Jun Izawa: “Slow dynamics of motor memory”
12:15 〜 13:00
CiNet 1F Conference Room
Associate Professor, Faculty of Engineering
Information and Systems Ph.D. Program in Empowerment Informatics
University of Tsukuba
Host PI: Hiroshi Imamizu
For machine learning, modulating learning-parameters (e.g. learning rate) optimally for given environments and tasks is critical to stabilize and accelerate the learning, which we call “meta-learning”. Does the human brain have any identical ability? To ask this question, we let the human subjects to adapt for making straight movements under the two conflicting force environments over the consecutive two days and estimated the learning parameters of each of the seven adaptation blocks (conflicting group). These parameters were compared with those estimated from the other subjects’ group who experienced the same adaptation paradigms under the uniform force fields (uniform group). For this talk, I will present a preliminary data of this experiment. So far, we observed the tendency that, in the conflicting group, the adaptation rate for each force environment increased within one day whereas the interaction between two memories were decreased. In contrast, of the uniform group, whereas the learning rate increased, the interaction term increased. We found that the human meta-learning lead efficient motor adaptation for the given environment.
The Friday Lunch Seminar is CiNet's main regular meeting series, held every week at 12:15 in the beautiful main lecture theatre on the ground floor at CiNet. The talks are typically 40mins long and orientated towards an inter-disciplinary audience. They are informal, social, and most people bring their own lunch to eat during the talk. They are open to anyone who is feeling curious and wants to come, regardless of where you work.