MMEvalPro: Calibrating Multimodal Benchmarks Towards Trustworthy and Efficient Evaluation
This addresses the problem of unreliable evaluation in multimodal AI research, providing a more rigorous benchmark for researchers, but it is incremental as it builds on existing benchmarks with enhanced methodology.
The paper tackles the problem of systematic biases in multimodal benchmarks for evaluating Large Multimodal Models (LMMs), where Large Language Models (LLMs) without visual capabilities achieve non-trivial performance, undermining evaluation credibility. The result is MMEvalPro, a new benchmark with 6,414 questions that is more challenging, showing the best LMM lags behind human performance by 31.73% compared to an average gap of 8.03% in previous benchmarks, and more trustworthy, with the best LLM trailing the best LMM by 23.09% versus 14.64% previously.
Large Multimodal Models (LMMs) exhibit impressive cross-modal understanding and reasoning abilities, often assessed through multiple-choice questions (MCQs) that include an image, a question, and several options. However, many benchmarks used for such evaluations suffer from systematic biases. Remarkably, Large Language Models (LLMs) without any visual perception capabilities achieve non-trivial performance, undermining the credibility of these evaluations. To address this issue while maintaining the efficiency of MCQ evaluations, we propose MMEvalPro, a benchmark designed to avoid Type-I errors through a trilogy evaluation pipeline and more rigorous metrics. For each original question from existing benchmarks, human annotators augment it by creating one perception question and one knowledge anchor question through a meticulous annotation process. MMEvalPro comprises $2,138$ question triplets, totaling $6,414$ distinct questions. Two-thirds of these questions are manually labeled by human experts, while the rest are sourced from existing benchmarks (MMMU, ScienceQA, and MathVista). Compared with the existing benchmarks, our experiments with the latest LLMs and LMMs demonstrate that MMEvalPro is more challenging (the best LMM lags behind human performance by $31.73\%$, compared to an average gap of $8.03\%$ in previous benchmarks) and more trustworthy (the best LLM trails the best LMM by $23.09\%$, whereas the gap for previous benchmarks is just $14.64\%$). Our in-depth analysis explains the reason for the large performance gap and justifies the trustworthiness of evaluation, underscoring its significant potential for advancing future research.