Learning Improvement in Phase Oscillator Models

Please enable javascript in your browser.

Aghighi, Meysam | 2013

552 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 44132 (19)
University: Sharif University of Technology
Department: Computer Engineering
Advisor(s): Jalili, Mahdi
Abstract:
In the recent years, the problem of modeling a cognitive task using phase oscillators has been receiving a significant attention. In this view, single neurons are no longer elementary computational units. Rather, coherent oscillating groups of neurons are seen as nodes of networks performing cognitive tasks. From this assumption, we develop a model of stimulus-response learning and recognition. The most significant part of our work is defining learning methods for natural frequencies and coupling weights in a coupled phase oscillator network under Kuramoto conditions. In this thesis, we improved the previous models by not only emphasizing on the frequency of the oscillators but also taking into account the synaptic weights. We proposed a learning algorithm for the proposed model and applied it on a number of classification tasks including the Iris and Cars data sets. The results showed the superiority of the proposed model over the previous models in terms of classification rate. Also, the proposed model showed
Keywords:
Pattern Recognition ; Phase Oscillator ; Neural Oscillators Network ; Kuramoto Model

No TOC