Adaptively setting analog weights in a neural network and the li

Details Adaptively setting analog weights in a neural network and the li Adaptively setting analog weights in a neural network and the li

1992-07-06

1993-09-21

MacDonald, Allen H.

Boots, shoes, and leggings

364807, H03K 1900, H03K 1921

Patent

active

052476065

ABSTRACT:
A method for adaptively setting analog weights in analog cells of a neural network and the like. The process starts by addressing a synapse cell in the network. A target weight for said addressed synapse cell is selected, and the current weight present on the synapse cell is measured. The amplitude and duration of a voltage pulse to be applied to said synapse cell to adjust said synapse cell in the direction of said target weight is calculated using a set of coefficients representing the the physical characteristics of the synapse cell. The voltage pulse is applied to the addressed synapse cell and the new weight of the synapse cell is re-measured. If the synapse cell weight is within acceptable limits of the target weight, the values of the coefficients are saved and the next adjacent synapse cell is addressed until all synapse cells are set. If the synapse cell is not within acceptable limits, new values for the coefficients are calculated in relation to the re-measured weight. A new voltage pulse is generated and applied to the synapse cell. The process is repeated until the weight of the synapse cell is set within an acceptable limit of the target weight.

REFERENCES:
patent: 4660166 (1985-01-01), Hopfield
patent: 4760437 (1986-01-01), Denker et al.
patent: 4782460 (1987-04-01), Spencer
patent: 4802103 (1986-01-01), Faggin et al.
patent: 4903226 (1990-02-01), Tsiuidis
patent: 4904881 (1990-02-01), Castro
patent: 4947482 (1990-08-01), Brown
patent: 4950917 (1990-08-01), Holler et al.
patent: 4951239 (1990-08-01), Andes et al.
patent: 4956564 (1990-09-01), Holler et al.
patent: 4961002 (1990-10-01), Tam et al.
patent: 4962342 (1990-10-01), Mead et al.
patent: 5021693 (1991-06-01), Shima
patent: 5021988 (1991-06-01), Mashiko
patent: 5028810 (1991-07-01), Castro
patent: 5053638 (1991-10-01), Furutani et al.
patent: 5059814 (1991-10-01), Mead et al.
patent: 5059920 (1991-10-01), Anderson et al.
patent: 5148045 (1992-09-01), Oyanagi
Holler et al., "An Electrically Trainable Artificial Neural Network (ETANN) with 10240 Floating Gate", Proc. Int'l Annual Conf. on Neural Networks, Jun. 18-22, 1989, II-191-6.
Hansen et al., "A Time-Multiplexed Switched-Capacitor Circuit for Neural Network Applications", IEEE Int'l Symposium on Circuits and Systems, May 8-11, 1989 vol. 3, pp. 2177-2180.
Murray, "Pulse Arithmetic in VLSI Neural Networks", IEEE Micro, vol. 9 No. 6, Dec. 1989, pp. 67-74.
"VLSI For Artificial Intelligence", edited by Jose G. DelGado-Frias and Will R. Moore, Kluwer Academic Publishers, 1989, pp. 230-233.
"Programmable Analog Synapses for Micro Electronic Neural Networks Using A Hybrid Digital-Analog Approach", by F. J. Mack et al., IEEE International Conference on Neural Networks, Jul. 24-27, San Diego, CA.
"A Pipelined Associative Memory Implemented in VLSI", by Clark et al., IEEE Journal of Solid-State Circuits, vol. 24, No. 1, pp. 28-34, Feb. 1989.

Affiliated with

Also associated with

Rating

Adaptively setting analog weights in a neural network and the li does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Adaptively setting analog weights in a neural network and the li, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Adaptively setting analog weights in a neural network and the li will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-1056535