Title :
An association of simulated annealing and electrical simulator SPICE-PAC for learning of analog neural networks
Author :
Berthiau, G. ; Durbin, F. ; Haussy, J. ; Siarry, P.
Author_Institution :
Dept. Electron. et Instrum. Nucleaire, CEA, Gif-sur-Yvette, France
Abstract :
Simulated annealing adapted to continuous variables is used to determine the synaptic coefficients of an analog multilayer neural network, approximating any continuous function of one or several variables. The open electrical simulator SPICE-PAC driven by simulated annealing produces a globally optimal set of synaptic weights, in a reasonable time and without requiring heavy and inaccurate gradient computations. The authors illustrate and improve the weights-tuning strategy through two simple examples
Keywords :
SPICE; analogue processing circuits; circuit optimisation; learning (artificial intelligence); multilayer perceptrons; neural chips; simulated annealing; analog neural networks; electrical simulator SPICE-PAC; globally optimal set; learning; multilayer neural network; simulated annealing; synaptic coefficients; synaptic weights; weights-tuning strategy; Analog circuits; Circuit optimization; Circuit simulation; Computational modeling; Feedforward neural networks; Instruments; Multi-layer neural network; Neural networks; Resistors; Simulated annealing;
Conference_Titel :
Design Automation, 1993, with the European Event in ASIC Design. Proceedings. [4th] European Conference on
Conference_Location :
Paris
Print_ISBN :
0-8186-3410-3
DOI :
10.1109/EDAC.1993.386467
