LLMPhy: Parameter-Identifiable Physical Reasoning Combining Large Language Models and Physics Engines

ArXi:2411.08027v3 Announce Type: replace-cross Most learning-based approaches to complex physical reasoning sidestep the crucial problem of parameter identification (e.g., mass, friction) that governs scene dynamics, despite its importance in real-world applications such as collision avoidance and robotic manipulation. In this paper, we present LLMPhy, a black-box optimization framework that integrates large language models (LLMs) with physics simulators for physical reasoning.