Host: The Japan Society of Mechanical Engineers
Name : [in Japanese]
Date : June 28, 2023 - July 01, 2023
The synchronization phenomenon refers to the gradual synchronization of simple elements with different rhythms due to mutual influence and is observed across various fields, e.g., natural sciences to humanities and social sciences. Synchronization can be observed at various levels in the human brain, for example, epilepsy due to the abnormal synchronous firing of neurons, resulting in functional behaviors. Predicting and analyzing synchronization dynamics could be important for inducing or preventing functional behaviors of synchronization. However, most studies on synchronization have focused on complex oscillator systems, and few have predicted and analyzed the synchronization of basic pendulums using machine learning. In this study, we reproduce the synchronization of pendulums in a physics simulation with an easy condition setting and use machine learning to predict how the dynamics change under various conditions. We investigate which machine learning model is best for predicting synchronization phenomena.
