A New Modulated Hebbian Learning Rule -Method for Local Computation of a Principal Subspace
This paper presents one possible implementation of the transformation that performs linear mapping to a lower-dimensional subspace. Principal component subspace will be the one that will be analyzed. Comparing to some other well-known methods for yielding principal component subspace (e.g. Ojas Subspace Learning Algorithm 9-12), newly proposed method has one feature which could be seen as desirable from the biological point of view -all calculations are performed locally. In the other words, synaptic efficacy learning rule does not need the explicit information about the value of the other efficacies to make single efficacy modification. Synaptic efficacies are modified by implementation of Modulated Hebb-type (MH) learning rule. Slightly modified MH algorithm that can be used for Linear ICA will be also introduced. Possible similarity with part of the retinal circuit will be presented.
Marko Jankovic
Control Department,Electrical Engineering Institute Nikola Tesla,Koste Glavinica 8a, 11000 Beograd, Yugoslavia
国际会议
8th International Conference on Neural Information Processing(ICONIP 2001)(第八届国际神经信息处理大会)
上海
英文
508-513
2001-11-14（万方平台首次上网日期，不代表论文的发表时间）