VRBench: A Benchmark for Multi-Step Reasoning in Long Narrative Videos
This addresses the problem of limited evaluation tools for multi-step reasoning in long videos for researchers in AI and computer vision, though it is incremental as it builds on existing benchmark methodologies.
The authors tackled the lack of benchmarks for evaluating multi-step reasoning in long narrative videos by introducing VRBench, which includes 960 long videos and over 8,000 question-answer pairs, and found through evaluations of 12 LLMs and 19 VLMs that it provides valuable insights for advancing the field.
We present VRBench, the first long narrative video benchmark crafted for evaluating large models' multi-step reasoning capabilities, addressing limitations in existing evaluations that overlook temporal reasoning and procedural validity. It comprises 960 long videos (with an average duration of 1.6 hours), along with 8,243 human-labeled multi-step question-answering pairs and 25,106 reasoning steps with timestamps. These videos are curated via a multi-stage filtering process including expert inter-rater reviewing to prioritize plot coherence. We develop a human-AI collaborative framework that generates coherent reasoning chains, each requiring multiple temporally grounded steps, spanning seven types (e.g., event attribution, implicit inference). VRBench designs a multi-phase evaluation pipeline that assesses models at both the outcome and process levels. Apart from the MCQs for the final results, we propose a progress-level LLM-guided scoring metric to evaluate the quality of the reasoning chain from multiple dimensions comprehensively. Through extensive evaluations of 12 LLMs and 19 VLMs on VRBench, we undertake a thorough analysis and provide valuable insights that advance the field of multi-step reasoning.