The present model has weaknesses. It may well wrestle with correctly simulating the physics of a complex scene, and could not fully grasp specific situations of induce and result. For example, a person could take a bite out of a cookie, but afterward, the cookie might not have a Chunk mark.