15 search results - page 3 / 3 » Weight initialization methods for multilayer feedforward |
141
click to vote
DATE
2008
IEEE
15 years 9 months ago
2008
IEEE
This paper presents a novel architecture for on-chip neural network training using particle swarm optimization (PSO). PSO is an evolutionary optimization algorithm with a growing ...
245
click to vote
ICASSP
2011
IEEE
14 years 6 months ago
2011
IEEE
Deep Belief Networks (DBNs) are multi-layer generative models. They can be trained to model windows of coefficients extracted from speech and they discover multiple layers of fea...
196
click to vote
JMLR
2006
15 years 3 months ago
2006
This paper introduces a learning method for two-layer feedforward neural networks based on sensitivity analysis, which uses a linear training algorithm for each of the two layers....
132
click to vote
STOC
1993
ACM
15 years 7 months ago
1993
ACM
Abstract. It is shown that high-order feedforward neural nets of constant depth with piecewisepolynomial activation functions and arbitrary real weights can be simulated for Boolea...
135
click to vote
NECO
2007
15 years 2 months ago
2007
We present a hierarchical architecture and learning algorithm for visual recognition and other visual inference tasks such as imagination, reconstruction of occluded images, and e...