A Silicon Primitive for Competitive Learning (2000)
| Venue: | in Advances in Neural Information Processing Systems 13 |
| Citations: | 3 - 2 self |
BibTeX
@INPROCEEDINGS{Hsu00asilicon,
author = {David Hsu and Miguel Figueroa Chris Diorio and Miguel Figueroa and Chris Diorio},
title = {A Silicon Primitive for Competitive Learning},
booktitle = {in Advances in Neural Information Processing Systems 13},
year = {2000},
pages = {713--719},
publisher = {MIT Press}
}
Years of Citing Articles
Bookmark
 |
 |
 |
 |
 |
 |
OpenURL
Abstract
Competitive learning is a technique for training classification and clustering networks. We have designed and fabricated an l 1transistor primitive, that we term an automaximizing bump circuit, that implements competitive learning dynamics. The circuit performs a similarity computation, affords nonvolatile storage, and implements simultaneous local adaptation and computation. We show that our primitive is suitable for implementing competitive learning in VLSI, and demonstrate its effectiveness in a standard clustering task.
Citations
No citations identified.
