Title :
A formulation of mean-field approximation for layered feedforward stochastic networks
Author :
Tanaka, Toshiyuki
Author_Institution :
Dept. of Electr. & Inf. Eng., Tokyo Metropolitan Univ., Japan
Abstract :
A formulation of mean-field approximation for layered feedforward stochastic networks is presented. This formulation provides a framework to approximately treat not only averages but also correlations within each layer, so that it can be regarded as an extension to the formulation by Saul and Jordan (1998)
Keywords :
feedforward neural nets; stochastic processes; averages; correlations; layered feedforward stochastic networks; mean-field approximation; Joining processes; Paints; Probability; Statistics; Stochastic processes;
Conference_Titel :
Neural Information Processing, 1999. Proceedings. ICONIP '99. 6th International Conference on
Conference_Location :
Perth, WA
Print_ISBN :
0-7803-5871-6
DOI :
10.1109/ICONIP.1999.845652
