Design of Neural Network with Autonomy and its Circuit Implementation

2016 Fiscal Year Final Research Report

• Project/Area Number: 26540127
• Research Category: Grant-in-Aid for Challenging Exploratory Research
• Allocation Type: Multi-year Fund
• Research Field: Soft computing
• Research Institution: The University of Tokushima
• Principal Investigator: Yoko Uwate 徳島大学, 大学院理工学研究部, 准教授 (80582642)
• Project Period (FY): 2014-04-01 – 2017-03-31
• Keywords: ニューラルネットワーク / 脳情報処理 / 複雑系ネットワーク / 自律性 / 同期現象

Outline of Final Research Achievements

In this study, we investigate autonomy when applying new brain information processing function clarified by neuroscience to artificial neural network. Recently, it has been clarified that firing pattern of neuron group shows polygonal structure in cognitive process. Therefore, we investigate the case where the advanced oscillator neuron model is combined with a polygonal shape.
Next, as an approach toward circuit implementation of a neural network, consider a system in which a neuron model is replaced with a one-dimensional map. Furthermore, we implement the circuit of the proposed neural network. By realizing it in a circuit, enormous data can be processed faster and various applications are expected.

Free Research Field

非線形回路工学