Avenue Pittsburgh, PA 15213 T: (412)268-4717 F: (412)268-6298 From dhouston at cs.cmu.edu Fri Jan 13 09:56:51 2012 From: dhouston at cs.cmu.edu (Dana Houston) Date: Fri, 13 Jan 2012 09:56:51 -0500 Subject: [Intelligence Seminar] Jan. 16, 3:30pm:, Presentation by Andreas Krause In-Reply-To: <4F0DBE66.2010808@cs.cmu.edu> References: <4F0DBE66.2010808@cs.cmu.edu> Message-ID: <4F104633.7050400@cs.cmu.edu> INTELLIGENCE SEMINAR > JANUARY 16 (MONDAY) AT 3:30PM, IN GHC 4405 > (PLEASE NOTE THE UNUSUAL DAY AND TIME) > > SPEAKER: ANDREAS KRAUSE (Swiss Federal Institute of Technology, Zurich) > Host: Carlos Guestrin > For meetings, contact Michelle Martin (michelle324 at cs.cmu.edu) > > SEQUENTIAL DECISION MAKING IN EXPERIMENTAL DESIGN > SUSTAINABILITY VIA ADAPTIVE SUBMODULARITY > > Joint work with Daniel Golovin > > Solving sequential decision problems under partial observability is a > fundamental but notoriously difficult challenge. I will introduce the > new concept of adaptive submodularity, generalizing the classical > notion of submodular set functions to adaptive policies. We prove > that, if a problem satisfies this property, a simple adaptive greedy > algorithm is guaranteed to be competitive with the optimal policy. > The concept allows us to recover, generalize, and extend existing > results in diverse applications, including sensor management, viral > marketing, and active learning. I will focus on two case studies. In > an application to Bayesian experimental design, we show how greedy > optimization of a novel adaptive submodular criterion outperforms > standard myopic techniques based on information gain and value of > information. I will also discuss how adaptive submodularity can help > to address problems in computational sustainability by presenting > results on conservation planning for three rare species in the Pacific > Northwest of the United States. > > BIO > > Andreas Krause received his Diplom in Computer Science and Mathematics > from the Technical University of Munich (2004) and his Ph.D. in > Computer Science from Carnegie Mellon University (2008). He joined the > California Institute of Technology as an assistant professor of > computer science in 2009, and is currently assistant professor in the > Department of Computer Science at the Swiss Federal Institute of > Technology, Zurich. His research is in adaptive systems that actively > acquire information, reason, and make decisions in large, distributed, > and uncertain domains, such as sensor networks and the web. Dr. Krause > is a 2010 National Academy of Sciences' Kavli Frontiers Fellow. He > received an NSF CAREER award, the Okawa Foundation Research Grant > recognizing top young researchers in telecommunications, as well as > awards at several premier conferences (AAAI, KDD, IPSN, ICML, UAI) and > the ASCE Journal of Water Resources Planning and Management. > -- Dana M. Houston Language Technologies Institute School of Computer Science Carnegie Mellon University 5405 Gates Hillman Complex 5000 Forbes Avenue Pittsburgh, PA 15213 T: (412)268-4717 F: (412)268-6298 From arielpro at cs.cmu.edu Wed Feb 29 22:52:16 2012 From: arielpro at cs.cmu.edu (Ariel Procaccia) Date: Wed, 29 Feb 2012 22:52:16 -0500 Subject: [Intelligence Seminar] Algorithmic economics talk: Michael Wellman, March 13 2012 @ NSH 3305 Message-ID: Hi everyone, Michael Wellman of the University of Michigan will visit CMU on March 13 (during Spring break), and will give a talk entitled "Empirical game-theoretic analysis for canonical auctions games" at noon in NSH 3305. Lunch will be served (hopefully not too late this time!). Please find the talk details here: http://www.cs.cmu.edu/~arielpro/algecon.html. Michael will be available for meetings on March 13. If you want to set up a meeting, please send an email to Gayle Bishop, gayle at cs.cmu.edu, with your availability. Please let Gayle know by March 5. Best regards, Ariel -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mailman.srv.cs.cmu.edu/pipermail/intelligence-seminar-announce/attachments/20120229/99410b8e/attachment-0001.html From arielpro at cs.cmu.edu Fri Mar 9 10:10:14 2012 From: arielpro at cs.cmu.edu (Ariel Procaccia) Date: Fri, 9 Mar 2012 10:10:14 -0500 Subject: [Intelligence Seminar] Reminder: Michael Wellman on March 13 Message-ID: Hi everyone, This is is a reminder that Michael Wellman will visit on Tuesday and will give a lunch seminar at noon in NSH 3305 (see below). In addition, he still has some open slots for meetings; please send an email to Gayle ( gayle at cs.cmu.edu) for an appointment. Best, Ariel On Wed, Feb 29, 2012 at 10:52 PM, Ariel Procaccia wrote: > Hi everyone, > > Michael Wellman of the University of Michigan will visit CMU on March 13 > (during Spring break), and will give a talk entitled "Empirical > game-theoretic analysis for canonical auctions games" at noon in NSH 3305. > Lunch will be served (hopefully not too late this time!). Please find the > talk details here: http://www.cs.cmu.edu/~arielpro/algecon.html. > > Michael will be available for meetings on March 13. If you want to set up > a meeting, please send an email to Gayle Bishop, gayle at cs.cmu.edu, with > your availability. Please let Gayle know by March 5. > > Best regards, > Ariel > -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mailman.srv.cs.cmu.edu/pipermail/intelligence-seminar-announce/attachments/20120309/802afd18/attachment.html From arielpro at cs.cmu.edu Tue Mar 13 10:52:52 2012 From: arielpro at cs.cmu.edu (Ariel Procaccia) Date: Tue, 13 Mar 2012 10:52:52 -0400 Subject: [Intelligence Seminar] Last reminder: Michael Wellman, today at noon @ NSH 3305 Message-ID: Michael Wellman will give an algorithmic economics talk today at noon in NSH 3305 (see http://www.cs.cmu.edu/~arielpro/algecon.html). It's Spring break so there are no excuses not to come if you happen to be around! Lunch will be served. See you at noon, Ariel On Wed, Feb 29, 2012 at 10:52 PM, Ariel Procaccia wrote: > Hi everyone, > > Michael Wellman of the University of Michigan will visit CMU on March 13 > (during Spring break), and will give a talk entitled "Empirical > game-theoretic analysis for canonical auctions games" at noon in NSH 3305. > Lunch will be served (hopefully not too late this time!). Please find the > talk details here: http://www.cs.cmu.edu/~arielpro/algecon.html. > > Michael will be available for meetings on March 13. If you want to set up > a meeting, please send an email to Gayle Bishop, gayle at cs.cmu.edu, with > your availability. Please let Gayle know by March 5. > > Best regards, > Ariel > -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mailman.srv.cs.cmu.edu/pipermail/intelligence-seminar-announce/attachments/20120313/3d089a60/attachment.html From dhouston at cs.cmu.edu Tue Mar 13 11:39:03 2012 From: dhouston at cs.cmu.edu (Dana Houston) Date: Tue, 13 Mar 2012 11:39:03 -0400 Subject: [Intelligence Seminar] March 20, 3:30pm:, Presentation by Gal Kaminka Message-ID: <4F5F6A17.5070300@cs.cmu.edu> INTELLIGENCE SEMINAR MARCH 20 AT 3:30PM, IN GHC 4303 SPEAKER: GAL KAMINKA (Bar Ilan University) Host: Manuela Veloso For meetings, contact Gayle Bishop (gayle at cs.cmu.edu) REUSABLE TEAMWORK FOR MULTI-ROBOT TEAMS For many years, multi-robot researchers have focused on specific application-inspired basic tasks (e.g., coverage, moving in formation, foraging, patrolling) as a way of studying cooperation between robots. But users want to see increasingly complex missions being tackled, which challenge this methodology: first, some missions cannot be easily decomposed into the familiar basic tasks, making previous knowledge non-reusable; second, the target operating environments challenge the typically sterile settings assumed in many previous works (such challenges include adversaries, multiple concurrent goals, human operators and users, and more). In this talk, I will argue that the reusable components in complex missions are often found not in the tasks, but in the interactions between robots; that is, while taskwork varies significantly, teamwork is largely generic. And while many multi-robot researchers have begun exploring generic task-allocation methods, I will report on my group's work over the last decade, identifying and developing other general mechanisms for teamwork, and integrating them at the architecture level to facilitate development of robust teams at reduced programming effort. I will sample some of our results in developing robots for missions ranging from robust formation maintenance, through patrolling, to soccer and urban search-and-rescue. BIO Gal A. Kaminka is an associate professor in the Computer Science Department, and the Brain Sciences Research Center, at Bar Ilan University (Israel). His research expertise includes multi-agent and multi-robot systems, teamwork and coordination, behavior and plan recognition, and modeling social behavior. He received his Ph.D. from the University of Southern California (2000), and spent two years as a post-doctorate fellow at Carnegie Mellon. Today, Prof. Kaminka leads the MAVERICK research group at Bar Ilan University, supervising over a dozen M.Sc. and Ph.D. students. He was awarded an IBM faculty award and top places at international robotics competitions. He served as the program chair of the 2008 Israeli Conference on Robotics, and the program co-chair of the 2010 International Joint Conference on Autonomous Agents and Multi-Agent Systems (AAMAS). He has served on the international executive bodies of the International Foundation of Autonomous Agents and Multi-Agent Systems (IFAAMAS) and the Association for Advancement of Artificial Intelligence (AAAI). Currently, he is spending his sabbatical as a Radcliffe Fellow at Harvard University's Radcliffe Institute for Advanced Study. -- Dana M. Houston Language Technologies Institute School of Computer Science Carnegie Mellon University 5405 Gates Hillman Complex 5000 Forbes Avenue Pittsburgh, PA 15213 T: (412)268-4717 F: (412)268-6298 From dhouston at cs.cmu.edu Thu Mar 15 12:32:31 2012 From: dhouston at cs.cmu.edu (Dana Houston) Date: Thu, 15 Mar 2012 12:32:31 -0400 Subject: [Intelligence Seminar] March 20, 3:30pm:, Presentation by Gal Kaminka In-Reply-To: <4F5F6A17.5070300@cs.cmu.edu> References: <4F5F6A17.5070300@cs.cmu.edu> Message-ID: <4F62199F.5090604@cs.cmu.edu> > INTELLIGENCE SEMINAR > MARCH 20 AT 3:30PM, IN GHC 4303 > > SPEAKER: GAL KAMINKA (Bar Ilan University) > Host: Manuela Veloso > For meetings, contact Dana Houston (dhouston at cs.cmu.edu) > > REUSABLE TEAMWORK FOR MULTI-ROBOT TEAMS > > For many years, multi-robot researchers have focused on specific > application-inspired basic tasks (e.g., coverage, moving in formation, > foraging, patrolling) as a way of studying cooperation between robots. > But users want to see increasingly complex missions being tackled, > which challenge this methodology: first, some missions cannot be > easily decomposed into the familiar basic tasks, making previous > knowledge non-reusable; second, the target operating environments > challenge the typically sterile settings assumed in many previous > works (such challenges include adversaries, multiple concurrent goals, > human operators and users, and more). > > In this talk, I will argue that the reusable components in complex > missions are often found not in the tasks, but in the interactions > between robots; that is, while taskwork varies significantly, teamwork > is largely generic. And while many multi-robot researchers have begun > exploring generic task-allocation methods, I will report on my group's > work over the last decade, identifying and developing other general > mechanisms for teamwork, and integrating them at the architecture > level to facilitate development of robust teams at reduced programming > effort. I will sample some of our results in developing robots for > missions ranging from robust formation maintenance, through > patrolling, to soccer and urban search-and-rescue. > > BIO > > Gal A. Kaminka is an associate professor in the Computer Science > Department, and the Brain Sciences Research Center, at Bar Ilan > University (Israel). His research expertise includes multi-agent and > multi-robot systems, teamwork and coordination, behavior and plan > recognition, and modeling social behavior. He received his Ph.D. from > the University of Southern California (2000), and spent two years as a > post-doctorate fellow at Carnegie Mellon. Today, Prof. Kaminka leads > the MAVERICK research group at Bar Ilan University, supervising over a > dozen M.Sc. and Ph.D. students. He was awarded an IBM faculty award > and top places at international robotics competitions. He served as > the program chair of the 2008 Israeli Conference on Robotics, and the > program co-chair of the 2010 International Joint Conference on > Autonomous Agents and Multi-Agent Systems (AAMAS). He has served on > the international executive bodies of the International Foundation of > Autonomous Agents and Multi-Agent Systems (IFAAMAS) and the > Association for Advancement of Artificial Intelligence (AAAI). > Currently, he is spending his sabbatical as a Radcliffe Fellow at > Harvard University's Radcliffe Institute for Advanced Study. > -- Dana M. Houston Language Technologies Institute School of Computer Science Carnegie Mellon University 5405 Gates Hillman Complex 5000 Forbes Avenue Pittsburgh, PA 15213 T: (412)268-4717 F: (412)268-6298 From dhouston at cs.cmu.edu Mon Mar 19 09:15:06 2012 From: dhouston at cs.cmu.edu (Dana Houston) Date: Mon, 19 Mar 2012 09:15:06 -0400 Subject: [Intelligence Seminar] March 20, 3:30pm:, Presentation by Gal Kaminka In-Reply-To: <4F62199F.5090604@cs.cmu.edu> References: <4F5F6A17.5070300@cs.cmu.edu> <4F62199F.5090604@cs.cmu.edu> Message-ID: <4F67315A.6000907@cs.cmu.edu> > >> INTELLIGENCE SEMINAR >> MARCH 20 AT 3:30PM, IN GHC 4303 >> >> SPEAKER: GAL KAMINKA (Bar Ilan University) >> Host: Manuela Veloso >> For meetings, contact Dana Houston (dhouston at cs.cmu.edu) >> >> REUSABLE TEAMWORK FOR MULTI-ROBOT TEAMS >> >> For many years, multi-robot researchers have focused on specific >> application-inspired basic tasks (e.g., coverage, moving in formation, >> foraging, patrolling) as a way of studying cooperation between robots. >> But users want to see increasingly complex missions being tackled, >> which challenge this methodology: first, some missions cannot be >> easily decomposed into the familiar basic tasks, making previous >> knowledge non-reusable; second, the target operating environments >> challenge the typically sterile settings assumed in many previous >> works (such challenges include adversaries, multiple concurrent goals, >> human operators and users, and more). >> >> In this talk, I will argue that the reusable components in complex >> missions are often found not in the tasks, but in the interactions >> between robots; that is, while taskwork varies significantly, teamwork >> is largely generic. And while many multi-robot researchers have begun >> exploring generic task-allocation methods, I will report on my group's >> work over the last decade, identifying and developing other general >> mechanisms for teamwork, and integrating them at the architecture >> level to facilitate development of robust teams at reduced programming >> effort. I will sample some of our results in developing robots for >> missions ranging from robust formation maintenance, through >> patrolling, to soccer and urban search-and-rescue. >> >> BIO >> >> Gal A. Kaminka is an associate professor in the Computer Science >> Department, and the Brain Sciences Research Center, at Bar Ilan >> University (Israel). His research expertise includes multi-agent and >> multi-robot systems, teamwork and coordination, behavior and plan >> recognition, and modeling social behavior. He received his Ph.D. from >> the University of Southern California (2000), and spent two years as a >> post-doctorate fellow at Carnegie Mellon. Today, Prof. Kaminka leads >> the MAVERICK research group at Bar Ilan University, supervising over a >> dozen M.Sc. and Ph.D. students. He was awarded an IBM faculty award >> and top places at international robotics competitions. He served as >> the program chair of the 2008 Israeli Conference on Robotics, and the >> program co-chair of the 2010 International Joint Conference on >> Autonomous Agents and Multi-Agent Systems (AAMAS). He has served on >> the international executive bodies of the International Foundation of >> Autonomous Agents and Multi-Agent Systems (IFAAMAS) and the >> Association for Advancement of Artificial Intelligence (AAAI). >> Currently, he is spending his sabbatical as a Radcliffe Fellow at >> Harvard University's Radcliffe Institute for Advanced Study. >> > -- Dana M. Houston Language Technologies Institute School of Computer Science Carnegie Mellon University 5405 Gates Hillman Complex 5000 Forbes Avenue Pittsburgh, PA 15213 T: (412)268-4717 F: (412)268-6298 From dhouston at cs.cmu.edu Mon Apr 16 08:36:44 2012 From: dhouston at cs.cmu.edu (Dana Houston) Date: Mon, 16 Apr 2012 08:36:44 -0400 Subject: [Intelligence Seminar] April 23, 10:30am:, Presentation by Paul Rosenbloom In-Reply-To: References: Message-ID: <4F8C125C.9040200@cs.cmu.edu> ========================================================= WE HAVE TWO INTELLIGENCE SEMINAR PRESENTATIONS NEXT WEEK: PAUL ROSENBLOOM: APRIL 23, 10:30AM, IN GHC 6501 KEVIN LEYTON-BROWN: APRIL 24, 3:30PM, IN GHC 4303 THE ANNOUNCEMENT BELOW IS FOR THE FIRST PRESENTATION ========================================================= INTELLIGENCE SEMINAR APRIL 23 AT 10:30AM, IN GHC 6501 (UNUSUAL DAY, TIME, AND PLACE) SPEAKER: PAUL ROSENBLOOM (University of Southern California) Host: Philip Lehman For meetings, contact June Fischerkeller (jfische at cs.cmu.edu) TOWARDS A GRAPHICAL COGNITIVE ARCHITECTURE FOR VIRTUAL HUMANS (AND INTELLIGENT AGENTS/ROBOTS) A cognitive architecture provides a hypothesis about the fixed structure (and its integration) underlying intelligent behavior, whether in natural or artificial systems. The overall goal of this effort is to leverage graphical models with their ability to uniformly yield state-of-the-art algorithms across symbol, probability, and signal processing in developing a new architecture that goes significantly beyond today's best in providing, and tightly integrating together, the capabilities required for virtual humans (and intelligent agents/robots). The current focus is on a graphical mixed (i.e., statistical relational) architecture that supports hybrid processing through its grounding in a continuous representation. Aspects of memory, problem solving, perception, imagery, learning, and natural language have been demonstrated to date in this architecture, although some are still mere beginnings. The talk will introduce the desiderata for this graphical architecture, explain the basics of its operation, and highlight progress on some of these capabilities. BIO Paul S. Rosenbloom is a Professor of Computer Science at the University of Southern California (USC) and a Project Leader at USC's Institute for Creative Technologies. He spent twenty years at USC's Information Sciences Institute, including a decade leading new directions and a stint as Deputy Director. Earlier he was an Assistant Professor of Computer Science and Psychology at Stanford University, and a Research Computer Scientist at Carnegie Mellon University. He received his B.S. in Mathematical Sciences (with distinction) from Stanford University and his M.S. and Ph.D. in Computer Science from Carnegie Mellon University. He is a Fellow of the Association for the Advancement of Artificial Intelligence (AAAI). Prof. Rosenbloom's research focuses on cognitive architectures; he was a co-PI of the Soar Project for fifteen years, and is currently developing a new approach based on graphical models. He has also been working to understand the nature and structure of computing as a scientific domain. -- Dana M. Houston Language Technologies Institute School of Computer Science Carnegie Mellon University 5405 Gates Hillman Complex 5000 Forbes Avenue Pittsburgh, PA 15213 T: (412)268-4717 F: (412)268-6298 From dhouston at cs.cmu.edu Mon Apr 16 08:39:11 2012 From: dhouston at cs.cmu.edu (Dana Houston) Date: Mon, 16 Apr 2012 08:39:11 -0400 Subject: [Intelligence Seminar] April 24, 3:30pm:, Presentation by Kevin Leyton-Brown In-Reply-To: References: Message-ID: <4F8C12EF.8040007@cs.cmu.edu> ========================================================= WE HAVE TWO INTELLIGENCE SEMINAR PRESENTATIONS NEXT WEEK: PAUL ROSENBLOOM: APRIL 23, 10:30AM, IN GHC 6501 KEVIN LEYTON-BROWN: APRIL 24, 3:30PM, IN GHC 4303 THE ANNOUNCEMENT BELOW IS FOR THE SECOND PRESENTATION ========================================================= INTELLIGENCE SEMINAR APRIL 24 AT 3:30PM, IN GHC 4303 SPEAKER: KEVIN LEYTON-BROWN (University of British Columbia) Host: Ariel Procaccia For meetings, contact Dana Houston (dhouston at cs.cmu.edu) BEYOND EQUILIBRIUM: PREDICTING HUMAN BEHAVIOR IN NORMAL FORM GAMES It is standard in multiagent settings to assume that agents will adopt Nash equilibrium strategies. However, studies in experimental economics demonstrate that Nash equilibrium is a poor description of human players initial behavior in normal form games. In this talk, I will describe a wide range of widely studied models from behavioral game theory (BGT). For what we believe is the first time, we evaluated each of these models in a meta-analysis, taking as our dataset large-scale and publicly available experimental data from the BGT literature. We also analyzed the parameters of the best performing model, and identified ways of modifying it--and, indeed, simplifying it--to improve performance. In the end, our work demonstrates two surprising facts: one BGT model was consistently the best, and people are smarter than behavioral game theorists had thought. BIO Kevin Leyton-Brown is an associate professor in computer science at the University of British Columbia. He holds a Ph.D. and M.S. from Stanford University (2003; 2001) and a B.S. from McMaster University (1998). Much of his work is at the intersection of computer science and microeconomics, addressing computational problems in economic contexts and incentive issues in multiagent systems. He also studies the application of machine learning to the automated design and analysis of algorithms for solving hard computational problems. He has co-written two books, "Multiagent Systems" and "Essentials of Game Theory," and over seventy peer-refereed technical articles. He is the program chair for the ACM Conference on Electronic Commerce (ACM-EC), and an associate editor of the Journal of Artificial Intelligence Research (JAIR), the Artificial Intelligence Journal (AIJ), and ACM Transactions on Economics and Computation. He split his 2010-2011 sabbatical between Makerere University in Kampala, Uganda, and the Institute for Advanced Studies at Hebrew University in Jerusalem, Israel. He has served as a consultant for Trading Dynamics Inc., Ariba Inc., and Cariocas Inc., and was a scientific advisor to Zite Inc. until it was acquired by CNN in 2011. -- Dana M. Houston Language Technologies Institute School of Computer Science Carnegie Mellon University 5405 Gates Hillman Complex 5000 Forbes Avenue Pittsburgh, PA 15213 T: (412)268-4717 F: (412)268-6298 From dhouston at cs.cmu.edu Thu Apr 19 13:44:55 2012 From: dhouston at cs.cmu.edu (Dana Houston) Date: Thu, 19 Apr 2012 13:44:55 -0400 Subject: [Intelligence Seminar] April 24, 3:30pm:, Presentation by Kevin Leyton-Brown In-Reply-To: <4F8C12EF.8040007@cs.cmu.edu> References: <4F8C12EF.8040007@cs.cmu.edu> Message-ID: <4F904F17.1000108@cs.cmu.edu> > ========================================================= > WE HAVE TWO INTELLIGENCE SEMINAR PRESENTATIONS NEXT WEEK: > PAUL ROSENBLOOM: APRIL 23, 10:30AM, IN GHC 6501 > KEVIN LEYTON-BROWN: APRIL 24, 3:30PM, IN GHC 4303 > THE ANNOUNCEMENT BELOW IS FOR THE SECOND PRESENTATION > ========================================================= > > INTELLIGENCE SEMINAR > APRIL 24 AT 3:30PM, IN GHC 4303 > > SPEAKER: KEVIN LEYTON-BROWN (University of British Columbia) > Host: Ariel Procaccia > For meetings, contact Dana Houston (dhouston at cs.cmu.edu) > > BEYOND EQUILIBRIUM: PREDICTING HUMAN BEHAVIOR IN NORMAL FORM GAMES > > It is standard in multiagent settings to assume that agents will adopt > Nash equilibrium strategies. However, studies in experimental economics > demonstrate that Nash equilibrium is a poor description of human players > initial behavior in normal form games. In this talk, I will describe a > wide > range of widely studied models from behavioral game theory (BGT). For > what > we believe is the first time, we evaluated each of these models in a > meta-analysis, taking as our dataset large-scale and publicly available > experimental data from the BGT literature. We also analyzed the > parameters > of the best performing model, and identified ways of modifying it--and, > indeed, simplifying it--to improve performance. In the end, our work > demonstrates two surprising facts: one BGT model was consistently the > best, and people are smarter than behavioral game theorists had thought. > > BIO > > Kevin Leyton-Brown is an associate professor in computer science at the > University of British Columbia. He holds a Ph.D. and M.S. from Stanford > University (2003; 2001) and a B.S. from McMaster University (1998). Much > of his work is at the intersection of computer science and > microeconomics, > addressing computational problems in economic contexts and incentive > issues in multiagent systems. He also studies the application of machine > learning to the automated design and analysis of algorithms for solving > hard computational problems. He has co-written two books, "Multiagent > Systems" and "Essentials of Game Theory," and over seventy peer-refereed > technical articles. He is the program chair for the ACM Conference on > Electronic Commerce (ACM-EC), and an associate editor of the Journal of > Artificial Intelligence Research (JAIR), the Artificial Intelligence > Journal (AIJ), and ACM Transactions on Economics and Computation. He > split > his 2010-2011 sabbatical between Makerere University in Kampala, Uganda, > and the Institute for Advanced Studies at Hebrew University in Jerusalem, > Israel. He has served as a consultant for Trading Dynamics Inc., Ariba > Inc., and Cariocas Inc., and was a scientific advisor to Zite Inc. until > it was acquired by CNN in 2011. > > -- Dana M. Houston Language Technologies Institute School of Computer Science Carnegie Mellon University 5405 Gates Hillman Complex 5000 Forbes Avenue Pittsburgh, PA 15213 T: (412)268-4717 F: (412)268-6298 From dhouston at cs.cmu.edu Thu Apr 19 13:45:15 2012 From: dhouston at cs.cmu.edu (Dana Houston) Date: Thu, 19 Apr 2012 13:45:15 -0400 Subject: [Intelligence Seminar] April 23, 10:30am:, Presentation by Paul Rosenbloom In-Reply-To: <4F8C125C.9040200@cs.cmu.edu> References: <4F8C125C.9040200@cs.cmu.edu> Message-ID: <4F904F2B.9020705@cs.cmu.edu> ========================================================= > WE HAVE TWO INTELLIGENCE SEMINAR PRESENTATIONS NEXT WEEK: > PAUL ROSENBLOOM: APRIL 23, 10:30AM, IN GHC 6501 > KEVIN LEYTON-BROWN: APRIL 24, 3:30PM, IN GHC 4303 > THE ANNOUNCEMENT BELOW IS FOR THE FIRST PRESENTATION > ========================================================= > > INTELLIGENCE SEMINAR > APRIL 23 AT 10:30AM, IN GHC 6501 > (UNUSUAL DAY, TIME, AND PLACE) > > SPEAKER: PAUL ROSENBLOOM (University of Southern California) > Host: Philip Lehman > For meetings, contact June Fischerkeller (jfische at cs.cmu.edu) > > TOWARDS A GRAPHICAL COGNITIVE ARCHITECTURE FOR VIRTUAL HUMANS (AND > INTELLIGENT AGENTS/ROBOTS) > > A cognitive architecture provides a hypothesis about the fixed > structure (and its integration) underlying intelligent behavior, > whether in natural or artificial systems. The overall goal of this > effort is to leverage graphical models with their ability to uniformly > yield state-of-the-art algorithms across symbol, probability, and > signal processing in developing a new architecture that goes > significantly beyond today's best in providing, and tightly > integrating together, the capabilities required for virtual humans > (and intelligent agents/robots). The current focus is on a graphical > mixed (i.e., statistical relational) architecture that supports hybrid > processing through its grounding in a continuous representation. > Aspects of memory, problem solving, perception, imagery, learning, and > natural language have been demonstrated to date in this architecture, > although some are still mere beginnings. The talk will introduce the > desiderata for this graphical architecture, explain the basics of its > operation, and highlight progress on some of these capabilities. > > BIO > > Paul S. Rosenbloom is a Professor of Computer Science at the > University of Southern California (USC) and a Project Leader at USC's > Institute for Creative Technologies. He spent twenty years at USC's > Information Sciences Institute, including a decade leading new > directions and a stint as Deputy Director. Earlier he was an Assistant > Professor of Computer Science and Psychology at Stanford University, > and a Research Computer Scientist at Carnegie Mellon University. He > received his B.S. in Mathematical Sciences (with distinction) from > Stanford University and his M.S. and Ph.D. in Computer Science from > Carnegie Mellon University. He is a Fellow of the Association for the > Advancement of Artificial Intelligence (AAAI). Prof. Rosenbloom's > research focuses on cognitive architectures; he was a co-PI of the > Soar Project for fifteen years, and is currently developing a new > approach based on graphical models. He has also been working to > understand the nature and structure of computing as a scientific > domain. > > > -- Dana M. Houston Language Technologies Institute School of Computer Science Carnegie Mellon University 5405 Gates Hillman Complex 5000 Forbes Avenue Pittsburgh, PA 15213 T: (412)268-4717 F: (412)268-6298 From dhouston at cs.cmu.edu Mon Apr 23 08:51:41 2012 From: dhouston at cs.cmu.edu (Dana Houston) Date: Mon, 23 Apr 2012 08:51:41 -0400 Subject: [Intelligence Seminar] April 23, 10:30am:, Presentation by Paul Rosenbloom In-Reply-To: <4F904F2B.9020705@cs.cmu.edu> References: <4F8C125C.9040200@cs.cmu.edu> <4F904F2B.9020705@cs.cmu.edu> Message-ID: <4F95505D.7010707@cs.cmu.edu> > > ========================================================= >> WE HAVE TWO INTELLIGENCE SEMINAR PRESENTATIONS NEXT WEEK: >> PAUL ROSENBLOOM: APRIL 23, 10:30AM, IN GHC 6501 >> KEVIN LEYTON-BROWN: APRIL 24, 3:30PM, IN GHC 4303 >> THE ANNOUNCEMENT BELOW IS FOR THE FIRST PRESENTATION >> ========================================================= >> >> INTELLIGENCE SEMINAR >> APRIL 23 AT 10:30AM, IN GHC 6501 >> (UNUSUAL DAY, TIME, AND PLACE) >> >> SPEAKER: PAUL ROSENBLOOM (University of Southern California) >> Host: Philip Lehman >> For meetings, contact June Fischerkeller (jfische at cs.cmu.edu) >> >> TOWARDS A GRAPHICAL COGNITIVE ARCHITECTURE FOR VIRTUAL HUMANS (AND >> INTELLIGENT AGENTS/ROBOTS) >> >> A cognitive architecture provides a hypothesis about the fixed >> structure (and its integration) underlying intelligent behavior, >> whether in natural or artificial systems. The overall goal of this >> effort is to leverage graphical models with their ability to uniformly >> yield state-of-the-art algorithms across symbol, probability, and >> signal processing in developing a new architecture that goes >> significantly beyond today's best in providing, and tightly >> integrating together, the capabilities required for virtual humans >> (and intelligent agents/robots). The current focus is on a graphical >> mixed (i.e., statistical relational) architecture that supports hybrid >> processing through its grounding in a continuous representation. >> Aspects of memory, problem solving, perception, imagery, learning, and >> natural language have been demonstrated to date in this architecture, >> although some are still mere beginnings. The talk will introduce the >> desiderata for this graphical architecture, explain the basics of its >> operation, and highlight progress on some of these capabilities. >> >> BIO >> >> Paul S. Rosenbloom is a Professor of Computer Science at the >> University of Southern California (USC) and a Project Leader at USC's >> Institute for Creative Technologies. He spent twenty years at USC's >> Information Sciences Institute, including a decade leading new >> directions and a stint as Deputy Director. Earlier he was an Assistant >> Professor of Computer Science and Psychology at Stanford University, >> and a Research Computer Scientist at Carnegie Mellon University. He >> received his B.S. in Mathematical Sciences (with distinction) from >> Stanford University and his M.S. and Ph.D. in Computer Science from >> Carnegie Mellon University. He is a Fellow of the Association for the >> Advancement of Artificial Intelligence (AAAI). Prof. Rosenbloom's >> research focuses on cognitive architectures; he was a co-PI of the >> Soar Project for fifteen years, and is currently developing a new >> approach based on graphical models. He has also been working to >> understand the nature and structure of computing as a scientific >> domain. >> >> >> > -- Dana M. Houston Language Technologies Institute School of Computer Science Carnegie Mellon University 5405 Gates Hillman Complex 5000 Forbes Avenue Pittsburgh, PA 15213 T: (412)268-4717 F: (412)268-6298 From dhouston at cs.cmu.edu Mon Apr 23 08:51:53 2012 From: dhouston at cs.cmu.edu (Dana Houston) Date: Mon, 23 Apr 2012 08:51:53 -0400 Subject: [Intelligence Seminar] April 24, 3:30pm:, Presentation by Kevin Leyton-Brown In-Reply-To: <4F904F17.1000108@cs.cmu.edu> References: <4F8C12EF.8040007@cs.cmu.edu> <4F904F17.1000108@cs.cmu.edu> Message-ID: <4F955069.2070900@cs.cmu.edu> > > > >> ========================================================= >> WE HAVE TWO INTELLIGENCE SEMINAR PRESENTATIONS NEXT WEEK: >> PAUL ROSENBLOOM: APRIL 23, 10:30AM, IN GHC 6501 >> KEVIN LEYTON-BROWN: APRIL 24, 3:30PM, IN GHC 4303 >> THE ANNOUNCEMENT BELOW IS FOR THE SECOND PRESENTATION >> ========================================================= >> >> INTELLIGENCE SEMINAR >> APRIL 24 AT 3:30PM, IN GHC 4303 >> >> SPEAKER: KEVIN LEYTON-BROWN (University of British Columbia) >> Host: Ariel Procaccia >> For meetings, contact Dana Houston (dhouston at cs.cmu.edu) >> >> BEYOND EQUILIBRIUM: PREDICTING HUMAN BEHAVIOR IN NORMAL FORM GAMES >> >> It is standard in multiagent settings to assume that agents will adopt >> Nash equilibrium strategies. However, studies in experimental economics >> demonstrate that Nash equilibrium is a poor description of human players >> initial behavior in normal form games. In this talk, I will describe a >> wide >> range of widely studied models from behavioral game theory (BGT). For >> what >> we believe is the first time, we evaluated each of these models in a >> meta-analysis, taking as our dataset large-scale and publicly available >> experimental data from the BGT literature. We also analyzed the >> parameters >> of the best performing model, and identified ways of modifying it--and, >> indeed, simplifying it--to improve performance. In the end, our work >> demonstrates two surprising facts: one BGT model was consistently the >> best, and people are smarter than behavioral game theorists had thought. >> >> BIO >> >> Kevin Leyton-Brown is an associate professor in computer science at the >> University of British Columbia. He holds a Ph.D. and M.S. from Stanford >> University (2003; 2001) and a B.S. from McMaster University (1998). Much >> of his work is at the intersection of computer science and >> microeconomics, >> addressing computational problems in economic contexts and incentive >> issues in multiagent systems. He also studies the application of machine >> learning to the automated design and analysis of algorithms for solving >> hard computational problems. He has co-written two books, "Multiagent >> Systems" and "Essentials of Game Theory," and over seventy peer-refereed >> technical articles. He is the program chair for the ACM Conference on >> Electronic Commerce (ACM-EC), and an associate editor of the Journal of >> Artificial Intelligence Research (JAIR), the Artificial Intelligence >> Journal (AIJ), and ACM Transactions on Economics and Computation. He >> split >> his 2010-2011 sabbatical between Makerere University in Kampala, Uganda, >> and the Institute for Advanced Studies at Hebrew University in >> Jerusalem, >> Israel. He has served as a consultant for Trading Dynamics Inc., Ariba >> Inc., and Cariocas Inc., and was a scientific advisor to Zite Inc. until >> it was acquired by CNN in 2011. >> >> > -- Dana M. Houston Language Technologies Institute School of Computer Science Carnegie Mellon University 5405 Gates Hillman Complex 5000 Forbes Avenue Pittsburgh, PA 15213 T: (412)268-4717 F: (412)268-6298 From dhouston at cs.cmu.edu Mon Sep 24 09:45:52 2012 From: dhouston at cs.cmu.edu (Dana Houston) Date: Mon, 24 Sep 2012 09:45:52 -0400 Subject: [Intelligence Seminar] Sept. 28, 2:00pm: CMACS Seminar, , Presentation by Thomas Forster Message-ID: <50606410.1030902@cs.cmu.edu> ==================================================== THIS ANNOUNCEMENT IS FOR CMACS SEMINAR, WHICH MAY BE OF INTEREST TO THE INTELLIGENCE SEMINAR AUDIENCE ==================================================== CMACS SEMINAR (http://cmacs.cs.cmu.edu/seminars/) SEPTEMBER 28 AT 2:00PM, IN GHC 8102 SPEAKER: THOMAS FORSTER (http://www.dpmms.cam.ac.uk/~tf/ ) Department of Pure Mathematics and Mathematical Statistics University of Cambridge, United Kingdom For meetings, contact Charlotte Yano (yano at cs.cmu.edu ) Meetings are available from September 24 to October 4 REPRESENTING CONTINUOUS GAMES AS DISCRETE GAMES Typical examples of continuous games are pursuer-evader games: lion-antelope, homicidal chauffeur, etc. The players do not take turns to make moves, and they can make their moves at any time. Discrete (combinatorial) games are exemplified by chess, go, etc. These two classes of games enjoy two completely disjoint mathematical treatments. In this talk I illustrate how any pursuit-evader game can be represented as a discrete game - albeit one of imperfect information. BIO Thomas Forster did a first degree in Philosophy and Music but his Ph.D. was in Mathematical Logic (Cambridge 1977), on Quine's Set Theory NF. He has spent most of his working life in Cambridge, though he was a Pittsburgh Centre Fellow in 2003. Now semiretired, he still lectures Part III in Cambridge, and supervises undergraduates and Ph.D. students, but travels more than hitherto. -- Dana M. Houston Language Technologies Institute School of Computer Science Carnegie Mellon University 5405 Gates Hillman Complex 5000 Forbes Avenue Pittsburgh, PA 15213 T: (412)268-4717 F: (412)268-6298 -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mailman.srv.cs.cmu.edu/pipermail/intelligence-seminar-announce/attachments/20120924/1a750935/attachment-0001.html From tom.mitchell at cs.cmu.edu Sun Oct 21 23:01:26 2012 From: tom.mitchell at cs.cmu.edu (Tom Mitchell) Date: Sun, 21 Oct 2012 23:01:26 -0400 Subject: [Intelligence Seminar] AI-related Seminar on "Software Individuals" by Erik Sandewall - Oct. 24 Message-ID: This seminar by Erik Sandewall, former Editor in Chief of the AI Journal, may be of interest to some. Contact Sharon Cavlovich to schedule a meeting with Erik if you like. best Tom ---------- Forwarded message ---------- From: Sharon Cavlovich Date: Fri, Oct 19, 2012 at 11:14 AM Subject: ML Special Seminar - Erik Sandewall - Oct. 24, 2012 To: ml-seminar at cs.cmu.edu, Catherine Copetas Please join us for a ML Special Seminar: Oct. 24, 2012 10:00am Gates Building 8102 Speaker: Erik Sandewall Host: Tom Mitchell, for appointments contact Sharon Cavlovich (sharonw at cs.cmu.edu) Title of talk: Software Individuals Abstract: Software systems with general intelligence need two characteristic properties, among others: an ability to learn, and an ability to accept responsibilities e.g. as the result of delegation. I propose that they should therefore be organized as *software individuals* which are cognitive intelligent agents that are designed with two properties of a technical character: they are designed for long life, to the order of years or more, and they are also designed with individuality. The latter property means that moving an individual and making a copy of it are essentially different operations. Longevity is a prerequisite for meaningful learning; individuality is a prerequisite for delegation and responsibility; and it turns out that the technical implementation of these characteristics are interrelated. In the talk I will discuss these issues and describe the Leonardo system which is a software platform for cognitive intelligent agents that has the described properties. Bio: Erik Sandewall is professor of Computer Science at Link?ping University in Sweden, and he has also worked as visiting researcher at Stanford, MIT, and the University of Toulouse in France. Active in AI research since the mid 1960's, he has combined work in Knowledge Representation (reasoning about actions, and defeasible inheritance) with applied projects concerning autonomous intelligent vehicles (cars and UAV's). He founded the Electronic Transactions on Artificial Intelligence which featured open peer review, and he has served as co-Editor-in-chief of the journal Artificial Intelligence. -- ~~~~~~~~~~~~~~~~~~~~~~~~~~~ Sharon Cavlovich Assistant to the Department Head Machine Learning Department Carnegie Mellon University 5000 Forbes Avenue Gates Hillman Complex 8215 Pittsburgh, PA 15213 (412)268-5196 http://www.cs.cmu.edu/~sharonw From arielpro at cs.cmu.edu Thu Oct 25 21:51:08 2012 From: arielpro at cs.cmu.edu (Ariel Procaccia) Date: Thu, 25 Oct 2012 21:51:08 -0400 Subject: [Intelligence Seminar] Joint algorithmic economics + intelligence seminar: Milind Tambe, Nov 6 @ noon, GHC 6115 Message-ID: Hi everyone, Milind Tambe, Prof. of CS at USC, will give a talk on Tuesday Nov 6 at noon in GHC 6115, entitled "Security and Game Theory: Key Algorithmic Principles, Deployed Applications, Lessons Learned". Milind has been doing fascinating work on this topic, which I view as one of the most exciting applications of algorithmic game theory, and one of the best examples of the usefulness of game theory in general; see also: http://agtb.wordpress.com/2012/06/17/is-game-theory-useful/. Lunch will be served! You can find the talk details (including abstract) on the seminar's website: http://www.cs.cmu.edu/~arielpro/algecon.html. Milind will be available on Nov 6 for meetings. If you would like to meet him, please send your constraints to Pat Loring (cc'd) at sawako at cs.cmu.edu. Best regards, Ariel -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mailman.srv.cs.cmu.edu/pipermail/intelligence-seminar-announce/attachments/20121025/3cc02492/attachment.html From dhouston at cs.cmu.edu Thu Nov 1 11:12:49 2012 From: dhouston at cs.cmu.edu (Dana Houston) Date: Thu, 01 Nov 2012 11:12:49 -0400 Subject: [Intelligence Seminar] November 6, 12:00 noon: , Presentation by Milind Tambe Message-ID: <50929171.10207@cs.cmu.edu> INTELLIGENCE SEMINAR NOVEMBER 6 AT 12:00 NOON, IN GHC 6115 (PLEASE NOTE THE UNUSUAL ROOM AND TIME) This talk is a joint presentation for Intelligence Seminar and Algorithmic Economics Seminar. SPEAKER: MILIND TAMBE (University of Southern California) Host: Ariel Procaccia For meetings, contact Pat Loring (sawako at cs.cmu.edu ) SECURITY AND GAME THEORY: KEY ALGORITHMIC PRINCIPLES, DEPLOYED APPLICATIONS, LESSONS LEARNED Security is a critical concern around the world, whether it is the challenge of protecting ports, airports, and other critical national infrastructure, or protecting wildlife and forests, or suppressing crime in urban areas. In many of these cases, limited security resources prevent full security coverage at all times. Instead, these limited resources must be scheduled, avoiding schedule predictability, while simultaneously taking into account different target priorities, the responses of the adversaries to the security posture, and potential uncertainty over adversary types. Computational game theory can help design such unpredictable security schedules. Indeed, casting the problem as a Bayesian Stackelberg game, we have developed new algorithms that are now deployed over multiple years in multiple applications for security scheduling: for the US coast guard in Boston and New York (and potentially other ports), for the Federal Air Marshals, for the Los Angeles Airport Police, with the Los Angeles Sheriff's Department for patrolling metro trains, with further applications under evaluation for the TSA and other agencies. These applications are leading to real-world use-inspired research in the emerging research area of security games. Specifically, the research challenges posed by these applications include scaling up security games to large-scale problems, handling significant adversarial uncertainty, dealing with bounded rationality of human adversaries, and other interdisciplinary challenges. This lecture will provide an overview of my research group's work in this area, outlining key algorithmic principles and research results, as well as a discussion of our deployed systems and lessons learned. BIO Milind Tambe is Helen N. and Emmett H. Jones Professor in Engineering, and Professor of Computer Science and Industrial and Systems Engineering at the University of Southern California (USC). He leads the Teamcore Research Group at USC, with research focused on agent-based and multi-agent systems. He is a fellow of AAAI, recipient of the ACM Autonomous Agents Research Award, as well as recipient of the Christopher Columbus Fellowship Foundation Homeland Security Award. In addition, he is also recipient of the "influential paper award" from the International Foundation for Agents and Multiagent Systems, the Rist Prize of the Military Operations Research Society, US Coast Guard First District's Operational Excellence Award, Certificate of Appreciation from the US Federal Air Marshals Service, special commendation given by the Los Angeles World Airports police from the city of Los Angeles, IBM Faculty Award, Okawa Foundation faculty research award, the RoboCup scientific challenge award, USC Viterbi School of Engineering use-inspired research award, USC Steven B. Sample Teaching and Mentoring award, and the ACM recognition of service award. Prof. Tambe and his research group's papers have been selected as best papers at a number of premier artificial intelligence conferences and workshops; these have included best paper awards at the International Conference on Autonomous Agents and Multiagent Systems and International Conference on Intelligent Virtual Agents. Additionally, algorithms developed by his Teamcore research group have been deployed for real-world use by several agencies including the US Coast Guard, the US Federal Air Marshals service, LAX Police, and the LA Sheriff's Department. He received his Ph.D. from the School of Computer Science at Carnegie Mellon University. -- Dana M. Houston Language Technologies Institute School of Computer Science Carnegie Mellon University 5405 Gates Hillman Complex 5000 Forbes Avenue Pittsburgh, PA 15213 T: (412)268-4717 F: (412)268-6298 -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mailman.srv.cs.cmu.edu/pipermail/intelligence-seminar-announce/attachments/20121101/9df9fe36/attachment.html From dhouston at cs.cmu.edu Mon Nov 5 11:45:54 2012 From: dhouston at cs.cmu.edu (Dana Houston) Date: Mon, 05 Nov 2012 11:45:54 -0500 Subject: [Intelligence Seminar] Joint OM/OR Seminar - Marek Petrik Message-ID: <5097ED42.90109@cs.cmu.edu> ==================================================== THIS ANNOUNCEMENT IS FOR OM/OR SEMINAR, WHICH MAY BE OF INTEREST TO THE INTELLIGENCE SEMINAR AUDIENCE ==================================================== OR/OMM Faculty and PHDs The following joint Operations Research and Operations Management and Manufacturing Seminar has been posted to the Tepper School seminar tracker. The schedule can be found under Manufacturing. There are slots left to fill, please take the time to complete the schedule. Name: Marek Petrik University: IBM T.J. Watson Research Center Date: Friday, November 16, 2012 Time: 10:30 a.m. - 12:00 p.m. Location: GSIA Faculty Conference Room 322 Title: Approximate Dynamic Programming By Minimizing Distributionally Robust Bounds Abstract: Approximate dynamic programming is a popular method for solving large Markov decision processes. I will describe a new class of approximate dynamic programming (ADP) methods---distributionally robust ADP---that address the curse of dimensionality by minimizing a pessimistic bound on the policy return. This approach turns ADP into an optimization problem, for which we derive new mathematical program formulations and analyze its properties. DRADP improves on the theoretical guarantees of existing ADP methods---it guarantees convergence and L1 norm-based error bounds. The empirical evaluation of DRADP shows that the theoretical guarantees translate well into good performance on benchmark problems. -- Dana M. Houston Language Technologies Institute School of Computer Science Carnegie Mellon University 5405 Gates Hillman Complex 5000 Forbes Avenue Pittsburgh, PA 15213 T: (412)268-4717 F: (412)268-6298 -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mailman.srv.cs.cmu.edu/pipermail/intelligence-seminar-announce/attachments/20121105/f1c7bb6d/attachment-0001.html From dhouston at cs.cmu.edu Mon Nov 5 13:54:10 2012 From: dhouston at cs.cmu.edu (Dana Houston) Date: Mon, 05 Nov 2012 13:54:10 -0500 Subject: [Intelligence Seminar] November 6, 12:00 noon: , Presentation by Milind Tambe Message-ID: <50980B52.6070803@cs.cmu.edu> INTELLIGENCE SEMINAR NOVEMBER 6 AT 12:00 NOON, IN GHC 6115 (PLEASE NOTE THE UNUSUAL ROOM AND TIME) This talk is a joint presentation for Intelligence Seminar and Algorithmic Economics Seminar. SPEAKER: MILIND TAMBE (University of Southern California) Host: Ariel Procaccia For meetings, contact Pat Loring (sawako at cs.cmu.edu ) SECURITY AND GAME THEORY: KEY ALGORITHMIC PRINCIPLES, DEPLOYED APPLICATIONS, LESSONS LEARNED Security is a critical concern around the world, whether it is the challenge of protecting ports, airports, and other critical national infrastructure, or protecting wildlife and forests, or suppressing crime in urban areas. In many of these cases, limited security resources prevent full security coverage at all times. Instead, these limited resources must be scheduled, avoiding schedule predictability, while simultaneously taking into account different target priorities, the responses of the adversaries to the security posture, and potential uncertainty over adversary types. Computational game theory can help design such unpredictable security schedules. Indeed, casting the problem as a Bayesian Stackelberg game, we have developed new algorithms that are now deployed over multiple years in multiple applications for security scheduling: for the US coast guard in Boston and New York (and potentially other ports), for the Federal Air Marshals, for the Los Angeles Airport Police, with the Los Angeles Sheriff's Department for patrolling metro trains, with further applications under evaluation for the TSA and other agencies. These applications are leading to real-world use-inspired research in the emerging research area of security games. Specifically, the research challenges posed by these applications include scaling up security games to large-scale problems, handling significant adversarial uncertainty, dealing with bounded rationality of human adversaries, and other interdisciplinary challenges. This lecture will provide an overview of my research group's work in this area, outlining key algorithmic principles and research results, as well as a discussion of our deployed systems and lessons learned. BIO Milind Tambe is Helen N. and Emmett H. Jones Professor in Engineering, and Professor of Computer Science and Industrial and Systems Engineering at the University of Southern California (USC). He leads the Teamcore Research Group at USC, with research focused on agent-based and multi-agent systems. He is a fellow of AAAI, recipient of the ACM Autonomous Agents Research Award, as well as recipient of the Christopher Columbus Fellowship Foundation Homeland Security Award. In addition, he is also recipient of the "influential paper award" from the International Foundation for Agents and Multiagent Systems, the Rist Prize of the Military Operations Research Society, US Coast Guard First District's Operational Excellence Award, Certificate of Appreciation from the US Federal Air Marshals Service, special commendation given by the Los Angeles World Airports police from the city of Los Angeles, IBM Faculty Award, Okawa Foundation faculty research award, the RoboCup scientific challenge award, USC Viterbi School of Engineering use-inspired research award, USC Steven B. Sample Teaching and Mentoring award, and the ACM recognition of service award. Prof. Tambe and his research group's papers have been selected as best papers at a number of premier artificial intelligence conferences and workshops; these have included best paper awards at the International Conference on Autonomous Agents and Multiagent Systems and International Conference on Intelligent Virtual Agents. Additionally, algorithms developed by his Teamcore research group have been deployed for real-world use by several agencies including the US Coast Guard, the US Federal Air Marshals service, LAX Police, and the LA Sheriff's Department. He received his Ph.D. from the School of Computer Science at Carnegie Mellon University. -- Dana M. Houston Language Technologies Institute School of Computer Science Carnegie Mellon University 5405 Gates Hillman Complex 5000 Forbes Avenue Pittsburgh, PA 15213 T: (412)268-4717 F: (412)268-6298 -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mailman.srv.cs.cmu.edu/pipermail/intelligence-seminar-announce/attachments/20121105/d3888fd8/attachment.html From dhouston at cs.cmu.edu Thu Nov 8 12:01:01 2012 From: dhouston at cs.cmu.edu (Dana Houston) Date: Thu, 08 Nov 2012 12:01:01 -0500 Subject: [Intelligence Seminar] Mailing List Maintenance Message-ID: <509BE54D.8030709@cs.cmu.edu> List members, If you are no longer at CMU, and would like to be removed from this list, please reply. I will process your request as soon as possible. Thank you, Dana -- Dana M. Houston Language Technologies Institute School of Computer Science Carnegie Mellon University 5405 Gates Hillman Complex 5000 Forbes Avenue Pittsburgh, PA 15213 T: (412)268-4717 F: (412)268-6298 From dhouston at cs.cmu.edu Mon Nov 26 16:02:31 2012 From: dhouston at cs.cmu.edu (Dana Houston) Date: Mon, 26 Nov 2012 16:02:31 -0500 Subject: [Intelligence Seminar] Algorithmic economics seminar: David Parkes, Nov 27, 12 noon, NSH 3305 Message-ID: <50B3D8E7.1010201@cs.cmu.edu> ==================================================== THIS ANNOUNCEMENT IS FOR ALGORITHMIC ECONOMICS SEMINAR, WHICH MAY BE OF INTEREST TO THE INTELLIGENCE SEMINAR AUDIENCE ==================================================== ALGORITHMIC ECONOMICS SEMINAR NOVEMBER 27, AT 12:00 NOON, IN NSH 3305 David Parkes, Prof. of CS at Harvard, will give a talk on Tuesday Nov 27 at noon in NSH 3305, entitled "Mechanism Design as a Classification Problem". The award-winning paper on which the talk is based makes a surprising (and rather subtle) connection between machine learning and mechanism design. At CMU there is recently a surge of interest in the intersection of ML and the social sciences (see https://sites.google.com/site/mlsscmu/), so David's choice of topic seems especially timely. You can find the talk details (including abstract) on the seminar's website: http://www.cs.cmu.edu/~arielpro/algecon.html . David will be available on Nov 27 for meetings. If you would like to meet him, please send your constraints to Pat Loring at sawako at cs.cmu.edu . -- Dana M. Houston Language Technologies Institute School of Computer Science Carnegie Mellon University 5405 Gates Hillman Complex 5000 Forbes Avenue Pittsburgh, PA 15213 T: (412)268-4717 F: (412)268-6298 -------------- next part -------------- An HTML attachment was scrubbed... URL: