2022 Fiscal Year Final Research Report
Physical reservoir computing with the brain
| Project/Area Number |
20H04252
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Review Section |
Basic Section 61040:Soft computing-related
|
|---|
| Research Institution | The University of Tokyo |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2020-04-01 – 2023-03-31
|
| Keywords | 脳 / 情報処理 / 神経活動 |
|---|
| Outline of Final Research Achievements |
Reservoir computing has attracted attention as a training algorithm for large-scale recurrent neural networks. In particular, physical reservoir computing, which utilizes arbitrary large-degree-of-freedom dynamical systems as computational resources, has recently become an important research topic. In this study, we demonstrated physical reservoir computing with the brain and quantified the amount of computational resources in the brain. In dissociate culture of neurons and the auditory cortex of rats for physical reservoir, we estimated the information processing capacity. We also investigated whether self-organized neural circuits could improve the information processing capacity. This study verified the possibility that self-organization and plasticity in the brain plays critical roles in reservoir computing.
|
| Free Research Field |
神経工学
|
| Academic Significance and Societal Importance of the Research Achievements |
脳の情報処理をリザバー計算の枠組みで考察する研究が盛んであるが，脳そのものを物理リザバー計算に用いた研究はほとんどない．本研究の学術的独自性は，脳にリザバー計算が実装されているという仮説を検証する構成論的な試みにある．本研究の学術的創造性は，リザバー計算という概念により，人工ニューラルネットワークを用いたモデル研究と，実際の脳の生理学的研究とを結びつける学際性にある．このような学際性は，古くはパーセプトロンと小脳の学習機構 (1980年代)，コネクショニズムと海馬・新皮質からなる学習機構 (1990年代)，そして最近では深層学習と脳の階層的表現などのように，革新的な研究の原動力となってきた．
|
