From awd at cs.cmu.edu Wed Feb 1 11:59:33 2006 From: awd at cs.cmu.edu (Artur Dubrawski) Date: Wed, 01 Feb 2006 11:59:33 -0500 Subject: [Research] Yet another correction! Meeting on Thursday will be in NSH 1507 (the usual place) In-Reply-To: <1138723755.1791.51.camel@logo.auton.cs.cmu.edu> References: <1138722831.1791.27.camel@logo.auton.cs.cmu.edu> <1138723755.1791.51.camel@logo.auton.cs.cmu.edu> Message-ID: <1138813173.1791.635.camel@logo.auton.cs.cmu.edu> Where: NSH 1507 Other stuff remains unchanged. Artur On Tue, 2006-01-31 at 11:09, Artur Dubrawski wrote: > Correction! > > I've just received a bid from a volunteer who wants to > give a brief talk followed by a brainstorming session > this Thursday. So, after all, we will have a lab meeting > this week. > > Speaker: Kaustav Das > When: Thursday, February 2nd, 2006, 4:30pm > Where: NSH 3305 *NOTE THE UNUSUAL PLACE* > Food: Yes. > > Artur > > On Tue, 2006-01-31 at 10:53, Artur Dubrawski wrote: > > Hello, > > > > Though there won't be a regular lab meeting this week, > > I encourage everyone to attend Robin's CALD KDD project talk > > to be held today at 2pm in NSH 3305. It will be about > > T-Cubes and it will be much refined version of his latest > > lab meeting performance (our logo will be on the slides > > this time I was assured :-). > > Please come and support Robin if you can. > > > > Artur > > > > _______________________________________________ > > Research mailing list > > Research at autonlab.org > > https://www.autonlab.org/mailman/listinfo/research > > From awd at cs.cmu.edu Mon Feb 6 17:43:20 2006 From: awd at cs.cmu.edu (Artur Dubrawski) Date: Mon, 06 Feb 2006 17:43:20 -0500 Subject: [Research] Fwd: JRSA Student Presenation Contest In-Reply-To: <5E24DACD44122FC98C43D0DE@heinz-ab0q.heinz.win.cmu.edu> References: <5E24DACD44122FC98C43D0DE@heinz-ab0q.heinz.win.cmu.edu> Message-ID: <1139265800.17978.263.camel@logo.auton.cs.cmu.edu> To Auton Lab students working on anti-terrorism-related stuff: This may be a chance to increase visibility of your work, involving a visit Denver, CO. > > Justice Research and Statistics Association > > > > > 2006 Student Presentation Contest > > > > JRSA is pleased to announce the Third Annual Student Presentation Contest > in conjunction with the 2006 Bureau of Justice Statistics/Justice Research > and Statistics Association National Conference. The winner will have all > expenses paid to the conference, which will be held October 12?13 in > Denver, Colorado, and will present research findings as part of a panel > session. > > > > Entry criteria: > * > To be eligible, an entrant must be a graduate student in an accredited > university program in the United States. > * > Only one entry per student is allowed. > * > Work submitted must be applied research that was conducted during 2005 > and/or 2006 toward fulfillment of a graduate degree. > * > The research must fall within a criminal justice area such as corrections, > courts, sentencing, civil justice, cybercrime, computer crime, terrorism, > reentry, domestic violence, juvenile justice, drug use, victimization, or > law enforcement. > * > The paper to be presented must not have been published in a professional > journal. > * > The winner must be able to present the paper in person at the BJS/JRSA 2006 > conference, and is expected to participate in the entire two-day conference. > > Entry guidelines: > * > A letter of intent to submit a paper, including the topic, should be > submitted by Friday, March 31, to studentcontest at jrsa.org. > * > Papers must be submitted electronically by 5:00 p.m. eastern time on > Wednesday, May 31, to studentcontest at jrsa.org. > * > Papers must not exceed 25 double-spaced pages in 12-pt. font, exclusive of > references and appendices. > * > A 50-100 word abstract must be provided. > * > The cover page must include: the author?s name (if there are multiple > authors, all must be graduate students, only one may enter the contest and > that name must be clearly identified); contact information, including email > address; the name of the graduate school; the time period during which the > research was carried out; and the date the paper was completed. If the > paper has been presented at other conferences, indicate the conference and > date. > > > > Submissions will be rated by an independent panel of reviewers. > Determination of the winner will be made by JRSA and announced in August. > JRSA reserves the right not to award a prize if no entry is deemed > qualified. All decisions will be considered final. > > > > Staff of JRSA, BJS, and the state Statistical Analysis Centers are > ineligible to apply. Questions should be directed to > studentcontest at jrsa.org. > > > > The Justice Research and Statistics Association is a national nonprofit > organization. For more information about JRSA, see our Web site at > www.jrsa.org. > > ---------- End Forwarded Message ---------- > > > From awd at cs.cmu.edu Tue Feb 7 09:42:17 2006 From: awd at cs.cmu.edu (Artur Dubrawski) Date: Tue, 07 Feb 2006 09:42:17 -0500 Subject: [Research] the upcoming lab meetings Message-ID: <1139323337.19629.16.camel@logo.auton.cs.cmu.edu> Dear Autonians, Here's the schedule for the next two Thursdays: February 9th, 4:30pm, NSH 1507 Daniel Neill Disease Surveillance by Detection of Anomalous Spatial Clusters February 16th, 4:30pm, NSH 1507 Ashwin Tengli Auton Martini: Take ASL, Classifiers, Sparse Data, add Algorithmic & Representational Speedups. Serve shaken, not stirred. Artur PS Please contact Mike Baysek if you could not attend any of those in person. Lawrenceville, Downtown, California, NYC or Wean Hall: it does not matter how exotic is the place where you are. Mike has the power to bring you back for the meeting. From awd at cs.cmu.edu Thu Feb 9 22:36:05 2006 From: awd at cs.cmu.edu (Artur Dubrawski) Date: Thu, 9 Feb 2006 22:36:05 -0500 Subject: [Research] food for lab meetings Message-ID: <000301c62df3$206635b0$0200000a@scs.ad.cs.cmu.edu> Dear Autonians, We significantly overshoot the volumes of the lab meeting food orders. Would you please kindly inform Kristen of your planned absences so that she could plan accordingly. Thanks! One could tell it's late because I have another crazy idea :-) How about if we had "a person on duty" whose duty would be to take care of leftovers and cleanup of the tables after the meeting? If I do not hear vetoes, let us initiate shifts in the alphabetical order (by last name) starting next week with Brigham. He would be followed by Brent and then Patrick and so on. Drew could take care of the first meeting to actually happen on Tuesday. And I think it would be fair to spare Mike from this as he is already busy setting up and down the videoconferencing stuff at every meeting. Does that make sense? Artur PS. This is the list of people who according to my records are frequenting our meetings. Please let me and Kristen know if I forgot someone (or if there is a person too many :-) and also if you have any specific constraints that we could benefit from knowing when we order food in the future: Brigham Anderson Drew Bagnell (not attending on Thursdays) Mike Baysek Brent Bryan Patrick Choi Karen Chen Kaustav Das Artur Dubrawski Khalid El-Arini Anna Goldenberg Adam Goode Geoff Gordon Jeremy Kubica Joey Liang Ting Liu Daniel Neill John Ostlund Robin Sabhnani Purna Sarkar Sajid Siddiqi Ajit Singh Ashwin Tengli Alice Zheng From qsliang at cmu.edu Fri Feb 10 09:34:27 2006 From: qsliang at cmu.edu (Joey (QuanSheng) Liang) Date: Fri, 10 Feb 2006 09:34:27 -0500 Subject: [Research] food for lab meetings In-Reply-To: <000301c62df3$206635b0$0200000a@scs.ad.cs.cmu.edu> References: <000301c62df3$206635b0$0200000a@scs.ad.cs.cmu.edu> Message-ID: <43ECA473.2050400@cmu.edu> I like the idea of shifting. We need to make sure the person, who will be on duty, will get the notice from either Kristen or the last person on duty; it's hard to remember the sequence. If the person has the difficulty, just passes/returns the token to the next or to Kristen. (personal thought) Artur Dubrawski wrote: > Dear Autonians, > > We significantly overshoot the volumes of the lab meeting food orders. > Would you please kindly inform Kristen of your planned absences > so that she could plan accordingly. Thanks! > > One could tell it's late because I have another crazy idea :-) > How about if we had "a person on duty" whose duty would > be to take care of leftovers and cleanup of the tables > after the meeting? If I do not hear vetoes, let us initiate shifts > in the alphabetical order (by last name) starting next week with > Brigham. He would be followed by Brent and then Patrick and so on. > Drew could take care of the first meeting to actually happen on Tuesday. > And I think it would be fair to spare Mike from this as he > is already busy setting up and down the videoconferencing stuff > at every meeting. > Does that make sense? > > Artur > > PS. This is the list of people who according to my records are > frequenting our meetings. Please let me and Kristen know if > I forgot someone (or if there is a person too many :-) and > also if you have any specific constraints that we could benefit > from knowing when we order food in the future: > > Brigham Anderson > Drew Bagnell (not attending on Thursdays) > Mike Baysek > Brent Bryan > Patrick Choi > Karen Chen > Kaustav Das > Artur Dubrawski > Khalid El-Arini > Anna Goldenberg > Adam Goode > Geoff Gordon > Jeremy Kubica > Joey Liang > Ting Liu > Daniel Neill > John Ostlund > Robin Sabhnani > Purna Sarkar > Sajid Siddiqi > Ajit Singh > Ashwin Tengli > Alice Zheng > > _______________________________________________ > Research mailing list > Research at autonlab.org > https://www.autonlab.org/mailman/listinfo/research > -- Joey (QuanSheng) Liang Auton Lab, RI, SCS, Carnegie Mellon University P: http://www.ChemiLab.net/ W: http://www.AutonLab.org/ From awm at cs.cmu.edu Sat Feb 11 13:15:32 2006 From: awm at cs.cmu.edu (Andrew W Moore) Date: Sat, 11 Feb 2006 13:15:32 -0500 Subject: [Research] FW: Summer Internships at Yahoo Message-ID: <016201c62f37$26b990c0$96ce0280@auton.cs.cmu.edu> -----Original Message----- From: Dmitry Pavlov [mailto:pavlovd at ics.uci.edu] Sent: Tuesday, February 07, 2006 2:53 AM To: awm at cs.cmu.edu Subject: Summer Internships Hello Professor Moore, You may not remember me, we briefly met at the UAI conference a while back. I am a big fan of your work, particularly ADTrees. We are looking for qualified summer interns to work with us this summer. If you know of anyone looking for the summer position please forward the following message to them. Also, please feel free to distribute it further, Thanks much in advance Dima Summer Intern, Yahoo! Search Content Analysis Team Location . Santa Clara, CA Essential Responsibilities We are looking for a highly motivated PhD student specializing in machine learning to help develop algorithms and software systems for analyzing web content. You will be part of Yahoo's content analysis team, an advanced development group responsible for web scale document classification, clustering, attribute extraction and matching applications that push the frontiers of text mining. The content analysis team is focused on lexical, syntactic and semantic analysis of text documents. As part of the team you will have access to Yahoo's enormous amount of data including query logs, product description, and web page index. Qualified candidates should be proficient in C/C++ development, have a solid knowledge of machine learning, data mining and information retrieval. A hands-on experience with data clustering, classification, regression, recommender systems, optimization, text mining, natural language processing, analysis, pattern recognition is desired. Required Qualifications/Education . Masters degree in Computer Science or a related field . Knowledge of at least one of the following: text mining, natural language processing, information retrieval, statistical analysis, machine learning. . Experience in: C/C++, Unix operating system and development tools, scripting languages. . Understanding of algorithm efficiency issues. . Excellent communication and problem solving skills. Keywords: data mining, machine learning, text mining, document classification, document clustering, natural language processing, regression analysis, support vector machine, neural network, decision tree, collaborative filter, statistics, information retrieval Please send your resumes to Dr. Dmitry Pavlov at dpavlov at yahoo-inc.com with the subject line "Summer Internship 2006". From awd at cs.cmu.edu Tue Feb 14 16:13:33 2006 From: awd at cs.cmu.edu (Artur Dubrawski) Date: Tue, 14 Feb 2006 16:13:33 -0500 Subject: [Research] SCHEDULE CHANGE Re: the upcoming lab meetings In-Reply-To: <1139323337.19629.16.camel@logo.auton.cs.cmu.edu> References: <1139323337.19629.16.camel@logo.auton.cs.cmu.edu> Message-ID: <1139951613.10290.94.camel@logo.auton.cs.cmu.edu> Hello, There will be no lab meeting this Thursday. Ashwin will present his work on either Tuesday Feb 21 or Tuesday Feb 28 (the final decision will be made in a couple of days from now). In either case, we will use the "regular" Tuesday spot, i.e. 12noon, NSH 1507. Also, there won't be a lab meeting on Thursday Feb 23 due to CALD research day, unless I successfully twist an arm of a friendly industrial visitor, who would come to town for that occasion, and make them talk at the meeting. If that works, please be prepared for a very short notice. But, alas, we do have a regular meeting scheduled for Thursday, March 02. The speaker will be Purna Sarkar and she will talk about incredibly cool things (the exact title will be available soon: the speaker herself will make it known to us in an email). Artur On Tue, 2006-02-07 at 09:42, Artur Dubrawski wrote: > Dear Autonians, > > Here's the schedule for the next two Thursdays: > > February 9th, 4:30pm, NSH 1507 > Daniel Neill > Disease Surveillance by Detection of Anomalous Spatial Clusters > > February 16th, 4:30pm, NSH 1507 > Ashwin Tengli > Auton Martini: Take ASL, Classifiers, Sparse Data, add Algorithmic > & Representational Speedups. Serve shaken, not stirred. > > Artur > > PS Please contact Mike Baysek if you could not attend any of those > in person. Lawrenceville, Downtown, California, NYC or Wean Hall: > it does not matter how exotic is the place where you are. Mike has > the power to bring you back for the meeting. > > > _______________________________________________ > Research mailing list > Research at autonlab.org > https://www.autonlab.org/mailman/listinfo/research > From psarkar at cs.cmu.edu Wed Feb 15 08:18:31 2006 From: psarkar at cs.cmu.edu (Purnamrita Sarkar) Date: Wed, 15 Feb 2006 08:18:31 -0500 Subject: [Research] upcoming lab meetings Message-ID: <003101c63232$50e8de70$ecb90280@sp.cs.cmu.edu> Hello everyone! In the next to next lab-meeting (March 2nd) I will talk about "Euclidean Embedding of Author-Word Co-occurrence Data Through Time" Here goes an abstract of it : We address the problem of embedding entities into Euclidean space over time based on co-occurrence count data. We model the probability of the data given the (hidden) Euclidean coordinates of the entities. This results in a factored state space model with real-valued hidden parent nodes and discrete children. An approximation on the log partition function of the observation model makes it conjugate to the Normal distribution, allowing us to formulate the dynamic model as a Kalman filter. Qualitative results on co-occurrences of authors and words in the NIPS corpus show that our model results in efficient and meaningful embeddings of large datasets. See you ! Purna -------------- next part -------------- An HTML attachment was scrubbed... URL: From awd at cs.cmu.edu Thu Feb 16 10:54:42 2006 From: awd at cs.cmu.edu (Artur Dubrawski) Date: Thu, 16 Feb 2006 10:54:42 -0500 Subject: [Research] Reminder: No Lab meeting today Message-ID: <1140105282.18468.69.camel@logo.auton.cs.cmu.edu> From awd at cs.cmu.edu Thu Feb 16 17:09:33 2006 From: awd at cs.cmu.edu (Artur Dubrawski) Date: Thu, 16 Feb 2006 17:09:33 -0500 Subject: [Research] SCHEDULE CHANGE Re: the upcoming lab meetings In-Reply-To: <1139951613.10290.94.camel@logo.auton.cs.cmu.edu> References: <1139323337.19629.16.camel@logo.auton.cs.cmu.edu> <1139951613.10290.94.camel@logo.auton.cs.cmu.edu> Message-ID: <1140127773.18468.163.camel@logo.auton.cs.cmu.edu> Ashwin will present his work on Tuesday Feb 28 at 12noon in NSH 1507. This will be the first lab meeting counting from now. The same week we will have Purna's presentation on Thursday. Artur On Tue, 2006-02-14 at 16:13, Artur Dubrawski wrote: > Hello, > > There will be no lab meeting this Thursday. > > Ashwin will present his work on either Tuesday Feb 21 > or Tuesday Feb 28 (the final decision will be made in > a couple of days from now). In either case, we will > use the "regular" Tuesday spot, i.e. 12noon, NSH 1507. > > Also, there won't be a lab meeting on Thursday Feb 23 > due to CALD research day, unless I successfully > twist an arm of a friendly industrial visitor, who > would come to town for that occasion, and make them > talk at the meeting. If that works, please be prepared > for a very short notice. > > But, alas, we do have a regular meeting scheduled for > Thursday, March 02. > The speaker will be Purna Sarkar and she will talk > about incredibly cool things (the exact title > will be available soon: the speaker herself > will make it known to us in an email). > > Artur > > On Tue, 2006-02-07 at 09:42, Artur Dubrawski wrote: > > Dear Autonians, > > > > Here's the schedule for the next two Thursdays: > > > > February 9th, 4:30pm, NSH 1507 > > Daniel Neill > > Disease Surveillance by Detection of Anomalous Spatial Clusters > > > > February 16th, 4:30pm, NSH 1507 > > Ashwin Tengli > > Auton Martini: Take ASL, Classifiers, Sparse Data, add Algorithmic > > & Representational Speedups. Serve shaken, not stirred. > > > > Artur > > > > PS Please contact Mike Baysek if you could not attend any of those > > in person. Lawrenceville, Downtown, California, NYC or Wean Hall: > > it does not matter how exotic is the place where you are. Mike has > > the power to bring you back for the meeting. > > > > > > _______________________________________________ > > Research mailing list > > Research at autonlab.org > > https://www.autonlab.org/mailman/listinfo/research > > > > _______________________________________________ > Research mailing list > Research at autonlab.org > https://www.autonlab.org/mailman/listinfo/research > From awm at cs.cmu.edu Sat Feb 18 13:43:00 2006 From: awm at cs.cmu.edu (Andrew W Moore) Date: Sat, 18 Feb 2006 13:43:00 -0500 Subject: [Research] FW: Ting and collaboration Message-ID: <000901c634bb$24a9bbf0$96ce0280@auton.cs.cmu.edu> Another internship opportunity. FX == Fujitsu+Xerox. Matt is a nice guy. -----Original Message----- From: Matthew Cooper [mailto:cooper at fxpal.com] Sent: Wednesday, February 15, 2006 4:49 PM To: Andrew W Moore Subject: Re: Ting and collaboration Hi Andrew, I hope you are doing well. As a reminder, I worked with Ting Liu in the summer of 2004 at FX Palo Alto Labs. I am again in the position of seeking a summer intern, and after meeting with Jeremy Kubica who interviewed last week here, I was reminded of the how impressive the students in your group are. I intend to work on a project this summer to automatically annotate video data, again in the context of the large scale TRECVID evaluation. We are looking for someone mainly to work on statistical modeling and classification, probably using conditional random fields, to classify video shots as exhibiting (or not) a semantic attribute. As in Ting's case, we are mainly looking for help with the statistical parts of the project, the low-level data processing will be complete before the summer so experience with video data is not required. If you can recommend any students, or other faculty that we may wish to contact, I'd very much appreciate it. I understand from Ting that you're starting a new collaboration with Google. Congratulations. Perhaps we can meet some time if you're out to the west coast. Thanks for your help and consideration. Matt Cooper ________________________________________________________ Matthew Cooper cooper at fxpal.com Andrew W Moore wrote: > Matt, > > Thanks for the message. I'd love to meet up. I've just emailed the > folks organizing my visit to ask whether there's time on their schedule. > > Thanks, > > Andrew > From qsliang at cmu.edu Tue Feb 21 10:12:31 2006 From: qsliang at cmu.edu (Joey Quansheng Liang) Date: Tue, 21 Feb 2006 10:12:31 -0500 Subject: [Research] Changes of h/utils and h/dset Message-ID: <43FB2DDF.4070504@cmu.edu> 1. Added a new function in h/utils/am_string.h(c): /** Convert a string to double value, if possible input: const char *: The target null-terminated string output: double *: The target address holding the double value The value will not be changed if function returns 'false' return: bool: 'true' if succeed, or else 'false' Example of converting: "213.21" => 213.21 "$212,212" => 212212 "($23,323,212)" => -2.33232e+007 "(843" is not a number, return 'false' " -$54,432" => -54432 "34,234%" => 342.34 "2.34e-1" => 0.234 "2.34E1.2" => 37.0865 "50/3" => 16.6667 Some combinations are not allowed, like: "(34.21E2)" : Neither a scientific nor accounting notation "3.1415926/25" : Not a fraction but a formula Some combinations are allowed although it may not be so meaningful, like: "$3278%" => 32.78 */ bool string_to_double(const char *s, double *pValue); 2. The datset realization now uses the function above to convert symbols to doubles. So many new formatted symbols can be realized now. For my curiousness, I did a performance test (Linux, debug): 1. Generated a 10,000 rows by 20 cols datset with random numbers; 2. Forcedly loaded the datset all as symbolic; 3. Made a copy of the loaded datset; 4. Realized all 20 columns by using old and new realization code, for each datset; The old code took 7 seconds, the new code took 6 seconds. That's great! Let me know if there is any question. -- Joey (QuanSheng) Liang Auton Lab, RI, SCS, Carnegie Mellon University W: http://www.AutonLab.org/ P: http://www.ChemiLab.net/ From awd at cs.cmu.edu Tue Feb 21 11:18:11 2006 From: awd at cs.cmu.edu (Artur Dubrawski) Date: Tue, 21 Feb 2006 11:18:11 -0500 Subject: [Research] Reminder: no lab meetings this week Message-ID: <1140538691.9376.119.camel@logo.auton.cs.cmu.edu> From awd at cs.cmu.edu Tue Feb 21 17:33:19 2006 From: awd at cs.cmu.edu (Artur Dubrawski) Date: Tue, 21 Feb 2006 17:33:19 -0500 Subject: [Research] matlab Message-ID: <001f01c63736$d09a69a0$9ee6ed80@scs.ad.cs.cmu.edu> Hello, Does anyone need or anticipates a need to use matlab in the lab environment now or in the near future? Mike and I are thinking about not renewing our current licenses so please speak up if you'd like to stop us from doing so. Thanks Artur From awd at cs.cmu.edu Wed Feb 22 10:44:17 2006 From: awd at cs.cmu.edu (Artur Dubrawski) Date: Wed, 22 Feb 2006 10:44:17 -0500 Subject: [Research] there WILL be a lab meeting tomorrow! Message-ID: <1140623057.19356.57.camel@logo.auton.cs.cmu.edu> Hello, I am sorry for the late notice, but in fact we will have a lab meeting this Thursday and a guest speaker who, without much arm-twisting, agreed to talk about the topics explained below. Kristen: please make sure we have room available and let everyone know if it won't be the usual NSH 1507. And please get us some food :-) Artur ----- Piero Bonissone, General Electric Global Research "Knowledge, Time and Decisions (How, When, and What): A Framework for Computational Intelligence Applications in Predictive Health Management" This effort presents a framework for understanding the requirements, limitations, and performance of a large variety of Computational Intelligence (or Soft Computing) models - based on Fuzzy Sets (Instance-based, C-means) evolutionary (GA, MOEA, GP), Neural Nets (Feedforward NN, ANFIS,SOM), Statistical (Random Forests, Kernel-Based approaches), Information Theory (Kolmogorov Complexity), etc. We apply these models to the area of Prognostics and Health Management (for equipment such as aircraft engines, CT scanners, locomotives, etc.). To analyze the output (decisions) of the models, we focus on two factors: 1) the time horizon for the decision (tactical, operational, strategic) and 2) the degree of availability of domain knowledge to construct the model. For the latter we use linguistics as a metaphor. We analyze the progression from simple lexicon to annotated lexicon, morphology, syntax, semantics, and pragmatics and compare it with the injection of domain knowledge in PHM. This knowledge start with event messages and is extended with event messages taxonomy, normal/failure labels, failure signatures, key variables (features) and first principle relationships, and context-based model selection. >From the combination of data and knowledge we can perform anomaly detection, anomaly identification, failure mode analysis (diagnostics), estimation of remaining useful life (prognostics), and optimal health management decisions. We illustrate key combinations of this space (Time x Knowledge) with short problem descriptions and related SC-based solutions. From kristens at cs.cmu.edu Wed Feb 22 13:37:16 2006 From: kristens at cs.cmu.edu (Kristen Schrauder) Date: Wed, 22 Feb 2006 13:37:16 -0500 Subject: [Research] there WILL be a lab meeting tomorrow! In-Reply-To: <1140623057.19356.57.camel@logo.auton.cs.cmu.edu> Message-ID: <007701c637df$04129840$fadc0280@adm.ri.cmu.edu> We will be in NSH 1507 as usual and there will be food :) -----Original Message----- From: Artur Dubrawski [mailto:awd at cs.cmu.edu] Sent: Wednesday, February 22, 2006 10:44 AM To: research at autonlab.org Cc: Kristen Schrauder Subject: there WILL be a lab meeting tomorrow! Hello, I am sorry for the late notice, but in fact we will have a lab meeting this Thursday and a guest speaker who, without much arm-twisting, agreed to talk about the topics explained below. Kristen: please make sure we have room available and let everyone know if it won't be the usual NSH 1507. And please get us some food :-) Artur ----- Piero Bonissone, General Electric Global Research "Knowledge, Time and Decisions (How, When, and What): A Framework for Computational Intelligence Applications in Predictive Health Management" This effort presents a framework for understanding the requirements, limitations, and performance of a large variety of Computational Intelligence (or Soft Computing) models - based on Fuzzy Sets (Instance-based, C-means) evolutionary (GA, MOEA, GP), Neural Nets (Feedforward NN, ANFIS,SOM), Statistical (Random Forests, Kernel-Based approaches), Information Theory (Kolmogorov Complexity), etc. We apply these models to the area of Prognostics and Health Management (for equipment such as aircraft engines, CT scanners, locomotives, etc.). To analyze the output (decisions) of the models, we focus on two factors: 1) the time horizon for the decision (tactical, operational, strategic) and 2) the degree of availability of domain knowledge to construct the model. For the latter we use linguistics as a metaphor. We analyze the progression from simple lexicon to annotated lexicon, morphology, syntax, semantics, and pragmatics and compare it with the injection of domain knowledge in PHM. This knowledge start with event messages and is extended with event messages taxonomy, normal/failure labels, failure signatures, key variables (features) and first principle relationships, and context-based model selection. >From the combination of data and knowledge we can perform anomaly detection, anomaly identification, failure mode analysis (diagnostics), estimation of remaining useful life (prognostics), and optimal health management decisions. We illustrate key combinations of this space (Time x Knowledge) with short problem descriptions and related SC-based solutions. From kristens at cs.cmu.edu Thu Feb 23 14:57:02 2006 From: kristens at cs.cmu.edu (Kristen Schrauder) Date: Thu, 23 Feb 2006 14:57:02 -0500 Subject: [Research] there WILL be a lab meeting tomorrow! Message-ID: <009201c638b3$53697740$fadc0280@adm.ri.cmu.edu> Just a reminder, there will be a lab meeting today (2/23) in NSH 1507 at 4:30pm. Thanks! -----Original Message----- From: Kristen Schrauder [mailto:kristens at cs.cmu.edu] Sent: Wednesday, February 22, 2006 1:37 PM To: 'research at autonlab.org' Cc: 'awd at cs.cmu.edu' Subject: RE: there WILL be a lab meeting tomorrow! We will be in NSH 1507 as usual and there will be food :) -----Original Message----- From: Artur Dubrawski [mailto:awd at cs.cmu.edu] Sent: Wednesday, February 22, 2006 10:44 AM To: research at autonlab.org Cc: Kristen Schrauder Subject: there WILL be a lab meeting tomorrow! Hello, I am sorry for the late notice, but in fact we will have a lab meeting this Thursday and a guest speaker who, without much arm-twisting, agreed to talk about the topics explained below. Kristen: please make sure we have room available and let everyone know if it won't be the usual NSH 1507. And please get us some food :-) Artur ----- Piero Bonissone, General Electric Global Research "Knowledge, Time and Decisions (How, When, and What): A Framework for Computational Intelligence Applications in Predictive Health Management" This effort presents a framework for understanding the requirements, limitations, and performance of a large variety of Computational Intelligence (or Soft Computing) models - based on Fuzzy Sets (Instance-based, C-means) evolutionary (GA, MOEA, GP), Neural Nets (Feedforward NN, ANFIS,SOM), Statistical (Random Forests, Kernel-Based approaches), Information Theory (Kolmogorov Complexity), etc. We apply these models to the area of Prognostics and Health Management (for equipment such as aircraft engines, CT scanners, locomotives, etc.). To analyze the output (decisions) of the models, we focus on two factors: 1) the time horizon for the decision (tactical, operational, strategic) and 2) the degree of availability of domain knowledge to construct the model. For the latter we use linguistics as a metaphor. We analyze the progression from simple lexicon to annotated lexicon, morphology, syntax, semantics, and pragmatics and compare it with the injection of domain knowledge in PHM. This knowledge start with event messages and is extended with event messages taxonomy, normal/failure labels, failure signatures, key variables (features) and first principle relationships, and context-based model selection. >From the combination of data and knowledge we can perform anomaly detection, anomaly identification, failure mode analysis (diagnostics), estimation of remaining useful life (prognostics), and optimal health management decisions. We illustrate key combinations of this space (Time x Knowledge) with short problem descriptions and related SC-based solutions. From awd at cs.cmu.edu Fri Feb 24 20:45:52 2006 From: awd at cs.cmu.edu (Artur Dubrawski) Date: Fri, 24 Feb 2006 20:45:52 -0500 Subject: [Research] Reminder: two lab meetings next week Message-ID: <1140831952.26921.284.camel@logo.auton.cs.cmu.edu> Hello, We will have Ashwin presenting on Tuesday (at noon) and Purna on Thursday (the usual time). Artur From tengli at cs.cmu.edu Mon Feb 27 18:04:08 2006 From: tengli at cs.cmu.edu (Ashwin Tengli) Date: Mon, 27 Feb 2006 18:04:08 -0500 Subject: [Research] summary for lab meeting tomorrow In-Reply-To: <1140831952.26921.284.camel@logo.auton.cs.cmu.edu> References: <1140831952.26921.284.camel@logo.auton.cs.cmu.edu> Message-ID: <44038568.40003@cs.cmu.edu> Hi, I will be presenting a collection of new classifiers (knn,naive bayes, decision tree, logistic regression) based on a new data-structure called abdat to handle both dense and sparse attributes and set of wrapper structures of classifiers. I will briefly talk about how more classifiers can be added to the collection. In the remaining time I will discuss creating ball-trees in metric space and improving speed of decision trees on sparse attributes. - Ashwin From awd at cs.cmu.edu Tue Feb 28 11:58:26 2006 From: awd at cs.cmu.edu (Artur Dubrawski) Date: Tue, 28 Feb 2006 11:58:26 -0500 Subject: [Research] Reminder: Lab meting starting in a minute! Message-ID: <1141145905.11107.0.camel@logo.auton.cs.cmu.edu> From kristens at cs.cmu.edu Tue Feb 28 15:48:12 2006 From: kristens at cs.cmu.edu (Kristen Schrauder) Date: Tue, 28 Feb 2006 15:48:12 -0500 Subject: [Research] FW: Reminder: two lab meetings next week Message-ID: <010501c63ca8$4db7c7e0$fadc0280@adm.ri.cmu.edu> FYI, the lab meeting on Thursday will be in NSH 4201 at 4:30pm (we've been bumped by other meetings). It will probably be a little snug as it is a much smaller room than 1507. Sorry for the inconvenience! > -----Original Message----- > From: Artur Dubrawski [mailto:awd at cs.cmu.edu] > Sent: Friday, February 24, 2006 8:46 PM > To: research at autonlab.org > Cc: Kristen Schrauder > Subject: Reminder: two lab meetings next week > > Hello, > > We will have Ashwin presenting on Tuesday (at noon) > and Purna on Thursday (the usual time). > > Artur > > > > > >