MathVerse: Does Your Multi-modal LLM Truly See the Diagrams in Visual Math Problems?

TL;DR

MathVerse is a comprehensive benchmark designed to evaluate whether multi-modal large language models genuinely understand diagrams in visual math problems, using diverse problem versions and a detailed reasoning assessment strategy.

Contribution

The paper introduces MathVerse, a new benchmark with 2,612 math problems and a novel Chain-of-Thought evaluation method for assessing visual understanding in MLLMs.

Findings

MLLMs struggle with visual diagram comprehension in math problems.

The benchmark reveals gaps in current MLLMs' reasoning capabilities.

Detailed error analysis guides future model improvements.

Abstract

The remarkable progress of Multi-modal Large Language Models (MLLMs) has garnered unparalleled attention, due to their superior performance in visual contexts. However, their capabilities in visual math problem-solving remain insufficiently evaluated and understood. We investigate current benchmarks to incorporate excessive visual content within textual questions, which potentially assist MLLMs in deducing answers without truly interpreting the input diagrams. To this end, we introduce MathVerse, an all-around visual math benchmark designed for an equitable and in-depth evaluation of MLLMs. We meticulously collect 2,612 high-quality, multi-subject math problems with diagrams from publicly available sources. Each problem is then transformed by human annotators into six distinct versions, each offering varying degrees of information content in multi-modality, contributing to 15K test samples in total. This approach allows MathVerse to comprehensively assess whether and how much MLLMs can truly understand the visual diagrams for mathematical reasoning. In addition, we propose a Chain-of-Thought (CoT) evaluation strategy for a fine-grained assessment of the output answers. Rather than naively judging True or False, we employ GPT-4(V) to adaptively extract crucial reasoning steps, and then score each step with detailed error analysis, which can reveal the intermediate CoT reasoning quality by MLLMs. We hope the MathVerse benchmark may provide unique insights to guide the future development of MLLMs. Project page: https://mathverse-cuhk.github.io