Deep Learning for Classical Mechanics

Mugeon Song, Maverick S. H. Oh, Yongjun Ahn, Keun-Young Kim

Deep learning has been widely and actively used in various research areas. Recently, in the subject so-called gauge/gravity duality, a new deep learning technique which deals with classical equations of motion has been proposed. This method is a little different from standard deep learning techniques in the sense that not only do we have the right final answers but also obtain physical understanding of learning parameters. Building on this idea, we apply the deep learning technique to simple classical mechanics problems. The type of problems we address is how to find the unknown force, by the deep learning technique, only from the initial and final data sets. We demonstrate that our deep learning technique is successful for simple cases: one dimensional velocity or position-dependent force. In our opinion, this method has a big potential for wider applications to physics and computer science both in education and research.

Knowledge Graph

arrow_drop_up

Comments

Sign up or login to leave a comment