DocumentCode
1458396
Title
On the infeasibility of training neural networks with small mean-squared error
Author
Vu, Van H.
Author_Institution
Dept. of Math., Yale Univ., New Haven, CT, USA
Volume
44
Issue
7
fYear
1998
fDate
11/1/1998 12:00:00 AM
Firstpage
2892
Lastpage
2900
Abstract
We demonstrate that the problem of training neural networks with small mean-squared error is computationally intractable. This answers a question posed by Jones (1997)
Keywords
learning (artificial intelligence); mean square error methods; neural nets; neural networks training; small mean-squared error neural networks; Computer errors; Computer networks; Computer science; Interpolation; Mathematics; NP-hard problem; Neural networks; Polynomials; Seminars; Springs;
fLanguage
English
Journal_Title
Information Theory, IEEE Transactions on
Publisher
ieee
ISSN
0018-9448
Type
jour
DOI
10.1109/18.737520
Filename
737520
Link To Document