Test-Time Warmup for Multimodal Large Language Models
This addresses the issue of limited training data for MLLMs, offering an incremental improvement for multimodal reasoning tasks.
The paper tackles the problem of weak performance in Multimodal Large Language Models (MLLMs) on complex reasoning tasks by proposing a Test-Time Warmup method that adapts the model per test instance using weakly supervised auxiliary tasks, resulting in relative performance improvements of 4.03% on MMMU, 5.28% on VQA-Rad, and 1.63% on GQA.
Multimodal Large Language Models (MLLMs) hold great promise for advanced reasoning at the intersection of text and images, yet they have not fully realized this potential. MLLMs typically integrate an LLM, a vision encoder, and a connector that maps the vision encoder's embeddings into the LLM's text embedding space. Although each component is pretrained on massive datasets with billions of samples, the entire multimodal model is typically trained on only thousands (or a few million) samples, which can result in weak performance on complex reasoning tasks. To address these shortcomings, instead of relying on extensive labeled datasets for fine-tuning, we propose a Test-Time Warmup method that adapts the MLLM per test instance by leveraging data from weakly supervised auxiliary tasks. With our approach, we observe a relative performance improvement of 4.03% on MMMU, 5.28% on VQA-Rad, and 1.63% on GQA on the Llama-Vision-Instruct model. Our method demonstrates that 'warming up' before inference can enhance MLLMs' robustness across diverse reasoning tasks.