We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

(license)

Current browse context:

cond-mat
< prev | next >
new | recent | 0201

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo

Condensed Matter > Disordered Systems and Neural Networks

Title: Statistical Physics of Feedforward Neural Networks

Authors: Robert Urbanczik
(Submitted on 29 Jan 2002)
Abstract: The article is a lightly edited version of my habilitation thesis at the University Wuerzburg. My aim is to give a self contained, if concise, introduction to the formal methods used when off-line learning in feedforward networks is analyzed by statistical physics. However, due to its origin, the article is not a comprehensive review of the field but is highly skewed towards reporting my own research.
Comments: 44 pages
Subjects: Disordered Systems and Neural Networks (cond-mat.dis-nn); Quantitative Biology (q-bio)
Cite as: arXiv:cond-mat/0201530 [cond-mat.dis-nn]
  (or arXiv:cond-mat/0201530v1 [cond-mat.dis-nn] for this version)

Submission history

From: Robert Urbanczik [view email]
[v1] Tue, 29 Jan 2002 11:28:15 GMT (30kb)
Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Link back to: arXiv, form interface, contact.