M$^3$-VQA: A Benchmark for Multimodal, Multi-Entity, Multi-Hop Visual Question Answering

TL;DR

M$^3$-VQA is a new benchmark designed to evaluate multimodal large language models on complex multi-entity, multi-hop reasoning tasks involving visual and textual data, highlighting current challenges and improvements with structured retrieval.

Contribution

Introduces M$^3$-VQA, a challenging multimodal VQA benchmark with multi-entity, multi-hop questions and detailed evidence, to better assess and advance MLLMs' reasoning capabilities.

Findings

Models perform poorly without external knowledge.

Providing gold evidence significantly improves results.

Structured, reasoning-aware retrieval outperforms heuristic methods.

Abstract

We present M$^3$-VQA, a novel knowledge-based Visual Question Answering (VQA) benchmark, to enhance the evaluation of multimodal large language models (MLLMs) in fine-grained multimodal entity understanding and complex multi-hop reasoning. Unlike existing VQA datasets that focus on coarse-grained categories and simple reasoning over single entities, M$^3$-VQA introduces diverse multi-entity questions involving multiple distinct entities from both visual and textual sources. It requires models to perform both sequential and parallel multi-hop reasoning across multiple documents, supported by traceable, detailed evidence and a curated multimodal knowledge base. We evaluate 16 leading MLLMs under three settings: without external knowledge, with gold evidence, and with retrieval-augmented input. The poor results reveal significant challenges for MLLMs in knowledge acquisition and reasoning. Models perform poorly without external information but improve markedly when provided with precise evidence. Furthermore, reasoning-aware agentic retrieval surpasses heuristic methods, highlighting the importance of structured reasoning for complex multimodal understanding. M$^3$-VQA presents a more challenging evaluation for advancing the multimodal reasoning capabilities of MLLMs. Our code and dataset are available at https://github.com/CASIA-IVA-Lab/M3VQA.