© The Institution of Electrical Engineers
Results are presented demonstrating analogue neural network computation using VLSI pulse-stream devices. Synapse weight storage is dynamic, and the multiply and add functions are performed by a switched capacitor circuit. Used on a mobile robot localisation task, the neural VLSI devices achieve results within 1.2% of those of a SUN workstation.
References
-
-
1)
-
A.F. Murray ,
A.V.W. Smith ,
L. Tarassenko ,
J.G. Delgado-Frias ,
W.R. Moore
.
(1989)
Fully-programmable analogue VLSI devices for the implementation of neural networks, VLSI for artificial intelligence.
-
2)
-
Hubbard, W., Schwartz, D., Denker, J., Graf, H.P., Howard, R., Jackel, L., Straughn, B., Tennant, D.: `Electronic neural networks', Proc. AIP Conf. Neural Networks for Computing, 1986, Snowbird, p. 227–233.
-
3)
-
Murray, A.F., Hamilton, A., Reekie, H.M., Tarassenko, L.: `Pulse-stream arithmetic in programmable neural networks', Proc. Int. Symp. on Circuits and Systems, 1989, Portland, Oregon, p. 210–1212.
-
4)
-
Tarassenko, L., Brownlow, M.J., Murray, A.F.: `VLSI neural networks for autonomous robot navigation', Proc. Int. Neural Network Conf., 1990, Paris.
-
5)
-
Sivilotti, M.A., Emerling, M.R., Mead, C.A.: `VLSI architectures for implementation of neural networks', Proc. AIP Conf. Neural Networks for Computing, 1986, p. 408–413.
-
6)
-
A.F. Murray ,
A. Hamilton ,
L. Tarassenko ,
D.S. Touretzky
.
(1989)
Programmable analog pulse-firing networks, Advances in neural information processing systems.
-
7)
-
A.F. Murray ,
A.V.W. Smith
.
Asynchronous arithmetic for VLSI neural systems.
Electron. Lett.
,
642 -
643
-
8)
-
J. Millman ,
A. Grabel
.
(1987)
, Microelectronics.
-
9)
-
Graf, H.P., Jackel, L.D., Howard, R.E., Straughn, B., Denker, J.S., Hubbard, W.E., Tennant, D.M., Schwartz, D.: `VLSI implementation of a neural network memory with several hundreds of neurons', Proc. AIP Conf. on Neural Networks for Computing, 1986, Snowbird, p. 182–187.
-
10)
-
H.P. Graf ,
L.D. Jackel ,
W.E. Hubbard
.
VLSI implementation of a neural network model.
IEEE Computer
,
41 -
49
-
11)
-
M. Verleysen ,
B. Sirletti ,
A.M. Vandemeulebroecke ,
P.G.A. Jespers
.
Neural networks for high-storage content-addressable memory: VLSI circuit and learning algorithm.
IEEE J. Solid-State Circuits
,
562 -
569
-
12)
-
Boahen, K.A., Andreou, A.G., Pouliquen, P.O., Pavasovic, A.: `Architectures for associative memories using current-mode analog MOS circuits', Advanced Research in VLSI, Proc. Stanford Conference, 1989, p. 175–194.
http://iet.metastore.ingenta.com/content/journals/10.1049/el_19900834
Related content
content/journals/10.1049/el_19900834
pub_keyword,iet_inspecKeyword,pub_concept
6
6