4255 search results - page 492 / 851 » On Learning Boolean Functions |
173
click to vote
EDM
2010
15 years 8 months ago
2010
Learning software is not designed for data analysis and mining. Because usage data is not stored in a systematic way, its thorough analysis requires long and tedious preprocessing....
175
click to vote
ICONIP
2008
15 years 8 months ago
2008
Abstract. While injecting fault during training has long been demonstrated as an effective method to improve fault tolerance of a neural network, not much theoretical work has been...
217
click to vote
PAKM
2000
15 years 8 months ago
2000
Acknowledging the "untidiness of knowledge work", we agree that organizational learning calls for flexible and adaptable IT support. However, recurring situations which ...
163
click to vote
NIPS
1996
15 years 8 months ago
1996
Real-valued random hidden variables can be useful for modelling latent structure that explains correlations among observed variables. I propose a simple unit that adds zero-mean G...
162
click to vote
HAIS
2008
Springer
15 years 7 months ago
2008
Springer
The widespread use of artificial neural networks and the difficult work regarding the correct specification (tuning) of parameters for a given problem are the main aspects that mot...