Establishing mutual communication between organism and machine using deep learning

2019 Fiscal Year Final Research Report

• Project/Area Number: 17K00329
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Soft computing
• Research Institution: Hokkaido University
• Principal Investigator: Iizuka Hiroyuki 北海道大学, 情報科学研究院, 准教授 (30396832)
• Co-Investigator(Kenkyū-buntansha): 山本 雅人 北海道大学, 情報科学研究院, 教授 (40292057)
• Project Period (FY): 2017-04-01 – 2020-03-31
• Keywords: 生物機械間相互作用 / 双方向相互作用 / メダカ

Outline of Final Research Achievements

We constructed a mutual interaction system between machine and fish (medaka). The system can generate and show visual stimuli of machine movements in response to the movements of the real medaka which are detected in real time. Using this experimental apparatus, we compared the real medaka behaviors to the artificial movements generated by reactive (mutual) or non-reactive (unidirectional) way. Real medaka were more interested in reactive artificial movements and spent more time on the screen showing the reactive artificial movements than non-reactive one. In other words, it was possible to induce specific behaviors in real medaka under the mutual interaction. It can be said that the communication-like interaction between the artifact and the medaka has been realized.

Free Research Field

人工生命

Academic Significance and Societal Importance of the Research Achievements

今までの人工物を用いた生物の振る舞いを明らかにする実験では，本当の生物の動きを記録したものを再生するか，あらかじめ人工的に作られた動きを再生するプレイバック実験しか行われてこなかった．これは我々の行った実験のnon-reactive条件に相当する．それに対して，我々の実験では，リアルタイムに生物の動きを計測し，瞬時に人工物の動きを生成することで，人工物と生物の間に双方向相互作用を成立させることに成功した．結果としては，生物は双方向と単方向の相互作用で異なる振る舞いを見せ，生物の新たな実験ステージへと開いた．