Unsupervised Intuitive Physics from Visual Observations
Sebastien Ehrhardt, Aron Monszpart, Niloy Mitra, Andrea, Vedaldi

TL;DR
This paper introduces a method to learn intuitive physics directly from raw visual data without supervision or simulators, enabling prediction of object trajectories in complex real-world environments.
Contribution
It presents the first approach to learn physics predictors solely from raw videos using unsupervised object tracking and motion prediction, without external supervision or simulators.
Findings
Successfully predicts object trajectories in synthetic datasets.
Introduces ROLL4REAL, a new real-world dataset for physics learning.
Achieves reliable trajectory extrapolation from raw videos without supervision.
Abstract
While learning models of intuitive physics is an increasingly active area of research, current approaches still fall short of natural intelligences in one important regard: they require external supervision, such as explicit access to physical states, at training and sometimes even at test times. Some authors have relaxed such requirements by supplementing the model with an handcrafted physical simulator. Still, the resulting methods are unable to automatically learn new complex environments and to understand physical interactions within them. In this work, we demonstrated for the first time learning such predictors directly from raw visual observations and without relying on simulators. We do so in two steps: first, we learn to track mechanically-salient objects in videos using causality and equivariance, two unsupervised learning principles that do not require auto-encoding. Second,…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
