Filtered-CoPhy: Unsupervised Learning of Counterfactual Physics in Pixel Space

TL;DR

Filtered-CoPhy introduces an unsupervised approach for learning counterfactual physical dynamics directly in pixel space, enabling long-term video forecasting without supervision of object positions, by leveraging a hybrid latent representation.

Contribution

The paper presents a novel unsupervised method for counterfactual physics learning in pixel space using a hybrid latent representation combining dense features, keypoints, and latent vectors, outperforming existing baselines.

Findings

Outperforms strong baselines in pixel-space physics prediction

Effectively captures physical dynamics with a hybrid latent representation

Introduces a new challenging counterfactual video prediction benchmark

Abstract

Learning causal relationships in high-dimensional data (images, videos) is a hard task, as they are often defined on low dimensional manifolds and must be extracted from complex signals dominated by appearance, lighting, textures and also spurious correlations in the data. We present a method for learning counterfactual reasoning of physical processes in pixel space, which requires the prediction of the impact of interventions on initial conditions. Going beyond the identification of structural relationships, we deal with the challenging problem of forecasting raw video over long horizons. Our method does not require the knowledge or supervision of any ground truth positions or other object or scene properties. Our model learns and acts on a suitable hybrid latent representation based on a combination of dense features, sets of 2D keypoints and an additional latent vector per keypoint. We show that this better captures the dynamics of physical processes than purely dense or sparse representations. We introduce a new challenging and carefully designed counterfactual benchmark for predictions in pixel space and outperform strong baselines in physics-inspired ML and video prediction.