Theme- 2 前川班 Team MAEKAWA

Algorithm development of χlogbot,which measures and intervenes autonomously

Conventional research on animal behavioral analysis has separated ecological and neurological studies on data measurement in physical space and information science studies on data analysis in cyberspace; therefore, the cycle of the measurement and analysis had to be repeatedly conducted until the whole aspect of the target behavior was clarified, presenting limitations to the methods. To address such problems, we will (1) research measurement and intervention methods using autonomous artificial intelligence, and (2) develop χlogbot hardware equipped with intervention methods.

For (1), we will work on reinforcement learning techniques to formulate intervention strategies to elucidate the hidden behavioral mechanisms (behavioral diagrams) of animals. In collaboration with each group, we will develop AI that intervenes and observes out of curiosity, and DeepHLχ that automatically searches for intervention methods to distinguish the differences among the behavior of comparison groups (e.g., diseased and healthy individuals), targeting various animals (e.g., mice, nematodes, fish, and insects). For study (2), we will collaborate with Team Hashimoto in A02-1 to develop the software for animal-borne χlogbot (for seabirds, bats, etc.), and for the development of the environmental χlogbot (for nematodes, mice, ants, etc.), we will work with Team Makino in A02-4 and Team Nishimori in A01-4, and then integrate them with study (1).

• 

A02-2 前川班 研究者

• 

  Principal Investigator Takuya Maekawa Takuya Maekawa

  Graduate School of Information Science and Technology, Osaka University Associate Professor

  Research field

  Ubiquitous Computing

  Research keywords

  Wearable/ubiquitous sensing, activity recognition, behavior analysis, biologging, indoor positioning

  Laboratory

  http://www-mmde.ist.osaka-u.ac.jp/~maekawa/index-e.html

• 

  Co-Investigator Kotaro Kimura Kotaro Kimura

  Graduate School of Science, Nagoya City University Professor

  Research field

  Neurobiology

  Research keywords

  sensory behavior, sensory information processing, imaging, nematode, optogenetics

  Laboratory

  https://en.kokimura-lab.org

• 

  Co-Investigator Koh Takeuchi Koh Takeuchi

  Graduate School of Informatics, Kyoto University Assistant Professor

  Research field

  Machine Learning, Data Mining

  Research keywords

  Spatio-Temporal Data, Pattern Extraction, Causal Inference

  Laboratory

  https://koh-t.github.io/index.html