MINERVA: Evaluating Complex Video Reasoning

TL;DR

MINERVA introduces a new video reasoning dataset with detailed reasoning traces to evaluate and analyze the reasoning capabilities of multimodal models, revealing common failure modes and advancing understanding of video comprehension.

Contribution

The paper presents MINERVA, a comprehensive video reasoning dataset with annotated reasoning traces, enabling detailed evaluation and analysis of multimodal models' reasoning abilities.

Findings

Models struggle with temporal localization.

Visual perception errors are common.

Logical errors are less frequent.

Abstract

Multimodal LLMs are turning their focus to video benchmarks, however most video benchmarks only provide outcome supervision, with no intermediate or interpretable reasoning steps. This makes it challenging to assess if models are truly able to combine perceptual and temporal information to reason about videos, or simply get the correct answer by chance or by exploiting linguistic biases. To remedy this, we provide a new video reasoning dataset called MINERVA for modern multimodal models. Each question in the dataset comes with 5 answer choices, as well as detailed, hand-crafted reasoning traces. Our dataset is multimodal, diverse in terms of video domain and length, and consists of complex multi-step questions. Extensive benchmarking shows that our dataset provides a challenge for frontier open-source and proprietary models. We perform fine-grained error analysis to identify common failure modes across various models, and create a taxonomy of reasoning errors. We use this to explore both human and LLM-as-a-judge methods for scoring video reasoning traces, and find that failure modes are primarily related to temporal localization, followed by visual perception errors, as opposed to logical or completeness errors. The dataset, along with questions, answer candidates and reasoning traces will be publicly available under https://github.com/google-deepmind/neptune?tab=readme-ov-file\#minerva.