An analog continuous-time micropower Hopfield net
- 17 December 2002
- proceedings article
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 3, 1860-1865
- https://doi.org/10.1109/icnn.1994.374441
Abstract
Presents the building blocks for an analog continuous-time micropower CMOS Hopfield associative memory with local adaptation and storage of weights, a synaptic connection circuit, an analog memory, a memory programming circuit and an artificial neuron. The building blocks are modular, allowing a net of any size to be built. The weight adaption is by the outer product rule. The analog memory works by exposure to UV-light. A six neuron net, intended for storage and retrieval of two patterns, was implemented in a standard 2μm CMOS processKeywords
This publication has 6 references indexed in Scilit:
- UV-activated conductances allow for multiple time scale learningIEEE Transactions on Neural Networks, 1993
- Analog VLSI Implementation of Neural SystemsPublished by Springer Nature ,1989
- Parallel Distributed ProcessingPublished by MIT Press ,1986
- Neurons with graded response have collective computational properties like those of two-state neurons.Proceedings of the National Academy of Sciences, 1984
- Neural networks and physical systems with emergent collective computational abilities.Proceedings of the National Academy of Sciences, 1982
- Photoemission of Electrons from Silicon into Silicon DioxidePhysical Review B, 1965