Rethinking Visual Prompting for Multimodal Large Language Models with External Knowledge
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909243737636864 |
|---|---|
| author | Lin, Yuanze Li, Yunsheng Chen, Dongdong Xu, Weijian Clark, Ronald Torr, Philip Yuan, Lu |
| author_facet | Lin, Yuanze Li, Yunsheng Chen, Dongdong Xu, Weijian Clark, Ronald Torr, Philip Yuan, Lu |
| contents | In recent years, multimodal large language models (MLLMs) have made significant strides by training on vast high-quality image-text datasets, enabling them to generally understand images well. However, the inherent difficulty in explicitly conveying fine-grained or spatially dense information in text, such as masks, poses a challenge for MLLMs, limiting their ability to answer questions requiring an understanding of detailed or localized visual elements. Drawing inspiration from the Retrieval-Augmented Generation (RAG) concept, this paper proposes a new visual prompt approach to integrate fine-grained external knowledge, gleaned from specialized vision models (e.g., instance segmentation/OCR models), into MLLMs. This is a promising yet underexplored direction for enhancing MLLMs' performance. Our approach diverges from concurrent works, which transform external knowledge into additional text prompts, necessitating the model to indirectly learn the correspondence between visual content and text coordinates. Instead, we propose embedding fine-grained knowledge information directly into a spatial embedding map as a visual prompt. This design can be effortlessly incorporated into various MLLMs, such as LLaVA and Mipha, considerably improving their visual understanding performance. Through rigorous experiments, we demonstrate that our method can enhance MLLM performance across nine benchmarks, amplifying their fine-grained context-aware capabilities. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_04681 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Rethinking Visual Prompting for Multimodal Large Language Models with External Knowledge Lin, Yuanze Li, Yunsheng Chen, Dongdong Xu, Weijian Clark, Ronald Torr, Philip Yuan, Lu Computer Vision and Pattern Recognition Artificial Intelligence Computation and Language Machine Learning In recent years, multimodal large language models (MLLMs) have made significant strides by training on vast high-quality image-text datasets, enabling them to generally understand images well. However, the inherent difficulty in explicitly conveying fine-grained or spatially dense information in text, such as masks, poses a challenge for MLLMs, limiting their ability to answer questions requiring an understanding of detailed or localized visual elements. Drawing inspiration from the Retrieval-Augmented Generation (RAG) concept, this paper proposes a new visual prompt approach to integrate fine-grained external knowledge, gleaned from specialized vision models (e.g., instance segmentation/OCR models), into MLLMs. This is a promising yet underexplored direction for enhancing MLLMs' performance. Our approach diverges from concurrent works, which transform external knowledge into additional text prompts, necessitating the model to indirectly learn the correspondence between visual content and text coordinates. Instead, we propose embedding fine-grained knowledge information directly into a spatial embedding map as a visual prompt. This design can be effortlessly incorporated into various MLLMs, such as LLaVA and Mipha, considerably improving their visual understanding performance. Through rigorous experiments, we demonstrate that our method can enhance MLLM performance across nine benchmarks, amplifying their fine-grained context-aware capabilities. |
| title | Rethinking Visual Prompting for Multimodal Large Language Models with External Knowledge |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Computation and Language Machine Learning |
| url | https://arxiv.org/abs/2407.04681 |