[CaCL] Bayesian modelling of cognition: pros and cons

[william schuler]

readings for next cacl:

> Psychol Bull. 2012 May;138(3):415-22. doi: 10.1037/a0026884.
>    How the Bayesians got their beliefs (and what those beliefs actually
>    are): comment on Bowers and Davis (2012).
>    Griffiths TL1, Chater N, Norris D, Pouget A.
>     http://www.yangzhiping.com/files/pubs/415.pdf
> 
>    Psychol Bull. 2012 May;138(3):423-6. doi: 10.1037/a0027750.
>    Is that what Bayesians believe? reply to Griffiths, Chater, Norris,
>    and Pouget (2012).
>    Bowers JS1, Davis CJ.
>     http://clm.utexas.edu/compjclub/papers/Bowers2012b.pdf



On Apr 1, 2014, at 7:20 AM, william schuler <schulerw at gmail.com> wrote:

> 
> ok, with one vote each, i'll break the tie in favor of the bayesian papers that mike suggested (since we may be able to do spectral learning in 5702), so lets read the following for wednesday:
> 
>>> Psychol Bull. 2012 May;138(3):389-414. doi: 10.1037/a0026450.
>>> Bayesian just-so stories in psychology and neuroscience.
>>> Bowers JS1, Davis CJ.
>>> http://www.yangzhiping.com/files/pubs/389.pdf
> 
> 
> wm
> 
> Sent from my iPad
> 
> On Mar 31, 2014, at 7:50 AM, william schuler <schulerw at gmail.com> wrote:
> 
>> hi all:
>> 
>> we have a couple options for cacl this week.
>> 
>> 1. mike has suggested a series of three philosophical papers on the merits of bayesian modeling for cognitive science (or lack thereof, because it's purely defined at the computational level), see below.
>> 
>> 2. i found a few short but rather mathematical papers on spectral learning (trendy new type of unsupervised learning using pca), starting with the paper below, which i think we could handle given what we learned about linear algebra in 5702.
>> 
>> A Spectral Algorithm for Learning Hidden Markov Models
>> http://www.cs.mcgill.ca/~colt2009/papers/011.pdf
>> 
>> also, if we want to keep meeting in may, we could do both.
>> 
>> please let me know your thoughts asap; i'd like to tally votes by tuesday so we have time to read.
>> 
>> wm
>> 
>> 
>> Sent from my iPad
>> 
>> Begin forwarded message:
>> 
>>> From: Michael White <mwhite at ling.osu.edu>
>>> Date: March 26, 2014 3:20:05 PM EDT
>>> To: William Schuler <schuler at ling.osu.edu>
>>> Subject: Fwd: [ProbModels] [AdHocPapers] Bayesian modelling of cognition: pros and cons
>>> 
>> 
>>> fyi
>>> 
>>> ---------- Forwarded message ----------
>>> From: Sharon Goldwater <sgwater at inf.ed.ac.uk>
>>> Date: Tue, Mar 18, 2014 at 6:42 AM
>>> Subject: [ProbModels] [AdHocPapers] Bayesian modelling of cognition: pros and cons
>>> To: statsnlp at inf.ed.ac.uk
>>> 
>>> 
>>> So, as I mentioned there are several recent back-and-forth
>>> opinion/argument articles on this recently.  After some consultation
>>> with Chris and Stella it looks like we'll go with this set:
>>> 
>>> Psychol Bull. 2012 May;138(3):389-414. doi: 10.1037/a0026450.
>>> Bayesian just-so stories in psychology and neuroscience.
>>> Bowers JS1, Davis CJ.
>>> http://www.yangzhiping.com/files/pubs/389.pdf
>>> 
>>> Psychol Bull. 2012 May;138(3):415-22. doi: 10.1037/a0026884.
>>> How the Bayesians got their beliefs (and what those beliefs actually
>>> are): comment on Bowers and Davis (2012).
>>> Griffiths TL1, Chater N, Norris D, Pouget A.
>>> http://www.yangzhiping.com/files/pubs/415.pdf
>>> 
>>> Psychol Bull. 2012 May;138(3):423-6. doi: 10.1037/a0027750.
>>> Is that what Bayesians believe? reply to Griffiths, Chater, Norris,
>>> and Pouget (2012).
>>> Bowers JS1, Davis CJ.
>>> http://clm.utexas.edu/compjclub/papers/Bowers2012b.pdf
>>> 
>>> These are long. My suggestion is to focus for this week on Parts 1-3
>>> of the first paper (up to p. 403) and the corresponding
>>> counterarguments in the second paper (to top of p. 418 and maybe the
>>> 'Merits' section also).  If we want we can then continue with the rest
>>> for next week (or not).
>>> 
>>> This is still a lot of reading so don't leave it to the last minute.
>>> 
>>> --
>>> The University of Edinburgh is a charitable body, registered in
>>> Scotland, with registration number SC005336.
>>> _______________________________________________
>>> statsnlp mailing list
>>> statsnlp at inf.ed.ac.uk
>>> http://lists.inf.ed.ac.uk/mailman/listinfo/statsnlp
>>> 
>>> The University of Edinburgh is a charitable body, registered in
>>> Scotland, with registration number SC005336.
>>> 
>>> 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osu.edu/pipermail/cacl/attachments/20140409/3e0e853c/attachment.html>