McCLELLAND, James JAY

<p class="default">Link is to Jay's web page at Stanford.</p>

<p>

Note: NO private info here.</p>

<p>

&nbsp;

PhD in 1975 from Univ of Penn; &nbsp;then served on faculty

&nbsp;of UCSD before going to CMU in 1984.

Jay was at CMU from 1984 until 2006 &nbsp;(22 years) &nbsp;before moving to Stanford in 2006.</p>

<p>

Nov 12, 2009 5:44 PM

See his 2004 book on Semantic Cognition: a PDP approach &nbsp;.</p>

<p>

PDP activity of neural populations. &nbsp;

Neocortex does semantic learning - very gradual.

The HC is needed to complement the NC -&nbsp;

theHC provides a mech for rapid learning of specific arbitrary pattern.

Results of this rapid learning are grad integrated into the neocortical system -

this accts for the retrograde amnesia seen in HC deficient pts.

Kids learn categories first; then subcategories, then sub sub categories;

In dementia this pattern reverses.

However, note that BIRD is more accessible than either Robin or Animal.

• &nbsp;see his work with Rogers (former student) that integrates this into a single model
  

that reflects interplay btwn frequency and concept similarity in determining the

semantic reps of conecpts and ease of assigning name to them.&nbsp;</p>

<p>

&nbsp;&nbsp;&nbsp;&nbsp;Topic II: (Jay's Lab interests) &nbsp;Dynamics of Decision Making

&nbsp;&nbsp;building on work with Marius Usher

(Usher and Mcclelland 01).&nbsp;

Perspective is that Human Decision Stes simlutaneously exhibit both

categorical and discrete characteristics. &nbsp;One state of another gets favored as evidence accumulates...

state is graded and can switch depending on evidence.

This work is linked to NEWSOME primate neurophys lab at Stfd; &nbsp;Jon Cohen and Phil HOlmes at Princeton etc.</p>

<p>

&nbsp;&nbsp;&nbsp;&nbsp;Last listed interest: &nbsp;Language.&nbsp;

&nbsp;NLP has far more graded structure. &nbsp;Collaboration with

Joan Bybee.&nbsp;</p>

<p>

&nbsp;&nbsp;&nbsp;&nbsp;PhD &nbsp;from U Penn in 1975; &nbsp;faculty at UCSD; &nbsp;to &nbsp;CMU in 1984

co-founder of Center for Neural Basis of Cognition (Pitt and CMU) &nbsp;til &nbsp;2006 then moved to Stfd

Prof of Psych and foudner of Center for MBC.</p>

<p>

Member of the NAS; &nbsp;APS Wm James Award; &nbsp;Pres Elect of the Federation of the Behav, Psych,a dn Cog Sci.</p>

<p>

&nbsp;</p>

<p>

10:04 AM, November 21, 2007&nbsp;</p>

<p>

&nbsp;best known for Parallel Distrib Processing...

connectionism; &nbsp;neural networks.</p>

<p>

fantastic - Jay Mcclelland is now in town at Stanford Psych Dept!

(for years he was David Rummelharts collaborator.&nbsp;</p>