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

TL;DR

The paper introduces STAR, a new benchmark dataset for evaluating situated reasoning in real-world videos, emphasizing situation abstraction and logic-grounded question answering to advance machine intelligence.

Contribution

It presents a novel benchmark with a hyper-graph based representation of situations, procedural question generation, and a diagnostic neuro-symbolic model for better understanding reasoning challenges.

Findings

Existing models perform poorly on the benchmark.

The neuro-symbolic model improves reasoning performance.

The dataset highlights the complexity of situated reasoning in videos.

Abstract

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 Benchmark). 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 of each question is represented by a functional program based on a 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.