Search papers, labs, and topics across Lattice.
PhysMind is a novel training-free framework that generates reusable, question-agnostic executable worlds from video inputs, enabling reliable physical reasoning. By employing object segmentation, mesh reconstruction, and 6D pose tracking, it constructs temporally consistent dynamic scenes and fits continuous-time dynamics without relying on traditional simulators. The framework significantly outperforms existing vision-language models, achieving a 38.23-point accuracy improvement on CLEVRER and surpassing GPT-5.5 by 19.25 points on counterfactual reasoning tasks.
PhysMind achieves a remarkable 38.23-point accuracy boost in physical reasoning tasks by transforming videos into executable worlds without the need for training.
Reliable physical reasoning from video requires understanding how objects move, interact, and respond to interventions. Existing vision-language models (VLMs) often struggle to interpret these dynamics and reason reliably about future and counterfactual outcomes. We introduce PhysMind, a training-free agentic framework that constructs one reusable, question-agnostic executable world per video. PhysMind recovers a temporally consistent dynamic scene through object segmentation, mesh reconstruction, and 6D pose tracking, then fits analytic continuous-time dynamics and latent physical parameters without unrolling a time-stepped simulator. Given a question, it inspects, continues, or edits the world and answers from the resulting trajectories and interactions. Relative to direct chain-of-thought (CoT) reasoning with the same VLM, PhysMind improves accuracy by 38.23 points on CLEVRER and 8.08 points on Physion++. On counterfactual questions, it exceeds the strongest evaluated VLM baseline, GPT-5.5, by 19.25 points.
