Abstract
A biologically motivated artificial neural network
The Biological bulletin (Lancaster), Vol.175(2), p.314
01/01/1988
Abstract
Most artificial neural networks utilize neuronal elements whose presynaptic strengths are modulated by the output, and are set by a number of different iterative, non-linear, or stochastic algorithms. The network described here, a dynamically stable associative learning (DYSTAL) network, is based on a biological neural network: the convergent visual and vestibular pathways which mediate associative learning of Hermissenda crassicornis . The DYSTAL network displays a number of desirable features. The network is self-adapting and the strength (weight) associated with each synapse is adjusted by a rule that only requires information regarding the pre- and post-synaptic neurons involved. Finally, the network can associate different patterns which are presented sequentially.
Details
- Title: Subtitle
- A biologically motivated artificial neural network
- Creators
- Daniel AlkonKim Tiplitz BlackwellTom P VoglVassilios Kountouris
- Resource Type
- Abstract
- Publication Details
- The Biological bulletin (Lancaster), Vol.175(2), p.314
- ISSN
- 0006-3185
- eISSN
- 1939-8697
- Comment
- DOI:10.1086/BBLv175n2p301. Abstracts of Papers Presented at the General Scientific Meetings of the Marine Biological Laboratory. The Biological Bulletin 1988 175:2, 301-319
- Language
- English
- Date published
- 01/01/1988
- Academic Unit
- Roy J. Carver Department of Biomedical Engineering; Iowa Neuroscience Institute
- Record Identifier
- 9984447841402771
Metrics
24 Record Views