Mean-field theory of learning: from dynamics to statics

Details Mean-field theory of learning: from dynamics to statics Mean-field theory of learning: from dynamics to statics

2000-06-15

arxiv.org/abs/cond-mat/0006251v1

Physics

Condensed Matter

Disordered Systems and Neural Networks

13 pages, 5 figures, to appear in "Advanced Mean Field Methods - Theory and Practice", edited by M. Opper and D. Saad, MIT Pre

Scientific paper

Using the cavity method and diagrammatic methods, we model the dynamics of batch learning of restricted sets of examples. Simulations of the Green's function and the cavity activation distributions support the theory well. The learning dynamics approaches a steady state in agreement with the static version of the cavity method. The picture of the rough energy landscape is reviewed.

Affiliated with

Also associated with

No associations

Rating

Mean-field theory of learning: from dynamics to statics does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.

If you have personal experience with Mean-field theory of learning: from dynamics to statics, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Mean-field theory of learning: from dynamics to statics will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFWR-SCP-O-547030