Visual Physics: Discovering Physical Laws from Videos
This addresses the challenge of automated scientific discovery from visual data, though it is incremental as it focuses on simple, known physical tasks.
The paper tackles the problem of teaching a machine to discover physical laws, such as projectile motion, from video streams without prior physics knowledge, achieving verification against textbook equations on elementary phenomena.
In this paper, we teach a machine to discover the laws of physics from video streams. We assume no prior knowledge of physics, beyond a temporal stream of bounding boxes. The problem is very difficult because a machine must learn not only a governing equation (e.g. projectile motion) but also the existence of governing parameters (e.g. velocities). We evaluate our ability to discover physical laws on videos of elementary physical phenomena, such as projectile motion or circular motion. These elementary tasks have textbook governing equations and enable ground truth verification of our approach.