Enhancing Multimodal Large Language Models with Multi-instance Visual Prompt Generator for Visual Representation Enrichment
This work addresses visual representation limitations in multimodal LLMs for vision-language tasks, but appears incremental as it builds on existing adapter methods like Q-former.
The paper tackles the problem of visual representation enrichment in multimodal large language models by proposing a Multi-instance Visual Prompt Generator (MIVPG) that incorporates instance correlation between images or patches, which improves Q-former performance on three public vision-language datasets.
Multimodal Large Language Models (MLLMs) have achieved SOTA performance in various visual language tasks by fusing the visual representations with LLMs leveraging some visual adapters. In this paper, we first establish that adapters using query-based Transformers such as Q-former is a simplified Multi-instance Learning method without considering instance heterogeneity/correlation. We then propose a general component termed Multi-instance Visual Prompt Generator (MIVPG) to incorporate enriched visual representations into LLMs by taking advantage of instance correlation between images or patches for the same sample. Quantatitive evaluation on three public vision-language (VL) datasets from different scenarios shows that the proposed MIVPG improves Q-former in main VL tasks.