From yongjiaw at engin.umich.edu Thu Jun 2 10:40:08 2005 From: yongjiaw at engin.umich.edu (yongjiaw) Date: Thu, 2 Jun 2005 10:40:08 -0400 Subject: [ACT-R-users] question - ACT-R's theories about chunk Message-ID: <020201c56781$092af720$260cd58d@raven> Hi, I'm interested in what is the complete theory about chunk, the declarative memory units, in ACT-R. The ACT-R theory is pretty strong about how chunk is retrieved. While, for other aspects, such as chunk creation and chunk modification, there seems not to be much resources (I mean mainly the tutorial). Seems that chunk can only be created via goal buffer, and the visual module, and there is no further restriction beyond that? Chunk can be destructively modified in goal buffer as well. What are the theories about those specifications? Please direct me to some more detailed document about that. Thanks a lot. -------------- next part -------------- An HTML attachment was scrubbed... URL: From jmfs at fe.up.pt Mon Jun 6 14:19:17 2005 From: jmfs at fe.up.pt (Jorge M. Santos) Date: Mon, 6 Jun 2005 19:19:17 +0100 Subject: [ACT-R-users] SUMMER SCHOOL NN2005 - NEURAL NETWORKS in CLASSIFICATION, REGRESSION and DATA MINING Message-ID: <002701c56ac4$ab33cc40$333ea8c0@jmfs> Apologies for multiple posting > We appreciate if you can forward this Announcement to potential > candidates. > FINAL ANNOUNCEMENT > ============================================================= > SUMMER SCHOOL NN2005 - with a special poster session and workshop (bring > your data sets). > > NEURAL NETWORKS in CLASSIFICATION, REGRESSION and DATA MINING > > July 4-8, 2005, ISEP - Porto, Portugal; http://www.nn.isep.ipp.pt; > nn-2005 at isep.ipp.pt > ============================================================= > > GENERAL INFORMATION > > The Summer School will be held at Porto, Portugal, jointly organized by > the Polytechnic School of Engineering of Porto (ISEP) and the Faculty of > Engineering, Porto University (FEUP). > > This year's edition, NN2005, includes a special WORKSHOP SESSION providing > a discussion forum where the participants can obtain peer guidance for > their projects. > > > Previous editions of the Summer School have reached a considerable > success. > Two opinions: > > "NN2003 adressed current topics in neural networks and machine learning in > a thorough and accessible manner ... presented relatively new material > that I found extremely useful ... there was something for experts as well > as for beginners" - Prof. Darrel Whitley, Colorado State University. > > > "The Summer School ... was of a most impressive quality. ... the students > were quickly brought up to a level where they could be introduced to more > advanced material ... excellent introductions to neural networks, support > vector machines, and genetic algorithms ... the students could further > master their skills by applying the course material to case studies in a > state-of-the art computer laboratory...a most impressive extensive guest > program ... I highly recommend this course" - Prof. Mark J. Embrechts, > Rensselaer Polytechnic Institute. > > > PROGRAMME COMMITTEE (PROVISIONAL) > > > ? Christopher Bishop (Assistant Director, Microsoft Research, Cambridge, > U.K.) > > ? Fernando Sereno (Assistant Professor, High School of Education, > Portugal) > > ? Joaquim Marques-de-S?, Programme Chair (Associate Professor, Dept. > Electr. and Comp. Engineering; Fac. of Engineering, University of Porto, > Portugal) > > ? Jos? Carlos Pr?ncipe (BellSouth Professor, Director Computational > NeuroEngineering, University of Florida, USA) > > ? Juergen Schmidhuber (Professor Scuola Universitaria Professionale, > Lugano, Switzerland) > > ? Lu?s Alexandre (Assistant Professor, Beira Interior University, > Portugal) > > ? Mark Embrechts (Associate Professor at the Decision Sciences and > Engineering Systems Faculty, Rensselaer Polytechnic Institute, RPI Troy, > USA) > > ? Paulo Cortez (Assistant Professor, University of Minho, Portugal) > > ? Steve R. Gunn (Senior Lecturer, Dept. of Electronics and Computer > Science, University of Southampton, UK) > > > LOCAL ORGANIZING COMMITTEE (LOC) > > - Alzira Faria - Assistant, Dept. Mathematics, ISEP, Portugal > > - Jorge M. Santos - Assistant, Dept. Mathematics, ISEP, Portugal > > - Rui Chibante - Assistant, Dept. Mathematics, ISEP, Portugal > > > COURSE CONTENTS > > Neural networks (NN) have become a very important tool in classification > and regression tasks. The applications are nowadays abundant, e.g. in the > engineering, economy and biology areas. The Summer School on NN is > dedicated to explain relevant NN paradigms, namely multilayer perceptrons > (MLP), radial basis function networks (RBF), support vector machines (SVM) > and entropy-based networks (ENN) used for classification and regression > tasks, illustrated with applications to real data. Specific topics are > also presented, namely relevance vector machines, recurrent networks, > co-operative structures of NN and data mining using NN. > > Classes include practical sessions with appropriate software tools. The > trainee has, therefore, the opportunity to apply the taught concepts and > become conversant with a broad range of NN topics and applications. A > special workshop session will provide a discussion forum where the > participants can obtain peer guidance for their projects. > > Official language is English. > > > PRELIMINARY PROGRAMME > > A preliminary programme and further informations about the classes are > available at the school webpage (http://www.nn.isep.ipp.pt) > > > POSTER and WORKSHOP SESSION > In the session "Workshop and Poster", we intend to discuss some of the > posters presented by the participants. > The purpose of the poster is to present some problem that the participant > wish to discuss with the lecturers and other participants, as a way of > getting some advise in the better way to solve it using neural networks. > Encouraging this form of participation as a means of interaction and > cooperative learning, we invite you to present a poster with a specific > problem, especially one involved with your research. The organizing committee will welcome that the participants bring along the data sets and a summary description of the most important problem they would be interested in order to find possible solutions in this school. > IMPORTANT DEADLINES > > Early Registration: 15 May 2005 > > Hotel booking: 15 June 2005 > > Summer School: 4-8 July 2005 > > All participants are required to register prior to the start of the School > - until the 15th of June - even if you choose to pay the late registration > fee at the registration desk. > Please note that only a LIMITED number of participants can be accepted. > > > REGISTRATION > > In order to attend the School you must fill in the registration form, > available at the School web page. Please note that if you have any guests > who would like to take part in the social programme, you must register > them as well, by filling in the corresponding field in the registration > form. > > > SCHOOL FEES > > The registration fee for participants amounts to: > > - Early registration fee (paid before the 15th of May) > > * 300 Euro (students, ISEP and FEUP staff) > > * 350 Euro (all other participants) > > - Late registration fee (paid after the 15th of May) > > * 350 Euro (students, ISEP and FEUP staff) > > * 400 Euro (all other participants) > > The registration fee includes: > > * school package (manuscripts, CD) > * coffee breaks > * daily lunch > * welcome reception > * school banquet > > NOTE: The registration fee for those who attended previous editions > amounts to 20 euro per lecture and includes the school package and > coffee-breaks. Please, contact the LOC for further details. > > > CONTACT ADDRESS > > NN2005 Secretariat > > Ms. Gabriela Afonso > Email: gafonso at fe.up.pt > > > Local Organizing Committee (LOC) - Summer School NN2005 > Departamento de Matem?tica > Instituto Superior de Engenharia do Porto > Rua Dr. Ant?nio Bernardino de Almeida 431 > 4200-072 PORTO / PORTUGAL > Email: nn-2005 at isep.ipp.pt > > http://www.nn.isep.ipp.pt > > Programme Chair: > Prof. Marques de S? > Tel. 225081828 - Email: jmsa at fe.up.pt > ======================================== > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From Raluca.Budiu at parc.com Thu Jun 9 15:53:32 2005 From: Raluca.Budiu at parc.com (Raluca.Budiu at parc.com) Date: Thu, 9 Jun 2005 12:53:32 PDT Subject: [ACT-R-users] ACT-R output from GLSA server@PARC Message-ID: <0B5954ACD6FE2340B81D1F79C91B192E1D9E3C@MOONRAKER.ad.parc.com> Some of you may have heard about PARC's effort to build an external GLSA/PMI server; it is now available at: http://glsa.parc.com/ and it produces ACT-R output (other formats are supported as well). GLSA (Generalized Latent Semantic Analysis) is a LSA-like method of computing word similarities, but it has the advantage of an adjustable, web-based corpus. It takes as input a list of word pairs and provides similarities between those words. Just very recently it started providing ACT-R output; this is an ACT-R file that defines a meaning chunk type and sets the Sij-s between words to their similarity value as computed by the server. The server is still in a development phase, but please feel free to experiment with it. Comments and suggestions are very welcome. Raluca Budiu From tkelley at arl.army.mil Thu Jun 9 16:52:52 2005 From: tkelley at arl.army.mil (Kelley, Troy (Civ,ARL/HRED)) Date: Thu, 9 Jun 2005 16:52:52 -0400 Subject: [ACT-R-users] ACT-R output from GLSA server@PARC Message-ID: Here is a sample output from the GLSA server at PARC for the word Love love love 0.9999997 love fun 0.16717705 love city 0.103955254 love place 0.2078263 love man 0.36086833 love friend 0.19896048 love neighbor -0.037641484 love woman 0.21863273 love fondness 0.0076576206 Interesting, love is more similar to the word "man" than "woman", and more similar to "man" than "fondness" or "friend". Troy -----Original Message----- From: act-r-users-admin at act-r.psy.cmu.edu [mailto:act-r-users-admin at act-r.psy.cmu.edu] On Behalf Of Raluca.Budiu at parc.com Sent: Thursday, June 09, 2005 3:54 PM To: act-r-users at act-r.psy.cmu.edu Subject: [ACT-R-users] ACT-R output from GLSA server at PARC Some of you may have heard about PARC's effort to build an external GLSA/PMI server; it is now available at: http://glsa.parc.com/ and it produces ACT-R output (other formats are supported as well). GLSA (Generalized Latent Semantic Analysis) is a LSA-like method of computing word similarities, but it has the advantage of an adjustable, web-based corpus. It takes as input a list of word pairs and provides similarities between those words. Just very recently it started providing ACT-R output; this is an ACT-R file that defines a meaning chunk type and sets the Sij-s between words to their similarity value as computed by the server. The server is still in a development phase, but please feel free to experiment with it. Comments and suggestions are very welcome. Raluca Budiu _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users From R.Belavkin at mdx.ac.uk Fri Jun 17 09:44:18 2005 From: R.Belavkin at mdx.ac.uk (Roman Belavkin) Date: Fri, 17 Jun 2005 14:44:18 +0100 Subject: [ACT-R-users] OPTIMIST conflict resolution for ACT-R Message-ID: <45272C367897174E94DEE1D40C5F0CA7017B30CF@MDX-CLX-DC1.uni.mdx.ac.uk> Hello dear ACT-R users, I have released the new version of the Optimist conflict resolution for ACT-R. You can access the files and documentation here: http://www.cs.mdx.ac.uk/staffpages/rvb/software/optimist/ It should work with both ACT-R 5 and 4 (there are two versions), and there is an example model. In brief, the algorithm provides a more dynamic and rule-specific noise which leads to a slightly different rate of converegence and formation of preferences to certain rules. There is also a simple reinfercement mechanism. The underlying theory is explained in corresponding papers. If conflict resolution and parameters learning plays a significant role in your models, then I suggest you to try this. I would also be glad to hear how the algorithm works for your models. Best wishes! Roman Bealvkin Middlesex University London NW4 4BT, UK From R.Belavkin at mdx.ac.uk Fri Jun 17 10:17:52 2005 From: R.Belavkin at mdx.ac.uk (Roman Belavkin) Date: Fri, 17 Jun 2005 15:17:52 +0100 Subject: [SPAM: 4.500] RE: [ACT-R-users] ACT-R output from GLSA server@PARC Message-ID: <45272C367897174E94DEE1D40C5F0CA7017B30D0@MDX-CLX-DC1.uni.mdx.ac.uk> Hi, I think the word `similar' is not really appropriate here. LSA just shows co-occurance of the two words really, and higher co-occurance for word man can be simply explained because the word man means also human, while woman is more specific and thus less ambiguous. Cheers, Roman -----Original Message----- From: act-r-users-admin at act-r.psy.cmu.edu on behalf of Kelley, Troy (Civ,ARL/HRED) Sent: Thu 6/9/2005 21:52 To: act-r-users at act-r.psy.cmu.edu Cc: Subject: [SPAM: 4.500] RE: [ACT-R-users] ACT-R output from GLSA server at PARC Here is a sample output from the GLSA server at PARC for the word Love love love 0.9999997 love fun 0.16717705 love city 0.103955254 love place 0.2078263 love man 0.36086833 love friend 0.19896048 love neighbor -0.037641484 love woman 0.21863273 love fondness 0.0076576206 Interesting, love is more similar to the word "man" than "woman", and more similar to "man" than "fondness" or "friend". Troy -----Original Message----- From: act-r-users-admin at act-r.psy.cmu.edu [mailto:act-r-users-admin at act-r.psy.cmu.edu] On Behalf Of Raluca.Budiu at parc.com Sent: Thursday, June 09, 2005 3:54 PM To: act-r-users at act-r.psy.cmu.edu Subject: [ACT-R-users] ACT-R output from GLSA server at PARC Some of you may have heard about PARC's effort to build an external GLSA/PMI server; it is now available at: http://glsa.parc.com/ and it produces ACT-R output (other formats are supported as well). GLSA (Generalized Latent Semantic Analysis) is a LSA-like method of computing word similarities, but it has the advantage of an adjustable, web-based corpus. It takes as input a list of word pairs and provides similarities between those words. Just very recently it started providing ACT-R output; this is an ACT-R file that defines a meaning chunk type and sets the Sij-s between words to their similarity value as computed by the server. The server is still in a development phase, but please feel free to experiment with it. Comments and suggestions are very welcome. Raluca Budiu _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users This message has been checked for viruses but the contents of an attachment may still contain software viruses, which could damage your computer system: you are advised to perform your own checks. Email communications with the University of Nottingham may be monitored as permitted by UK legislation. From R.Belavkin at mdx.ac.uk Fri Jun 17 10:40:18 2005 From: R.Belavkin at mdx.ac.uk (Roman Belavkin) Date: Fri, 17 Jun 2005 15:40:18 +0100 Subject: [ACT-R-users] on similarity and glsa Message-ID: <45272C367897174E94DEE1D40C5F0CA7017B30D1@MDX-CLX-DC1.uni.mdx.ac.uk> I have looked at the site, and it states it is using mutual information as a metric, which measures as we know statistical dependence (so, it is not just correlations betwen terms). Still I am not sure if similarity, as we understand it, and statistical dependene are the same things. For example, here are the similarities for the word apple: apple mac 0.23059334 apple microsoft 0.21670483 apple dkz 0.21022835 etc... So, the most `similar' word to apple is mac (or how about `dkz'?). To me, orange or a fruit seems more similar terms. What these numbers show is a degree of statistical dependence of two terms in the docements analysed. It would be an interesting project for the ACT-R community to investigate this difference. What do you think? Roman -----Original Message----- From: Roman Belavkin Sent: Fri 6/17/2005 15:17 To: Kelley, Troy (Civ,ARL/HRED); act-r-users at act-r.psy.cmu.edu Cc: Subject: RE: [SPAM: 4.500] RE: [ACT-R-users] ACT-R output from GLSA server at PARC Hi, I think the word `similar' is not really appropriate here. LSA just shows co-occurance of the two words really, and higher co-occurance for word man can be simply explained because the word man means also human, while woman is more specific and thus less ambiguous. Cheers, Roman -----Original Message----- From: act-r-users-admin at act-r.psy.cmu.edu on behalf of Kelley, Troy (Civ,ARL/HRED) Sent: Thu 6/9/2005 21:52 To: act-r-users at act-r.psy.cmu.edu Cc: Subject: [SPAM: 4.500] RE: [ACT-R-users] ACT-R output from GLSA server at PARC Here is a sample output from the GLSA server at PARC for the word Love love love 0.9999997 love fun 0.16717705 love city 0.103955254 love place 0.2078263 love man 0.36086833 love friend 0.19896048 love neighbor -0.037641484 love woman 0.21863273 love fondness 0.0076576206 Interesting, love is more similar to the word "man" than "woman", and more similar to "man" than "fondness" or "friend". Troy -----Original Message----- From: act-r-users-admin at act-r.psy.cmu.edu [mailto:act-r-users-admin at act-r.psy.cmu.edu] On Behalf Of Raluca.Budiu at parc.com Sent: Thursday, June 09, 2005 3:54 PM To: act-r-users at act-r.psy.cmu.edu Subject: [ACT-R-users] ACT-R output from GLSA server at PARC Some of you may have heard about PARC's effort to build an external GLSA/PMI server; it is now available at: http://glsa.parc.com/ and it produces ACT-R output (other formats are supported as well). GLSA (Generalized Latent Semantic Analysis) is a LSA-like method of computing word similarities, but it has the advantage of an adjustable, web-based corpus. It takes as input a list of word pairs and provides similarities between those words. Just very recently it started providing ACT-R output; this is an ACT-R file that defines a meaning chunk type and sets the Sij-s between words to their similarity value as computed by the server. The server is still in a development phase, but please feel free to experiment with it. Comments and suggestions are very welcome. Raluca Budiu _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users This message has been checked for viruses but the contents of an attachment may still contain software viruses, which could damage your computer system: you are advised to perform your own checks. Email communications with the University of Nottingham may be monitored as permitted by UK legislation. From R.Belavkin at mdx.ac.uk Fri Jun 17 10:58:55 2005 From: R.Belavkin at mdx.ac.uk (Roman Belavkin) Date: Fri, 17 Jun 2005 15:58:55 +0100 Subject: [ACT-R-users] on similarity Message-ID: <45272C367897174E94DEE1D40C5F0CA7017B30D4@MDX-CLX-DC1.uni.mdx.ac.uk> One more observation: The results depend greatly on the number of eigenvectors used (I understand this controls the number of dimension PCA reduces the space to). I found the `optimal' value for the word apple is about 50 dimensions, becuase the results are: apple carnation 0.6078935 apple fruit 0.5972236 apple blackberry 0.59559596 apple mac 0.5931967 apple orange 0.570567 apple sweet 0.5677523 apple palm 0.5576204 apple radish 0.55353457 apple intel 0.5511009 apple persimmon 0.54269683 They are quite different as you can see. Cheers! Roman -----Original Message----- From: act-r-users-admin at act-r.psy.cmu.edu on behalf of Roman Belavkin Sent: Fri 6/17/2005 15:17 To: Kelley, Troy (Civ,ARL/HRED); act-r-users at act-r.psy.cmu.edu Cc: Subject: RE: [SPAM: 4.500] RE: [ACT-R-users] ACT-R output from GLSA server at PARC Hi, I think the word `similar' is not really appropriate here. LSA just shows co-occurance of the two words really, and higher co-occurance for word man can be simply explained because the word man means also human, while woman is more specific and thus less ambiguous. Cheers, Roman -----Original Message----- From: act-r-users-admin at act-r.psy.cmu.edu on behalf of Kelley, Troy (Civ,ARL/HRED) Sent: Thu 6/9/2005 21:52 To: act-r-users at act-r.psy.cmu.edu Cc: Subject: [SPAM: 4.500] RE: [ACT-R-users] ACT-R output from GLSA server at PARC Here is a sample output from the GLSA server at PARC for the word Love love love 0.9999997 love fun 0.16717705 love city 0.103955254 love place 0.2078263 love man 0.36086833 love friend 0.19896048 love neighbor -0.037641484 love woman 0.21863273 love fondness 0.0076576206 Interesting, love is more similar to the word "man" than "woman", and more similar to "man" than "fondness" or "friend". Troy -----Original Message----- From: act-r-users-admin at act-r.psy.cmu.edu [mailto:act-r-users-admin at act-r.psy.cmu.edu] On Behalf Of Raluca.Budiu at parc.com Sent: Thursday, June 09, 2005 3:54 PM To: act-r-users at act-r.psy.cmu.edu Subject: [ACT-R-users] ACT-R output from GLSA server at PARC Some of you may have heard about PARC's effort to build an external GLSA/PMI server; it is now available at: http://glsa.parc.com/ and it produces ACT-R output (other formats are supported as well). GLSA (Generalized Latent Semantic Analysis) is a LSA-like method of computing word similarities, but it has the advantage of an adjustable, web-based corpus. It takes as input a list of word pairs and provides similarities between those words. Just very recently it started providing ACT-R output; this is an ACT-R file that defines a meaning chunk type and sets the Sij-s between words to their similarity value as computed by the server. The server is still in a development phase, but please feel free to experiment with it. Comments and suggestions are very welcome. Raluca Budiu _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users This message has been checked for viruses but the contents of an attachment may still contain software viruses, which could damage your computer system: you are advised to perform your own checks. Email communications with the University of Nottingham may be monitored as permitted by UK legislation. _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users From R.Belavkin at mdx.ac.uk Fri Jun 17 10:59:47 2005 From: R.Belavkin at mdx.ac.uk (Roman Belavkin) Date: Fri, 17 Jun 2005 15:59:47 +0100 Subject: [ACT-R-users] similarity Message-ID: <45272C367897174E94DEE1D40C5F0CA7017B30D5@MDX-CLX-DC1.uni.mdx.ac.uk> I have looked at the site, and it states it is using mutual information as a metric, which measures as we know statistical dependence (so, it is not just correlations betwen terms). Still I am not sure if similarity, as we understand it, and statistical dependene are the same things. For example, here are the similarities for the word apple: apple mac 0.23059334 apple microsoft 0.21670483 apple dkz 0.21022835 etc... So, the most `similar' word to apple is mac (or how about `dkz'?). To me, orange or a fruit seems more similar terms. What these numbers show is a degree of statistical dependence of two terms in the docements analysed. It would be an interesting project for the ACT-R community to investigate this difference. What do you think? Roman -----Original Message----- From: act-r-users-admin at act-r.psy.cmu.edu on behalf of Roman Belavkin Sent: Fri 6/17/2005 15:17 To: Kelley, Troy (Civ,ARL/HRED); act-r-users at act-r.psy.cmu.edu Cc: Subject: RE: [SPAM: 4.500] RE: [ACT-R-users] ACT-R output from GLSA server at PARC Hi, I think the word `similar' is not really appropriate here. LSA just shows co-occurance of the two words really, and higher co-occurance for word man can be simply explained because the word man means also human, while woman is more specific and thus less ambiguous. Cheers, Roman -----Original Message----- From: act-r-users-admin at act-r.psy.cmu.edu on behalf of Kelley, Troy (Civ,ARL/HRED) Sent: Thu 6/9/2005 21:52 To: act-r-users at act-r.psy.cmu.edu Cc: Subject: [SPAM: 4.500] RE: [ACT-R-users] ACT-R output from GLSA server at PARC Here is a sample output from the GLSA server at PARC for the word Love love love 0.9999997 love fun 0.16717705 love city 0.103955254 love place 0.2078263 love man 0.36086833 love friend 0.19896048 love neighbor -0.037641484 love woman 0.21863273 love fondness 0.0076576206 Interesting, love is more similar to the word "man" than "woman", and more similar to "man" than "fondness" or "friend". Troy -----Original Message----- From: act-r-users-admin at act-r.psy.cmu.edu [mailto:act-r-users-admin at act-r.psy.cmu.edu] On Behalf Of Raluca.Budiu at parc.com Sent: Thursday, June 09, 2005 3:54 PM To: act-r-users at act-r.psy.cmu.edu Subject: [ACT-R-users] ACT-R output from GLSA server at PARC Some of you may have heard about PARC's effort to build an external GLSA/PMI server; it is now available at: http://glsa.parc.com/ and it produces ACT-R output (other formats are supported as well). GLSA (Generalized Latent Semantic Analysis) is a LSA-like method of computing word similarities, but it has the advantage of an adjustable, web-based corpus. It takes as input a list of word pairs and provides similarities between those words. Just very recently it started providing ACT-R output; this is an ACT-R file that defines a meaning chunk type and sets the Sij-s between words to their similarity value as computed by the server. The server is still in a development phase, but please feel free to experiment with it. Comments and suggestions are very welcome. Raluca Budiu _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users This message has been checked for viruses but the contents of an attachment may still contain software viruses, which could damage your computer system: you are advised to perform your own checks. Email communications with the University of Nottingham may be monitored as permitted by UK legislation. _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users From rwilson+ at pitt.edu Fri Jun 17 11:15:46 2005 From: rwilson+ at pitt.edu (Roy Wilson) Date: Fri, 17 Jun 2005 11:15:46 -0400 Subject: [ACT-R-users] similarity In-Reply-To: <45272C367897174E94DEE1D40C5F0CA7017B30D5@MDX-CLX-DC1.uni.mdx.ac.uk> References: <45272C367897174E94DEE1D40C5F0CA7017B30D5@MDX-CLX-DC1.uni.mdx.ac.uk> Message-ID: <200506171115.48739.rwilson@pitt.edu> On Friday 17 June 2005 10:59, Roman Belavkin wrote: > I have looked at the site, and it states it is using mutual information as > a metric, which measures as we know statistical dependence (so, it is not > just correlations betwen terms). Still I am not sure if similarity, as we > understand it, and statistical dependene are the same things. For example, > here are the similarities for the word apple: > > apple mac 0.23059334 > apple microsoft 0.21670483 > apple dkz 0.21022835 > etc... > > So, the most `similar' word to apple is mac (or how about `dkz'?). To me, > orange or a fruit seems more similar terms. What these numbers show is a > degree of statistical dependence of two terms in the docements analysed. > > It would be an interesting project for the ACT-R community to investigate > this difference. What do you think? I don't know what the answer is, but Manning and Schutze have an fairly extensive discussion of these issues in (I think) "Statistical Foundations of Natural Language Processing" (MIT Press, 2000). > > Roman > > -----Original Message----- > From: act-r-users-admin at act-r.psy.cmu.edu on behalf of Roman Belavkin > Sent: Fri 6/17/2005 15:17 > To: Kelley, Troy (Civ,ARL/HRED); act-r-users at act-r.psy.cmu.edu > Cc: > Subject: RE: [SPAM: 4.500] RE: [ACT-R-users] ACT-R output from GLSA > server at PARC > > > > Hi, > > I think the word `similar' is not really appropriate here. LSA just shows > co-occurance of the two words really, and higher co-occurance for word man > can be simply explained because the word man means also human, while woman > is more specific and thus less ambiguous. > > > Cheers, > Roman > > -----Original Message----- > From: act-r-users-admin at act-r.psy.cmu.edu on behalf of Kelley, > Troy (Civ,ARL/HRED) Sent: Thu 6/9/2005 21:52 > To: act-r-users at act-r.psy.cmu.edu > Cc: > Subject: [SPAM: 4.500] RE: [ACT-R-users] ACT-R output from GLSA > server at PARC > > > > Here is a sample output from the GLSA server at PARC for the word > Love > > love love 0.9999997 > love fun 0.16717705 > love city 0.103955254 > love place 0.2078263 > love man 0.36086833 > love friend 0.19896048 > love neighbor -0.037641484 > love woman 0.21863273 > love fondness 0.0076576206 > > Interesting, love is more similar to the word "man" than "woman", > and more similar to "man" than "fondness" or "friend". > > Troy > > -----Original Message----- > From: act-r-users-admin at act-r.psy.cmu.edu > [mailto:act-r-users-admin at act-r.psy.cmu.edu] On Behalf Of > Raluca.Budiu at parc.com > Sent: Thursday, June 09, 2005 3:54 PM > To: act-r-users at act-r.psy.cmu.edu > Subject: [ACT-R-users] ACT-R output from GLSA server at PARC > > > > Some of you may have heard about PARC's effort to build an > external GLSA/PMI server; it is now available at: > > http://glsa.parc.com/ > > and it produces ACT-R output (other formats are supported as > well). > > GLSA (Generalized Latent Semantic Analysis) is a LSA-like method > of computing word similarities, but it has the advantage of an adjustable, > web-based corpus. It takes as input a list of word pairs and provides > similarities between those words. > > Just very recently it started providing ACT-R output; this is an > ACT-R file that defines a meaning chunk type and sets the Sij-s between > words to their similarity value as computed by the server. > > The server is still in a development phase, but please feel free > to experiment with it. Comments and suggestions are very welcome. > > Raluca Budiu > > _______________________________________________ > ACT-R-users mailing list > ACT-R-users at act-r.psy.cmu.edu > http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users > > _______________________________________________ > ACT-R-users mailing list > ACT-R-users at act-r.psy.cmu.edu > http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users > > This message has been checked for viruses but the contents of an > attachment may still contain software viruses, which could damage your > computer system: you are advised to perform your own checks. Email > communications with the University of Nottingham may be monitored as > permitted by UK legislation. > > > _______________________________________________ > ACT-R-users mailing list > ACT-R-users at act-r.psy.cmu.edu > http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users > > > _______________________________________________ > ACT-R-users mailing list > ACT-R-users at act-r.psy.cmu.edu > http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users -- Roy Wilson Learning Research Development Center University of Pittsburgh webpage: www.pitt.edu/~rwilson email: rwilson at pitt.edu From pirolli at parc.com Fri Jun 17 12:50:05 2005 From: pirolli at parc.com (Peter Pirolli) Date: Fri, 17 Jun 2005 09:50:05 PDT Subject: [ACT-R-users] similarity In-Reply-To: <200506171115.48739.rwilson@pitt.edu> Message-ID: <5.2.0.9.2.20050617094013.0325be90@hermes.parc.xerox.com> There may be some confusion over notions of similarity, so let me try to help. I would like to second Roy's pointer to the Manning and Schuetze book: It is an excellent introduction to Pointwise Mutual Information (PMI) and LSA (it's an excellent introduction to just about every basic concept and technique in statistical natural language processing). As I point out in the Pirolli (2005) Cognitive Science paper, PMI is approximately the same thing as association strength in ACT-R. PMI (like LSA) is very good at generating scores that correlate well with synonym judgements (e.g., as tested by the TOEFL test), which might be one way to define "similarity." The PMI scores in the GLSA server are used to set distances among word items that feed into the LSA computation (Ayman Farahat argues that this provides better results on a number of tests). The GLSA scores are dependent on the particulars of the document corpus over which they are computed, and we're interested in improving that corpus. --Pete At 08:15 AM 6/17/2005 -0700, Roy Wilson wrote: >On Friday 17 June 2005 10:59, Roman Belavkin wrote: > > I have looked at the site, and it states it is using mutual information as > > a metric, which measures as we know statistical dependence (so, it is not > > just correlations betwen terms). Still I am not sure if similarity, as we > > understand it, and statistical dependene are the same things. For example, > > here are the similarities for the word apple: > > > > apple mac 0.23059334 > > apple microsoft 0.21670483 > > apple dkz 0.21022835 > > etc... > > > > So, the most `similar' word to apple is mac (or how about `dkz'?). To me, > > orange or a fruit seems more similar terms. What these numbers show is a > > degree of statistical dependence of two terms in the docements analysed. > > > > It would be an interesting project for the ACT-R community to investigate > > this difference. What do you think? > >I don't know what the answer is, but Manning and Schutze have an fairly >extensive discussion of these issues in (I think) "Statistical Foundations of >Natural Language Processing" (MIT Press, 2000). > > > > > > Roman > > > > -----Original Message----- > > From: act-r-users-admin at act-r.psy.cmu.edu on behalf of Roman Belavkin > > Sent: Fri 6/17/2005 15:17 > > To: Kelley, Troy (Civ,ARL/HRED); act-r-users at act-r.psy.cmu.edu > > Cc: > > Subject: RE: [SPAM: 4.500] RE: [ACT-R-users] ACT-R output from GLSA > > server at PARC > > > > > > > > Hi, > > > > I think the word `similar' is not really appropriate here. LSA > just shows > > co-occurance of the two words really, and higher co-occurance for word man > > can be simply explained because the word man means also human, while woman > > is more specific and thus less ambiguous. > > > > > > Cheers, > > Roman > > > > -----Original Message----- > > From: act-r-users-admin at act-r.psy.cmu.edu on behalf of > Kelley, > > Troy (Civ,ARL/HRED) Sent: Thu 6/9/2005 21:52 > > To: act-r-users at act-r.psy.cmu.edu > > Cc: > > Subject: [SPAM: 4.500] RE: [ACT-R-users] ACT-R output > from GLSA > > server at PARC > > > > > > > > Here is a sample output from the GLSA server at PARC for the > word > > Love > > > > love love 0.9999997 > > love fun 0.16717705 > > love city 0.103955254 > > love place 0.2078263 > > love man 0.36086833 > > love friend 0.19896048 > > love neighbor -0.037641484 > > love woman 0.21863273 > > love fondness 0.0076576206 > > > > Interesting, love is more similar to the word "man" than > "woman", > > and more similar to "man" than "fondness" or "friend". > > > > Troy > > > > -----Original Message----- > > From: act-r-users-admin at act-r.psy.cmu.edu > > [mailto:act-r-users-admin at act-r.psy.cmu.edu] On Behalf Of > > Raluca.Budiu at parc.com > > Sent: Thursday, June 09, 2005 3:54 PM > > To: act-r-users at act-r.psy.cmu.edu > > Subject: [ACT-R-users] ACT-R output from GLSA server at PARC > > > > > > > > Some of you may have heard about PARC's effort to build an > > external GLSA/PMI server; it is now available at: > > > > http://glsa.parc.com/ > > > > and it produces ACT-R output (other formats are supported as > > well). > > > > GLSA (Generalized Latent Semantic Analysis) is a LSA-like > method > > of computing word similarities, but it has the advantage of an adjustable, > > web-based corpus. It takes as input a list of word pairs and provides > > similarities between those words. > > > > Just very recently it started providing ACT-R output; > this is an > > ACT-R file that defines a meaning chunk type and sets the Sij-s between > > words to their similarity value as computed by the server. > > > > The server is still in a development phase, but please > feel free > > to experiment with it. Comments and suggestions are very welcome. > > > > Raluca Budiu > > > > _______________________________________________ > > ACT-R-users mailing list > > ACT-R-users at act-r.psy.cmu.edu > > http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users > > > > _______________________________________________ > > ACT-R-users mailing list > > ACT-R-users at act-r.psy.cmu.edu > > http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users > > > > This message has been checked for viruses but the > contents of an > > attachment may still contain software viruses, which could damage your > > computer system: you are advised to perform your own checks. Email > > communications with the University of Nottingham may be monitored as > > permitted by UK legislation. > > > > > > _______________________________________________ > > ACT-R-users mailing list > > ACT-R-users at act-r.psy.cmu.edu > > http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users > > > > > > _______________________________________________ > > ACT-R-users mailing list > > ACT-R-users at act-r.psy.cmu.edu > > http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users > >-- >Roy Wilson >Learning Research Development Center >University of Pittsburgh >webpage: www.pitt.edu/~rwilson >email: rwilson at pitt.edu > >_______________________________________________ >ACT-R-users mailing list >ACT-R-users at act-r.psy.cmu.edu >http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users From farahat at parc.com Fri Jun 17 13:18:19 2005 From: farahat at parc.com (Ayman Farahat) Date: Fri, 17 Jun 2005 10:18:19 PDT Subject: [ACT-R-users] similarity In-Reply-To: <5.2.0.9.2.20050617094013.0325be90@hermes.parc.xerox.com> Message-ID: Hello The apple example is interesting because it illustrates how the corpus (or domain knowledge) can influence similarity. My feeling is that if we asked Steve Jobs (or for that matter a random person in the bay area) about the most similar term to apple, MAC would come up very high. An interesting example to try is "entropy". You will see that the similarity server captures the two senses of the term; the information theory and thermodynamics. The reason why we get better results with entropy has more to do with the corpus than with technique. Ayman > From: Peter Pirolli > Date: Fri, 17 Jun 2005 09:50:05 PDT > To: Roy Wilson , act-r-users at act-r.psy.cmu.edu > Cc: Ayman Farahat , Raluca.Budiu at parc.com, royer at parc.com > Subject: Re: [ACT-R-users] similarity > > As I point out > in the Pirolli (2005) Cognitive Science paper, PMI is approximately the > same thing as association strength in ACT-R. PMI (like LSA) is very good at > generating scores that correlate well with synonym judgements (e.g., as > tested by the TOEFL test), which might be one way to define From Raluca.Budiu at parc.com Fri Jun 17 13:28:32 2005 From: Raluca.Budiu at parc.com (Raluca.Budiu at parc.com) Date: Fri, 17 Jun 2005 10:28:32 PDT Subject: [ACT-R-users] on similarity Message-ID: <0B5954ACD6FE2340B81D1F79C91B192E1D9FBA@MOONRAKER.ad.parc.com> Yes, it's true that the results depend on the number of eigen vectors. The intuition is that, the more eigen vectors, the more noise (i.e., accidental co-occurrences) the system takes into account. With lower numbers of eigen vectors, more generalization is allowed. Ayman has run some tests and found out that the optimal number of eigen vectors is around 200-300 (this is also true for the traditional LSA). As for the results of the GLSA being called "similarities", that's a topic very much open to debate indeed. As Peter has said, there is some theoretical motivation to them being called so (that PMIs can be shown to roughly be equivalent with the original definition of strengths of association in ACT-R 4.0). Whether association is the same with similarity, that's another story. In ACT-R 5.0 I think there was a tacit understanding (maybe not so tacit, if I remember correctly some of the discussions at the ACT-R workshops and postgraduate summer school) that there is some need to experiment with Sji-s and see exactly what they should reflect (co-occurrence or perhaps some other measure of semantic similarity). Raluca From: Roman Belavkin Date: Jun 17, 2005 7:58 AM Subject: [ACT-R-users] on similarity To: act-r-users at act-r.psy.cmu.edu One more observation: The results depend greatly on the number of eigenvectors used (I understand this controls the number of dimension PCA reduces the space to). I found the `optimal' value for the word apple is about 50 dimensions, becuase the results are: apple carnation 0.6078935 apple fruit 0.5972236 apple blackberry 0.59559596 apple mac 0.5931967 apple orange 0.570567 apple sweet 0.5677523 apple palm 0.5576204 apple radish 0.55353457 apple intel 0.5511009 apple persimmon 0.54269683 They are quite different as you can see. Cheers! Roman -----Original Message----- From: act-r-users-admin at act-r.psy.cmu.edu on behalf of Roman Belavkin Sent: Fri 6/17/2005 15:17 To: Kelley, Troy (Civ,ARL/HRED); act-r-users at act-r.psy.cmu.edu Cc: Subject: RE: [SPAM: 4.500] RE: [ACT-R-users] ACT-R output from GLSA server at PARC Hi, I think the word `similar' is not really appropriate here. LSA just shows co-occurance of the two words really, and higher co-occurance for word man can be simply explained because the word man means also human, while woman is more specific and thus less ambiguous. Cheers, Roman -----Original Message----- From: act-r-users-admin at act-r.psy.cmu.edu on behalf of Kelley, Troy (Civ,ARL/HRED) Sent: Thu 6/9/2005 21:52 To: act-r-users at act-r.psy.cmu.edu Cc: Subject: [SPAM: 4.500] RE: [ACT-R-users] ACT-R output from GLSA server at PARC Here is a sample output from the GLSA server at PARC for the word Love love love 0.9999997 love fun 0.16717705 love city 0.103955254 love place 0.2078263 love man 0.36086833 love friend 0.19896048 love neighbor -0.037641484 love woman 0.21863273 love fondness 0.0076576206 Interesting, love is more similar to the word "man" than "woman", and more similar to "man" than "fondness" or "friend". Troy -----Original Message----- From: act-r-users-admin at act-r.psy.cmu.edu [mailto:act-r-users-admin at act-r.psy.cmu.edu] On Behalf Of Raluca.Budiu at parc.com Sent: Thursday, June 09, 2005 3:54 PM To: act-r-users at act-r.psy.cmu.edu Subject: [ACT-R-users] ACT-R output from GLSA server at PARC Some of you may have heard about PARC's effort to build an external GLSA/PMI server; it is now available at: http://glsa.parc.com/ and it produces ACT-R output (other formats are supported as well). GLSA (Generalized Latent Semantic Analysis) is a LSA-like method of computing word similarities, but it has the advantage of an adjustable, web-based corpus. It takes as input a list of word pairs and provides similarities between those words. Just very recently it started providing ACT-R output; this is an ACT-R file that defines a meaning chunk type and sets the Sij-s between words to their similarity value as computed by the server. The server is still in a development phase, but please feel free to experiment with it. Comments and suggestions are very welcome. Raluca Budiu _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users This message has been checked for viruses but the contents of an attachment may still contain software viruses, which could damage your computer system: you are advised to perform your own checks. Email communications with the University of Nottingham may be monitored as permitted by UK legislation. _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From Raluca.Budiu at parc.com Fri Jun 17 14:17:12 2005 From: Raluca.Budiu at parc.com (Raluca.Budiu at parc.com) Date: Fri, 17 Jun 2005 11:17:12 PDT Subject: [ACT-R-users] on similarity Message-ID: <0B5954ACD6FE2340B81D1F79C91B192E1D9FC4@MOONRAKER.ad.parc.com> Yes, it's true that the results depend on the number of eigen vectors. The intuition is that, the more eigen vectors, the more noise (i.e., accidental co-occurrences) the system takes into account. With lower numbers of eigen vectors, more generalization is allowed. Ayman has run some tests and found out that the optimal number of eigen vectors is around 200-300 (this is also true for the traditional LSA). As for the results of the GLSA being called "similarities", that's a topic very much open to debate indeed. As Peter has said, there is some theoretical motivation to them being called so (that PMIs can be shown to roughly be equivalent with the original definition of strengths of association in ACT-R 4.0). Whether association is the same with similarity, that's another story. In ACT-R 5.0 I think there was a tacit understanding (maybe not so tacit, if I remember correctly some of the discussions at the ACT-R workshops and postgraduate summer school) that there is some need to experiment with Sji-s and see exactly what they should reflect (co-occurrence or perhaps some other measure of semantic similarity). Raluca From: Roman Belavkin Date: Jun 17, 2005 7:58 AM Subject: [ACT-R-users] on similarity To: act-r-users at act-r.psy.cmu.edu One more observation: The results depend greatly on the number of eigenvectors used (I understand this controls the number of dimension PCA reduces the space to). I found the `optimal' value for the word apple is about 50 dimensions, becuase the results are: apple carnation 0.6078935 apple fruit 0.5972236 apple blackberry 0.59559596 apple mac 0.5931967 apple orange 0.570567 apple sweet 0.5677523 apple palm 0.5576204 apple radish 0.55353457 apple intel 0.5511009 apple persimmon 0.54269683 They are quite different as you can see. Cheers! Roman -----Original Message----- From: act-r-users-admin at act-r.psy.cmu.edu on behalf of Roman Belavkin Sent: Fri 6/17/2005 15:17 To: Kelley, Troy (Civ,ARL/HRED); act-r-users at act-r.psy.cmu.edu Cc: Subject: RE: [SPAM: 4.500] RE: [ACT-R-users] ACT-R output from GLSA server at PARC Hi, I think the word `similar' is not really appropriate here. LSA just shows co-occurance of the two words really, and higher co-occurance for word man can be simply explained because the word man means also human, while woman is more specific and thus less ambiguous. Cheers, Roman -----Original Message----- From: act-r-users-admin at act-r.psy.cmu.edu on behalf of Kelley, Troy (Civ,ARL/HRED) Sent: Thu 6/9/2005 21:52 To: act-r-users at act-r.psy.cmu.edu Cc: Subject: [SPAM: 4.500] RE: [ACT-R-users] ACT-R output from GLSA server at PARC Here is a sample output from the GLSA server at PARC for the word Love love love 0.9999997 love fun 0.16717705 love city 0.103955254 love place 0.2078263 love man 0.36086833 love friend 0.19896048 love neighbor -0.037641484 love woman 0.21863273 love fondness 0.0076576206 Interesting, love is more similar to the word "man" than "woman", and more similar to "man" than "fondness" or "friend". Troy -----Original Message----- From: act-r-users-admin at act-r.psy.cmu.edu [mailto:act-r-users-admin at act-r.psy.cmu.edu] On Behalf Of Raluca.Budiu at parc.com Sent: Thursday, June 09, 2005 3:54 PM To: act-r-users at act-r.psy.cmu.edu Subject: [ACT-R-users] ACT-R output from GLSA server at PARC Some of you may have heard about PARC's effort to build an external GLSA/PMI server; it is now available at: http://glsa.parc.com/ and it produces ACT-R output (other formats are supported as well). GLSA (Generalized Latent Semantic Analysis) is a LSA-like method of computing word similarities, but it has the advantage of an adjustable, web-based corpus. It takes as input a list of word pairs and provides similarities between those words. Just very recently it started providing ACT-R output; this is an ACT-R file that defines a meaning chunk type and sets the Sij-s between words to their similarity value as computed by the server. The server is still in a development phase, but please feel free to experiment with it. Comments and suggestions are very welcome. Raluca Budiu _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users This message has been checked for viruses but the contents of an attachment may still contain software viruses, which could damage your computer system: you are advised to perform your own checks. Email communications with the University of Nottingham may be monitored as permitted by UK legislation. _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From wfu at andrew.cmu.edu Fri Jun 17 16:02:54 2005 From: wfu at andrew.cmu.edu (Wai-Tat Fu) Date: Fri, 17 Jun 2005 16:02:54 -0400 Subject: [ACT-R-users] similarity In-Reply-To: References: Message-ID: <150DB8FCF9A834380421E5F1@[192.168.2.105]> Just want to add to Pete and Ayman's point about the use of PMI as Sji values in ACT-R. From that perspective, spreading activation will calculate the weighted sum of these PMI/Sji values with respect to your active goal. It is seldom the case that you have only one slot in your goal. In Ayman's example, if you have "apple", "Steve jobs", or "bay area" as slots in your goal, MAC will likely to be more active than fruits. This is similar to the use of multiple keywords when you google for some information on the web. Just as an example, I know that a number of HCI studies (e.g., menu search) used questionnaires to get subjective ratings as a measure of semantic relevance for their model. I think the tool will be very useful for this kind of research. --On Friday, June 17, 2005 10:18 AM -0700 Ayman Farahat wrote: > Hello > The apple example is interesting because it illustrates how the corpus (or > domain knowledge) can influence similarity. My feeling is that if we > asked Steve Jobs (or for that matter a random person in the bay area) > about the most similar term to apple, MAC would come up very high. > An interesting example to try is "entropy". You will see that the > similarity server captures the two senses of the term; the information > theory and thermodynamics. > The reason why we get better results with entropy has more to do with the > corpus than with technique. > Ayman > >> From: Peter Pirolli >> Date: Fri, 17 Jun 2005 09:50:05 PDT >> To: Roy Wilson , act-r-users at act-r.psy.cmu.edu >> Cc: Ayman Farahat , Raluca.Budiu at parc.com, >> royer at parc.com Subject: Re: [ACT-R-users] similarity >> >> As I point out >> in the Pirolli (2005) Cognitive Science paper, PMI is approximately the >> same thing as association strength in ACT-R. PMI (like LSA) is very good >> at generating scores that correlate well with synonym judgements (e.g., >> as tested by the TOEFL test), which might be one way to define > > > _______________________________________________ > ACT-R-users mailing list > ACT-R-users at act-r.psy.cmu.edu > http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users > From dbrumby at gmail.com Fri Jun 17 16:01:15 2005 From: dbrumby at gmail.com (Duncan Brumby) Date: Fri, 17 Jun 2005 13:01:15 -0700 Subject: [ACT-R-users] similarity In-Reply-To: References: <5.2.0.9.2.20050617094013.0325be90@hermes.parc.xerox.com> Message-ID: Hi Ayman, Hope all is going well. So your point that: >The reason why we get better results with entropy has more >to do with the corpus than with technique. Still presents the challenge for, say modeling a user searching a web page, because how can one make an a priori assumption about that users knowledge (or in this case, which training corpus should be selected). Further, if a model using a particular corpus does not fit the empirical data, it is then unclear whether this is because the model is incorrect or that the corpus was incorrect ... a similar case can of course be made when attributing the success of a model to fitting the data. --Duncan ~~~~~~~~~~~~~~~~~~~~~~ Duncan Brumby Intern Microsoft Research One Microsoft Way, Building 113 Redmond, WA 98052 phone: +1 (425) 706 8259 x68259 email: BrumbyDP at cardiff.ac.uk web: http://www.cardiff.ac.uk/psych On 6/17/05, Ayman Farahat wrote: > Hello > The apple example is interesting because it illustrates how the corpus (or > domain knowledge) can influence similarity. My feeling is that if we asked > Steve Jobs (or for that matter a random person in the bay area) about the > most similar term to apple, MAC would come up very high. > An interesting example to try is "entropy". You will see that the similarity > server captures the two senses of the term; the information theory and > thermodynamics. > The reason why we get better results with entropy has more to do with the > corpus than with technique. > Ayman > > > From: Peter Pirolli > > Date: Fri, 17 Jun 2005 09:50:05 PDT > > To: Roy Wilson , act-r-users at act-r.psy.cmu.edu > > Cc: Ayman Farahat , Raluca.Budiu at parc.com, royer at parc.com > > Subject: Re: [ACT-R-users] similarity > > > > As I point out > > in the Pirolli (2005) Cognitive Science paper, PMI is approximately the > > same thing as association strength in ACT-R. PMI (like LSA) is very good at > > generating scores that correlate well with synonym judgements (e.g., as > > tested by the TOEFL test), which might be one way to define > > > _______________________________________________ > ACT-R-users mailing list > ACT-R-users at act-r.psy.cmu.edu > http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users > From R.Belavkin at mdx.ac.uk Sat Jun 18 03:55:10 2005 From: R.Belavkin at mdx.ac.uk (Roman Belavkin) Date: Sat, 18 Jun 2005 08:55:10 +0100 Subject: [ACT-R-users] similarity Message-ID: <45272C367897174E94DEE1D40C5F0CA7017B30DA@MDX-CLX-DC1.uni.mdx.ac.uk> Hello Ayman, Indeed, Steve Jobs may have a bit different associations than myself, which begs for some explanation, and I think it is beyond a simple difference in the corpus we use. Perhaps term apple, obviously ambiguous, meaning two very different objects, corresponds to two different representations in the mind (i.e. chunks). And people are able to make this distinction very clearly, whereas a purely text-based technique, such as GLSA (or ICA, LSA, etc) has no means to do it. Or does it? I wonder if there has been any research on trying to capture this difference with LSA? Obviously, apple meaning fruit is `similar' to terms such as eat, orange, sweet and red, etc. While apple meaning mac would have associations with terms such as Steve Jobs, OS X, etc. The same term belongs to quite different clusters, therefore the method should be able to figure this out. cheers, Roman -----Original Message----- From: act-r-users-admin at act-r.psy.cmu.edu on behalf of Ayman Farahat Sent: Fri 6/17/2005 18:18 To: Peter Pirolli; Roy Wilson; act-r-users at act-r.psy.cmu.edu Cc: Raluca.Budiu at parc.com; royer at parc.com Subject: Re: [ACT-R-users] similarity Hello The apple example is interesting because it illustrates how the corpus (or domain knowledge) can influence similarity. My feeling is that if we asked Steve Jobs (or for that matter a random person in the bay area) about the most similar term to apple, MAC would come up very high. An interesting example to try is "entropy". You will see that the similarity server captures the two senses of the term; the information theory and thermodynamics. The reason why we get better results with entropy has more to do with the corpus than with technique. Ayman > From: Peter Pirolli > Date: Fri, 17 Jun 2005 09:50:05 PDT > To: Roy Wilson , act-r-users at act-r.psy.cmu.edu > Cc: Ayman Farahat , Raluca.Budiu at parc.com, royer at parc.com > Subject: Re: [ACT-R-users] similarity > > As I point out > in the Pirolli (2005) Cognitive Science paper, PMI is approximately the > same thing as association strength in ACT-R. PMI (like LSA) is very good at > generating scores that correlate well with synonym judgements (e.g., as > tested by the TOEFL test), which might be one way to define _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users From grayw at rpi.edu Sat Jun 18 08:42:23 2005 From: grayw at rpi.edu (Wayne Gray) Date: Sat, 18 Jun 2005 08:42:23 -0400 Subject: [ACT-R-users] Measures of semantic distance Message-ID: Friends, For those of you new to the topic of measures of semantic distance (MSDs) I can recommend a short and very well-written discussion of how the major systems differ: Lemaire, B., & Denhi?re, G. (2004). Incremental construction of an associative network from a corpus. In K. D. Forbus & D. Gentner & T. Regier (Eds.), 26th Annual Meeting of the Cognitive Science Society, CogSci2004. Hillsdale, NJ: Lawrence Erlbaum Publisher. There is also a very nice introduction to PMI-IR written by Peter D. Turney and published as a Lecture Note inn Computer Science. "Mining the Web for synonyms: PMI-IR versus LSA on TOEFL." Unfortunately I do not have the complete citation on this one. Wayne -- **Rensselaer**Rensselaer**Rensselaer**Rensselaer**Rensselaer** Wayne D. Gray; Professor of Cognitive Science Rensselaer Polytechnic Institute Carnegie Building (rm 108) ;;for all surface mail & deliveries 110 8th St.; Troy, NY 12180 EMAIL: grayw at rpi.edu, Office: 518-276-3315, Fax: 518-276-3017 for general information see: http://www.rpi.edu/~grayw/ for On-Line publications see: http://www.rpi.edu/~grayw/pubs/downloadable_pubs.htm for the CogWorks Lab see: http://www.cogsci.rpi.edu/cogworks/ If you just have formalisms or a model you are doing "operations research" or" AI", if you just have data and a good study you are doing "experimental psychology", and if you just have ideas you are doing "philosophy" -- it takes all three to do cognitive science. **Rensselaer**Rensselaer**Rensselaer**Rensselaer**Rensselaer** From stephane at gamard.net Sat Jun 18 08:59:19 2005 From: stephane at gamard.net (Stephane Gamard) Date: Sat, 18 Jun 2005 14:59:19 +0200 Subject: [ACT-R-users] Measures of semantic distance In-Reply-To: References: Message-ID: <58cafe07d2b331e975f4cf2421a0d4e4@gamard.net> just for information here is the citation ;-) Turney, Peter (2001) Mining the Web for Synonyms: PMI-IR versus LSA on TOEFL. In De Raedt, Luc and Flach, Peter, Eds. Proceedings Proceedings of the Twelfth European Conference on Machine Learning (ECML-2001), pages pp.?491-502, Freiburg, Germany. available on cogprints.org _Stephane On Jun 18, 2005, at 2:42 PM, Wayne Gray wrote: > Friends, > > For those of you new to the topic of measures of semantic distance > (MSDs) I can recommend a short and very well-written discussion of how > the major systems differ: > > Lemaire, B., & Denhi?re, G. (2004). Incremental construction of an > associative network from a corpus. In K. D. Forbus & D. Gentner & T. > Regier (Eds.), 26th Annual Meeting of the Cognitive Science Society, > CogSci2004. Hillsdale, NJ: Lawrence Erlbaum Publisher. > > > There is also a very nice introduction to PMI-IR written by Peter D. > Turney and published as a Lecture Note inn Computer Science. "Mining > the Web for synonyms: PMI-IR versus LSA on TOEFL." > > Unfortunately I do not have the complete citation on this one. > > Wayne > -- > **Rensselaer**Rensselaer**Rensselaer**Rensselaer**Rensselaer** > Wayne D. Gray; Professor of Cognitive Science > Rensselaer Polytechnic Institute > Carnegie Building (rm 108) ;;for all surface mail & deliveries > 110 8th St.; Troy, NY 12180 > > EMAIL: grayw at rpi.edu, Office: 518-276-3315, Fax: 518-276-3017 > > for general information see: http://www.rpi.edu/~grayw/ > > for On-Line publications see: > http://www.rpi.edu/~grayw/pubs/downloadable_pubs.htm > > for the CogWorks Lab see: http://www.cogsci.rpi.edu/cogworks/ > > If you just have formalisms or a model you are doing "operations > research" or" AI", if you just have data and a good study you are > doing "experimental psychology", and if you just have ideas you are > doing "philosophy" -- it takes all three to do cognitive science. > > **Rensselaer**Rensselaer**Rensselaer**Rensselaer**Rensselaer** > > _______________________________________________ > ACT-R-users mailing list > ACT-R-users at act-r.psy.cmu.edu > http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users > > -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: text/enriched Size: 2269 bytes Desc: not available URL: From dalrympl at mail.ucf.edu Sun Jun 19 07:39:07 2005 From: dalrympl at mail.ucf.edu (Ann Dalrymple) Date: Sun, 19 Jun 2005 07:39:07 -0400 Subject: [ACT-R-users] Re: Measures of semantic distance Message-ID: There is also some nice work done by ontologists connected with Word-Net - for example, Fernando Gomez, George Miller. In Word-Net, semantic distance is not merely a function of association frequency. Here's a site to get started with: http://wordnet.princeton.edu/ Today's Topics: 1. Measures of semantic distance (Wayne Gray) 2. Re: Measures of semantic distance (Stephane Gamard) --__--__-- Message: 1 Date: Sat, 18 Jun 2005 08:42:23 -0400 To: act-r-users+ at andrew.cmu.edu From: Wayne Gray Subject: [ACT-R-users] Measures of semantic distance =46riends, =46or those of you new to the topic of measures of=20 semantic distance (MSDs) I can recommend a short=20 and very well-written discussion of how the major=20 systems differ: Lemaire, B., & Denhi=E9re, G. (2004). Incremental=20 construction of an associative network from a=20 corpus. In K. D. Forbus & D. Gentner & T. Regier=20 (Eds.), 26th Annual Meeting of the Cognitive=20 Science Society, CogSci2004. Hillsdale, NJ:=20 Lawrence Erlbaum Publisher. There is also a very nice introduction to PMI-IR=20 written by Peter D. Turney and published as a=20 Lecture Note inn Computer Science. "Mining the=20 Web for synonyms: PMI-IR versus LSA on TOEFL." Unfortunately I do not have the complete citation on this one. Wayne -- **Rensselaer**Rensselaer**Rensselaer**Rensselaer**Rensselaer** Wayne D. Gray; Professor of Cognitive Science Rensselaer Polytechnic Institute Carnegie Building (rm 108) ;;for all surface mail & deliveries 110 8th St.; Troy, NY 12180 EMAIL: grayw at rpi.edu, Office: 518-276-3315, Fax: 518-276-3017 for general information see: http://www.rpi.edu/~grayw/ for On-Line publications see:=20 http://www.rpi.edu/~grayw/pubs/downloadable_pubs.htm for the CogWorks Lab see: http://www.cogsci.rpi.edu/cogworks/ If you just have formalisms or a model you are=20 doing "operations research" or" AI", if you just=20 have data and a good study you are doing=20 "experimental psychology", and if you just have=20 ideas you are doing "philosophy" -- it takes all=20 three to do cognitive science. **Rensselaer**Rensselaer**Rensselaer**Rensselaer**Rensselaer** --__--__-- Message: 2 Cc: act-r-users+ at andrew.cmu.edu From: Stephane Gamard Subject: Re: [ACT-R-users] Measures of semantic distance Date: Sat, 18 Jun 2005 14:59:19 +0200 To: Wayne Gray --Apple-Mail-5--941560708 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=ISO-8859-1; format=flowed just for information here is the citation ;-) Turney, Peter (2001) Mining the Web for Synonyms: PMI-IR versus LSA on=20= TOEFL. In De Raedt, Luc and Flach, Peter, Eds. Proceedings Proceedings=20= of the Twelfth European Conference on Machine Learning (ECML-2001),=20 pages pp.=A0491-502, Freiburg, Germany. available on cogprints.org _Stephane On Jun 18, 2005, at 2:42 PM, Wayne Gray wrote: > Friends, > > For those of you new to the topic of measures of semantic distance=20 > (MSDs) I can recommend a short and very well-written discussion of how=20= > the major systems differ: > > Lemaire, B., & Denhi=E9re, G. (2004). Incremental construction of an=20= > associative network from a corpus. In K. D. Forbus & D. Gentner & T.=20= > Regier (Eds.), 26th Annual Meeting of the Cognitive Science Society,=20= > CogSci2004. Hillsdale, NJ: Lawrence Erlbaum Publisher. > > > There is also a very nice introduction to PMI-IR written by Peter D.=20= > Turney and published as a Lecture Note inn Computer Science. "Mining=20= > the Web for synonyms: PMI-IR versus LSA on TOEFL." > > Unfortunately I do not have the complete citation on this one. > > Wayne > -- > **Rensselaer**Rensselaer**Rensselaer**Rensselaer**Rensselaer** > Wayne D. Gray; Professor of Cognitive Science > Rensselaer Polytechnic Institute > Carnegie Building (rm 108) ;;for all surface mail & deliveries > 110 8th St.; Troy, NY 12180 > > EMAIL: grayw at rpi.edu, Office: 518-276-3315, Fax: 518-276-3017 > > for general information see: http://www.rpi.edu/~grayw/ > > for On-Line publications see:=20 > http://www.rpi.edu/~grayw/pubs/downloadable_pubs.htm > > for the CogWorks Lab see: http://www.cogsci.rpi.edu/cogworks/ > > If you just have formalisms or a model you are doing "operations=20 > research" or" AI", if you just have data and a good study you are=20 > doing "experimental psychology", and if you just have ideas you are=20 > doing "philosophy" -- it takes all three to do cognitive science. > > **Rensselaer**Rensselaer**Rensselaer**Rensselaer**Rensselaer** > > _______________________________________________ > ACT-R-users mailing list > ACT-R-users at act-r.psy.cmu.edu > http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users > > --Apple-Mail-5--941560708 Content-Transfer-Encoding: quoted-printable Content-Type: text/enriched; charset=ISO-8859-1 just for information here is the citation ;-) ArialTurney, Peter (2001) Mining the Web for Synonyms: PMI-IR versus LSA on TOEFL. In De Raedt, Luc and Flach, Peter, Eds. Proceedings Proceedings of the Twelfth European Conference on Machine Learning (ECML-2001), pages pp.=A0491-502, Freiburg, Germany. available on cogprints.org _Stephane On Jun 18, 2005, at 2:42 PM, Wayne Gray wrote: Friends, For those of you new to the topic of measures of semantic distance (MSDs) I can recommend a short and very well-written discussion of how the major systems differ: Lemaire, B., & Denhi=E9re, G. (2004). Incremental construction of an associative network from a corpus. In K. D. Forbus & D. Gentner & T. Regier (Eds.), 26th Annual Meeting of the Cognitive Science Society, CogSci2004. Hillsdale, NJ: Lawrence Erlbaum Publisher. There is also a very nice introduction to PMI-IR written by Peter D. Turney and published as a Lecture Note inn Computer Science. "Mining the Web for synonyms: PMI-IR versus LSA on TOEFL." Unfortunately I do not have the complete citation on this one. Wayne -- **Rensselaer**Rensselaer**Rensselaer**Rensselaer**Rensselaer** Wayne D. Gray; Professor of Cognitive Science Rensselaer Polytechnic Institute Carnegie Building (rm 108) ;;for all surface mail & deliveries 110 8th St.; Troy, NY 12180 EMAIL: grayw at rpi.edu, Office: 518-276-3315, Fax: 518-276-3017 for general information see: http://www.rpi.edu/~grayw/ for On-Line publications see: http://www.rpi.edu/~grayw/pubs/downloadable_pubs.htm for the CogWorks Lab see: http://www.cogsci.rpi.edu/cogworks/ If you just have formalisms or a model you are doing "operations research" or" AI", if you just have data and a good study you are doing "experimental psychology", and if you just have ideas you are doing "philosophy" -- it takes all three to do cognitive science. **Rensselaer**Rensselaer**Rensselaer**Rensselaer**Rensselaer** _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users = --Apple-Mail-5--941560708-- --__--__-- _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users End of ACT-R-users Digest ----- "Emotion resolves resource conflicts in systems that work with multiple goals in finite time with limited resources [and] controls attention and behavior in light of goals" - Herbert Simon, father of AI. ----- 407-381-0892 (phone) 1-443-342-0330 (fax) From dalrympl at mail.ucf.edu Mon Jun 20 06:19:31 2005 From: dalrympl at mail.ucf.edu (Ann Dalrymple) Date: Mon, 20 Jun 2005 06:19:31 -0400 Subject: [ACT-R-users] Re: Measures of semantic distance Message-ID: ----- "Emotion resolves resource conflicts in systems that work with multiple goals in finite time with limited resources [and] controls attention and behavior in light of goals" - Herbert Simon, father of AI. ----- 407-381-0892 (phone) 1-443-342-0330 (fax) >>> "Ann Dalrymple" 06/19/05 7:39 AM >>> There is also some nice work done by ontologists connected with Word-Net - for example, Fernando Gomez, George Miller. In Word-Net, semantic distance is not merely a function of association frequency. Here's a site to get started with: http://wordnet.princeton.edu/ Today's Topics: 1. Measures of semantic distance (Wayne Gray) 2. Re: Measures of semantic distance (Stephane Gamard) --__--__-- Message: 1 Date: Sat, 18 Jun 2005 08:42:23 -0400 To: act-r-users+ at andrew.cmu.edu From: Wayne Gray Subject: [ACT-R-users] Measures of semantic distance =46riends, =46or those of you new to the topic of measures of=20 semantic distance (MSDs) I can recommend a short=20 and very well-written discussion of how the major=20 systems differ: Lemaire, B., & Denhi=E9re, G. (2004). Incremental=20 construction of an associative network from a=20 corpus. In K. D. Forbus & D. Gentner & T. Regier=20 (Eds.), 26th Annual Meeting of the Cognitive=20 Science Society, CogSci2004. Hillsdale, NJ:=20 Lawrence Erlbaum Publisher. There is also a very nice introduction to PMI-IR=20 written by Peter D. Turney and published as a=20 Lecture Note inn Computer Science. "Mining the=20 Web for synonyms: PMI-IR versus LSA on TOEFL." Unfortunately I do not have the complete citation on this one. Wayne -- **Rensselaer**Rensselaer**Rensselaer**Rensselaer**Rensselaer** Wayne D. Gray; Professor of Cognitive Science Rensselaer Polytechnic Institute Carnegie Building (rm 108) ;;for all surface mail & deliveries 110 8th St.; Troy, NY 12180 EMAIL: grayw at rpi.edu, Office: 518-276-3315, Fax: 518-276-3017 for general information see: http://www.rpi.edu/~grayw/ for On-Line publications see:=20 http://www.rpi.edu/~grayw/pubs/downloadable_pubs.htm for the CogWorks Lab see: http://www.cogsci.rpi.edu/cogworks/ If you just have formalisms or a model you are=20 doing "operations research" or" AI", if you just=20 have data and a good study you are doing=20 "experimental psychology", and if you just have=20 ideas you are doing "philosophy" -- it takes all=20 three to do cognitive science. **Rensselaer**Rensselaer**Rensselaer**Rensselaer**Rensselaer** --__--__-- Message: 2 Cc: act-r-users+ at andrew.cmu.edu From: Stephane Gamard Subject: Re: [ACT-R-users] Measures of semantic distance Date: Sat, 18 Jun 2005 14:59:19 +0200 To: Wayne Gray --Apple-Mail-5--941560708 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=ISO-8859-1; format=flowed just for information here is the citation ;-) Turney, Peter (2001) Mining the Web for Synonyms: PMI-IR versus LSA on=20= TOEFL. In De Raedt, Luc and Flach, Peter, Eds. Proceedings Proceedings=20= of the Twelfth European Conference on Machine Learning (ECML-2001),=20 pages pp.=A0491-502, Freiburg, Germany. available on cogprints.org _Stephane On Jun 18, 2005, at 2:42 PM, Wayne Gray wrote: > Friends, > > For those of you new to the topic of measures of semantic distance=20 > (MSDs) I can recommend a short and very well-written discussion of how=20= > the major systems differ: > > Lemaire, B., & Denhi=E9re, G. (2004). Incremental construction of an=20= > associative network from a corpus. In K. D. Forbus & D. Gentner & T.=20= > Regier (Eds.), 26th Annual Meeting of the Cognitive Science Society,=20= > CogSci2004. Hillsdale, NJ: Lawrence Erlbaum Publisher. > > > There is also a very nice introduction to PMI-IR written by Peter D.=20= > Turney and published as a Lecture Note inn Computer Science. "Mining=20= > the Web for synonyms: PMI-IR versus LSA on TOEFL." > > Unfortunately I do not have the complete citation on this one. > > Wayne > -- > **Rensselaer**Rensselaer**Rensselaer**Rensselaer**Rensselaer** > Wayne D. Gray; Professor of Cognitive Science > Rensselaer Polytechnic Institute > Carnegie Building (rm 108) ;;for all surface mail & deliveries > 110 8th St.; Troy, NY 12180 > > EMAIL: grayw at rpi.edu, Office: 518-276-3315, Fax: 518-276-3017 > > for general information see: http://www.rpi.edu/~grayw/ > > for On-Line publications see:=20 > http://www.rpi.edu/~grayw/pubs/downloadable_pubs.htm > > for the CogWorks Lab see: http://www.cogsci.rpi.edu/cogworks/ > > If you just have formalisms or a model you are doing "operations=20 > research" or" AI", if you just have data and a good study you are=20 > doing "experimental psychology", and if you just have ideas you are=20 > doing "philosophy" -- it takes all three to do cognitive science. > > **Rensselaer**Rensselaer**Rensselaer**Rensselaer**Rensselaer** > > _______________________________________________ > ACT-R-users mailing list > ACT-R-users at act-r.psy.cmu.edu > http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users > > --Apple-Mail-5--941560708 Content-Transfer-Encoding: quoted-printable Content-Type: text/enriched; charset=ISO-8859-1 just for information here is the citation ;-) ArialTurney, Peter (2001) Mining the Web for Synonyms: PMI-IR versus LSA on TOEFL. In De Raedt, Luc and Flach, Peter, Eds. Proceedings Proceedings of the Twelfth European Conference on Machine Learning (ECML-2001), pages pp.=A0491-502, Freiburg, Germany. available on cogprints.org _Stephane On Jun 18, 2005, at 2:42 PM, Wayne Gray wrote: Friends, For those of you new to the topic of measures of semantic distance (MSDs) I can recommend a short and very well-written discussion of how the major systems differ: Lemaire, B., & Denhi=E9re, G. (2004). Incremental construction of an associative network from a corpus. In K. D. Forbus & D. Gentner & T. Regier (Eds.), 26th Annual Meeting of the Cognitive Science Society, CogSci2004. Hillsdale, NJ: Lawrence Erlbaum Publisher. There is also a very nice introduction to PMI-IR written by Peter D. Turney and published as a Lecture Note inn Computer Science. "Mining the Web for synonyms: PMI-IR versus LSA on TOEFL." Unfortunately I do not have the complete citation on this one. Wayne -- **Rensselaer**Rensselaer**Rensselaer**Rensselaer**Rensselaer** Wayne D. Gray; Professor of Cognitive Science Rensselaer Polytechnic Institute Carnegie Building (rm 108) ;;for all surface mail & deliveries 110 8th St.; Troy, NY 12180 EMAIL: grayw at rpi.edu, Office: 518-276-3315, Fax: 518-276-3017 for general information see: http://www.rpi.edu/~grayw/ for On-Line publications see: http://www.rpi.edu/~grayw/pubs/downloadable_pubs.htm for the CogWorks Lab see: http://www.cogsci.rpi.edu/cogworks/ If you just have formalisms or a model you are doing "operations research" or" AI", if you just have data and a good study you are doing "experimental psychology", and if you just have ideas you are doing "philosophy" -- it takes all three to do cognitive science. **Rensselaer**Rensselaer**Rensselaer**Rensselaer**Rensselaer** _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users = --Apple-Mail-5--941560708-- --__--__-- _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users End of ACT-R-users Digest ----- "Emotion resolves resource conflicts in systems that work with multiple goals in finite time with limited resources [and] controls attention and behavior in light of goals" - Herbert Simon, father of AI. ----- 407-381-0892 (phone) 1-443-342-0330 (fax) _______________________________________________ ACT-R-users mailing list ACT-R-users at act-r.psy.cmu.edu http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users From rsun at rpi.edu Fri Jun 10 14:11:54 2005 From: rsun at rpi.edu (Professor Ron Sun) Date: Fri, 10 Jun 2005 14:11:54 -0400 Subject: [ACT-R-users] Post-doctoral position available Message-ID: <11ce3d66603b6438b038818b482cccc6@rpi.edu> I am looking for a post-doctoral researcher, to join in a basic research project investigating cognitive modeling, cognitive architectures, and human skill learning. The starting date is September 1, 2005 (although a slight delay, say by a couple of months, is also possible). This will be a full-time research position, with the expectation that you devote all your time to project-related research work (not your own research topics). Prospective applicants should have a finished Ph.D degree before starting, by September 1, 2005 (or shortly thereafter). They should have a strong background in computer science (the equivalent of a BS in computer science), with strong Java programming skills, and have prior exposure to psychology and cognitive science (with background in human and machine learning, motivation, and meta-cognition preferred), and other related areas. Prospective applicants with interests in cognitive science should apply by EMAILing me: (1) a complete vitae, and (2) samples of best prior writings (especially published papers), and also FAX me (3) GRE/TOEFL scores (if available), and other pertinent information. Make sure to also FAX me (4) copies of all transcripts of all the BS, MS, Ph.D programs you previously attended. Also FAX me (5) the reference letters (if available). RPI is a top-tier research university. The new CogSci department has identified research as its primary missions. The department is conducting research in a number of areas: cognitive architectures, cognitive modeling, human and machine learning, multi-agent interactions and social simulation, neural networks and connectionist models, human and machine reasoning, cognitive engineering, perception and motor control, and so on. See the Web page below regarding my own research (in the afore-described area and in other areas): http://www.cogsci.rpi.edu/~rsun Two sample publications are downloadable from my Web site: R. Sun, P. Slusarz, and C. Terry, The interaction of the explicit and the implicit in skill learning: A dual-process approach. Psychological Review, Vol.112, No.1, pp.159-192. 2005. http://www.cogsci.rpi.edu/~rsun/sun-pr2005-f.pdf R. Sun, E. Merrill, and T. Peterson, From implicit skills to explicit knowledge: A bottom-up model of skill learning. Cognitive Science, Vol.25, No.2, pp.203-244. 2001. http://www.cogsci.rpi.edu/~rsun/sun.cs01.pdf Apply as soon as possible. Completed applications will be considered as they come in, until the position is filled. ======================================================== Professor Ron Sun Cognitive Science Department Rensselaer Polytechnic Institute 110 Eighth Street, Carnegie 302A Troy, NY 12180, USA phone: 518-276-3409 fax: 518-276-3017 email: rsun at rpi.edu web: http://www.cogsci.rpi.edu/~rsun ======================================================= From cimca at canberra.edu.au Sat Jun 18 00:00:09 2005 From: cimca at canberra.edu.au (cimca) Date: Sat, 18 Jun 2005 14:00:09 +1000 Subject: [ACT-R-users] CFP: IEEE inco-operated International Conference on Computational Intelligence for Modelling, Control and Automation Message-ID: <088D96D6E422174EAABF7B27D94F3EE702AD16E8@hera.ucstaff.win.canberra.edu.au> CALL FOR PAPERS International Conference on Computational Intelligence for Modelling, Control and Automation 28 - 30 November 2005 Vienna, Austria http://www.ise.canberra.edu.au/conferences/cimca05/ In co-operation with: IEEE Computational Intelligence Society Conference Proceedings will be published as books by IEEE (The Institute of Electrical and Electronic Engineering) Sponsored by: European Society for Fuzzy Logic and Technology - EUFLAT International Association for Fuzzy Set in Management and Economy - SIGEF Japan Society for Fuzzy Theory and Intelligent Informatics - SOFT Taiwan Fuzzy Systems Association - TFSA World Wide Web Business Intelligence - W3BI Hungarian Fuzzy Association - HFA University of Canberra Jointly with International Conference on Intelligent Agents, Web Technologies and Internet Commerce http://www.ise.canberra.edu.au/conferences/iawtic05/ Honorary Chair: Lotfi A. Zadeh, University of California, USA Stephen Grossberg, Boston University, USA The international conference on computational intelligence for modelling, control and automation will be held in Vienna, Austria on 28 to 30 November 2005. The conference provides a medium for the exchange of ideas between theoreticians and practitioners to address the important issues in computational intelligence, modelling, control and automation. The conference will consist of both plenary sessions and contributory sessions, focusing on theory, implementation and applications of computational intelligence techniques to modelling, control and automation. For contributory sessions, papers (4 pages or more) are being solicited. Several well-known keynote speakers will address the conference. Conference Proceedings will be published as books by IEEE (The Institute of Electrical and Electronic Engineering) in USA and will be index world wide. All papers will be peer reviewed by at least two reviewers. Topics of the conference include, but are not limited to, the following areas: Modern and Advanced Control Strategies: Neural Networks Control, Fuzzy Logic Control, Genetic Algorithms and Evolutionary Control, Model-Predictive Control, Adaptive and Optimal Control, Intelligent Control Systems, Robotics and Automation, Fault Diagnosis, Intelligent agents, Industrial Automations Hybrid Systems: Fuzzy Evolutionary Systems, Fuzzy Expert Systems, Fuzzy Neural Systems, Neural Genetic Systems, Neural-Fuzzy-Genetic Systems, Hybrid Systems for Optimisation Data Analysis, Prediction and Model Identification: Signal Processing, Prediction and Time Series Analysis, System Identification, Data Fusion and Mining, Knowledge Discovery, Intelligent Information Systems, Image Processing, and Image Understanding, Parallel Computing applications in Identification & Control, Pattern Recognition, Clustering and Classification Decision Making and Information Retrieval: Case-Based Reasoning, Decision Analysis, Intelligent Databases & Information Retrieval, Dynamic Systems Modelling, Decision Support Systems, Multi-criteria Decision Making, Qualitative and Approximate-Reasoning Paper Submission Papers will be selected based on their originality, significance, correctness, and clarity of presentation. Papers (4 pages or more) should be submitted to the following e-mail or the following address: CIMCA'2005 Secretariat School of Information Sciences and Engineering University of Canberra, Canberra, 2616, ACT, Australia E-mail: cimca at canberra.edu.au Electronic submission of papers (either by E-mail or through conference website) is preferred. Draft papers should present original work, which has not been published or being reviewed for other conferences. Important Dates 31 August 2005 Submission of draft papers 30 September 2005 Notification of acceptance 21 October 2005 Deadline for camera-ready copies of accepted papers 28-30 November 2005 Conference sessions Special Sessions and Tutorials Special sessions and tutorials will be organised at the conference. The conference is calling for special sessions and tutorial proposals. All special session proposals should be sent to the conference chair (by email to: masoud.mohammadian at canberra.edu.au) on or before 5th of August 2005. CIMCA'05 will also include a special poster session devoted to recent work and work-in-progress. Abstracts are solicited for this session. Abstracts (3 pages limit) may be submitted up to 30 days before the conference date. Visits and social events Sightseeing visits will be arranged for the delegates and guests. A separate program will be arranged for companions during the conference. Further Information For further information either contact cimca at ise.canberra.edu.au or see the conference homepage at: http://www.ise.canberra.edu.au/conferences/cimca05/default.htm Organising Committee Chair: Masoud Mohammadian, University of Canberra, Australia International Program Committee: H. Adeli, The Ohio State University, USA W. Pedrycz, University of Manitoba, Canada A. Agah, The University of Kansas, USA T. Fukuda, Nagoya University, Japan J. Bezdek, University of West Florida, USA R. C. Eberhart, Purdue University, USA F. Herrera, University of Granada, Spain T. Furuhashi, Nagoya University, Japan A. Agah, The University of Kansas, US E. Andr?, Universit?t Augsburg, Germany A. Kandel, University of South Florida, USA J. P. Bigus, IBM T. J. Watson Research Center, USA J. Liu, Hong Kong Baptist University, Hong Kong A. Namatame, National Defense Academy, Japan K. Sycara, Carnegie Mellon University, USA B. Kosko, University of Southern California, USA T. Baeck, Informatic Centrum Dortmund, Germany K. Hirota, Tokyo Institute of Technology, Japan E. Oja, Helsinki University of Technology, Finland H. R. Berenji, NASA Ames Research Center, USA H. Liljenstrom, Royal Institute of Technology, Sweden A. Bulsari, AB Nonlinear Solutions OY, Finland J. Fernandez de Ca?ete, University of Malaga, Spain W. Duch, Nicholas Copernicus University, Poland E. Tulunay, Middle East Technical University, Turkey C. Kuroda, Tokyo Institute of Technology, Japan T. Yamakawa, Kyushu Institute of Technology, Japan J. Liu, Hong Kong Baptist University, Hong Kong A. Namatame, National Defense Academy, Japan A. Aamodt, Norwegian University of Science & Technology, Norway International Liaison: Canada and USA Liaison: Robert John, De Montfort University, UK Nasser Jazdi, Institut f?r Automatisierungs- und Softwaretechnik, Germany Europe Liaison: Dr. Eng. Djamel Khadraoui, Centre de Recherche Public, Luxembourg Frank Zimmer, SES ASTRA, Luxembourg Asia Liaison: Renzo Gobbin, University of Canberra, Australia R. Amin Sarker, ADFA, Australia Local Arrangements and Public Relation: Zohreh Pahlavani, AVIP, Austria C Meier, Australia Publicity: C. Meier, Australia Zohreh Pahlavani, AVIP, Austria Publication: Masoud Mohammadian, Australia In cooperation with: University of Canberra, (Masoud Mohammadian) Universidad Carlos III de Madrid, (Jos?-Luis Fern?ndez-Villaca?as Mart?n) University of Guelph, (Simon X. Yang) From maarten at science.uva.nl Fri Jun 24 03:41:16 2005 From: maarten at science.uva.nl (maarten at science.uva.nl) Date: Fri, 24 Jun 2005 09:41:16 +0200 Subject: [ACT-R-users] Measures of semantic distance In-Reply-To: References: Message-ID: <42BBB91C.2010901@science.uva.nl> Maybe you find the following general approach to semantic distance interesting: http://www.newscientist.com/channel/info-tech/mg18524846.100 http://www.arxiv.org/abs/%20cs.CL/0412098 - maarten van someren Wayne Gray wrote: > Friends, > > For those of you new to the topic of measures of semantic distance > (MSDs) I can recommend a short and very well-written discussion of how > the major systems differ: > > Lemaire, B., & Denhi?re, G. (2004). Incremental construction of an > associative network from a corpus. In K. D. Forbus & D. Gentner & T. > Regier (Eds.), 26th Annual Meeting of the Cognitive Science Society, > CogSci2004. Hillsdale, NJ: Lawrence Erlbaum Publisher. > > > There is also a very nice introduction to PMI-IR written by Peter D. > Turney and published as a Lecture Note inn Computer Science. "Mining the > Web for synonyms: PMI-IR versus LSA on TOEFL." > > Unfortunately I do not have the complete citation on this one. > > Wayne