STAR: A Benchmark for Situated Reasoning in Real-World Videos


Published on


Reasoning in the real world is not divorced from situations. How to capture the present knowledge from surrounding situations and perform reasoning accordingly is crucial and challenging for machine intelligence. This paper introduces a new benchmark that evaluates the situated reasoning ability via situation abstraction and logic-grounded question answering for real-world videos, called Situated Reasoning in Real-World Videos (STAR). This benchmark is built upon the real-world videos associated with human actions or interactions, which are naturally dynamic, compositional, and logical. The dataset includes four types of questions, including interaction, sequence, prediction, and feasibility. We represent the situations in real-world videos by hyper-graphs connecting extracted atomic entities and relations (e.g., actions, persons, objects, and relationships). Besides visual perception, situated reasoning also requires structured situation comprehension and logical reasoning. Questions and answers are procedurally generated. The answering logic represents by a functional program based on the situation hyper-graph. We compare various existing video reasoning models and find that they all struggle on this challenging situated reasoning task. We further propose a diagnostic neuro-symbolic model that can disentangle visual perception, situation abstraction, language understanding, and functional reasoning to understand the challenges of this benchmark.

This paper has been published at NeurIPS 2021 Dataset and Benchmarks trackĀ 

Please cite our work using the BibTeX below.

title={{STAR}: A Benchmark for Situated Reasoning in Real-World Videos},
author={Bo Wu and Shoubin Yu and Zhenfang Chen and Joshua B. Tenenbaum and Chuang Gan},
booktitle={Thirty-fifth Conference on Neural Information Processing Systems Datasets and Benchmarks Track (Round 2)},
Close Modal