2023 Fiscal Year Final Research Report
Frontiers in brain and life sciences on active information gain in an uncertain environment
Project Area | Frontiers in brain and life sciences on active information gain in an uncertain environment |
Project/Area Number |
21H05167
|
Research Category |
Grant-in-Aid for Transformative Research Areas (B)
|
Allocation Type | Single-year Grants |
Review Section |
Transformative Research Areas, Section (IV)
|
Research Institution | Nagoya University |
Principal Investigator |
|
Co-Investigator(Kenkyū-buntansha) |
雨森 賢一 京都大学, 高等研究院, 特定拠点准教授 (70344471)
本田 直樹 広島大学, 統合生命科学研究科(理), 教授 (30515581)
|
Project Period (FY) |
2021-08-23 – 2024-03-31
|
Keywords | 自由エネルギー原理 / 予測符号化 / 能動的推論 / 限定合理性 / David Marr / 神経回路 / 葛藤 / 認知 |
Outline of Final Research Achievements |
The free energy principle, which evolves under perfect rationality, has limitations in modeling living organisms, including humans. Unlike AI, which has unlimited computational resources, it is in principle difficult for living organisms with limited knowledge and computational capacity to achieve perfect rationality. We overcame the problems of the free energy principle and clarified how a system with bounded rationality can optimize predictions and generate behavioral strategies from the perspective of Marr's three levels: computational, algorithmic, and implementational levels. The core group was responsible for organizing the Osakada, Amemori, and Honda groups, promoting collaboration and the exchange of knowledge and techniques among them, and translating our research results into the real world. We invited advisors and researchers related to our research to regular group meetings and reflected their comments to the research and group management.
|
Free Research Field |
脳神経科学
|
Academic Significance and Societal Importance of the Research Achievements |
我々は、限定合理性のもとでの脳内計算原理の一端を明らかにし、柔軟で自律的な知性を実現するための手がかりを得た。神経疾患や精神疾患は、予測や葛藤の脳内情報処理のバランス異常として捉えることができる。本領域の研究は、神経疾患や精神疾患の診断法や新たな治療戦略の創出に貢献できると期待される。さらに、自らの予測のあいまいさに基づいて、目的志向の行動を切り替え、自律的・能動的に情報獲得を行う「あいまい脳」を有するAIの計算原理が導けると期待される。
|