Continuous-valued probabilistic behavior in a VLSI generative model

Author

Hsin Chen ; Fleury, P.C.D. ; Murray, A.F.

Author_Institution

Dept. of Electr. Eng., Nat. Tsing-Hua Univ., Hsin-Chu

Volume

17

Issue

3

fYear

2006

fDate

5/1/2006 12:00:00 AM

Firstpage

755

Lastpage

770

Abstract

This paper presents the VLSI implementation of the continuous restricted Boltzmann machine (CRBM), a probabilistic generative model that is able to model continuous-valued data with a simple and hardware-amenable training algorithm. The full CRBM system consists of stochastic neurons whose continuous-valued probabilistic behavior is mediated by injected noise. Integrating on-chip training circuits, the full CRBM system provides a platform for exploring computation with continuous-valued probabilistic behavior in VLSI. The VLSI CRBM´s ability both to model and to regenerate continuous-valued data distributions is examined and limitations on its performance are highlighted and discussed

Keywords

Boltzmann machines; VLSI; learning (artificial intelligence); stochastic processes; VLSI generative model; continuous restricted Boltzmann machine; continuous-valued data distributions; continuous-valued probabilistic behavior; hardware-amenable training algorithm; injected noise; on-chip training circuits; probabilistic generative model; stochastic neurons; Bioelectric phenomena; Circuit noise; Embedded computing; Embedded system; Intelligent systems; Neurons; Power system modeling; Stochastic resonance; Very large scale integration; Working environment noise; Boltzmann machine; continuous-valued probabilistic VLSI; noise; on-chip training; probabilistic generative model; stochastic computation; Algorithms; Artificial Intelligence; Computer Simulation; Equipment Design; Equipment Failure Analysis; Models, Statistical; Neural Networks (Computer); Pattern Recognition, Automated; Semiconductors; Signal Processing, Computer-Assisted;

fLanguage

English

Journal_Title

Neural Networks, IEEE Transactions on

Publisher

ieee

ISSN

1045-9227

Type

jour

DOI

10.1109/TNN.2006.873278

Filename

1629097