ContPhy: Continuum Physical Concept Learning and Reasoning from Videos
This addresses the problem of limited physical commonsense in AI for researchers, though it is incremental as it builds on existing benchmarks by adding continuum scenarios.
The authors introduced ContPhy, a benchmark for evaluating machine physical commonsense across diverse properties like mass and density, and found that current AI models perform poorly on it, highlighting a lack of continuum physical understanding. They also proposed ContPRO, an oracle model combining particle-based dynamics with large language models to improve predictions and reasoning.
We introduce the Continuum Physical Dataset (ContPhy), a novel benchmark for assessing machine physical commonsense. ContPhy complements existing physical reasoning benchmarks by encompassing the inference of diverse physical properties, such as mass and density, across various scenarios and predicting corresponding dynamics. We evaluated a range of AI models and found that they still struggle to achieve satisfactory performance on ContPhy, which shows that the current AI models still lack physical commonsense for the continuum, especially soft-bodies, and illustrates the value of the proposed dataset. We also introduce an oracle model (ContPRO) that marries the particle-based physical dynamic models with the recent large language models, which enjoy the advantages of both models, precise dynamic predictions, and interpretable reasoning. ContPhy aims to spur progress in perception and reasoning within diverse physical settings, narrowing the divide between human and machine intelligence in understanding the physical world. Project page: https://physical-reasoning-project.github.io