Reproduce “Thinking With Spatial–Social Code for Social Reasoning in Videos” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Xiaoyuan Zhang, Zhaowei Zhang, Wenbo Li, Alan Yuille, Tianmin Shu, Jieneng Chen
SSC separates physical events from participants’ observations and information history. It links records of entities, events, and perceptual access to attributed observations, messages, and hypothesized beliefs. The authors evaluate it on a diagnostic, video question-answering, and Habitat/EnactToM tasks. It reaches 95.0% paired correctness and improves benchmark accuracy, but participant states remain estimates, and syntax alone is not shown necessary.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.