MI-Pruner: Crossmodal Mutual Information-guided Token Pruner for Efficient MLLMs
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866915913088892928 |
|---|---|
| author | Li, Jiameng Tiulpin, Aleksei Blaschko, Matthew B. |
| author_facet | Li, Jiameng Tiulpin, Aleksei Blaschko, Matthew B. |
| contents | For multimodal large language models (MLLMs), visual information is relatively sparse compared with text. As a result, research on visual pruning emerges for efficient inference. Current approaches typically measure token importance based on the attention scores in the visual encoder or in the LLM decoder, then select visual tokens with high attention scores while pruning others. In this paper, we pursue a different and more surgical approach. Instead of relying on mechanism-specific signals, we directly compute Mutual Information (MI) between visual and textual features themselves, prior to their interaction. This allows us to explicitly measure crossmodal dependency at the feature levels. Our MI-Pruner is simple, efficient and non-intrusive, requiring no access to internal attention maps or architectural modifications. Experimental results demonstrate that our approach outperforms previous attention-based pruning methods with minimal latency. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_03072 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | MI-Pruner: Crossmodal Mutual Information-guided Token Pruner for Efficient MLLMs Li, Jiameng Tiulpin, Aleksei Blaschko, Matthew B. Computer Vision and Pattern Recognition For multimodal large language models (MLLMs), visual information is relatively sparse compared with text. As a result, research on visual pruning emerges for efficient inference. Current approaches typically measure token importance based on the attention scores in the visual encoder or in the LLM decoder, then select visual tokens with high attention scores while pruning others. In this paper, we pursue a different and more surgical approach. Instead of relying on mechanism-specific signals, we directly compute Mutual Information (MI) between visual and textual features themselves, prior to their interaction. This allows us to explicitly measure crossmodal dependency at the feature levels. Our MI-Pruner is simple, efficient and non-intrusive, requiring no access to internal attention maps or architectural modifications. Experimental results demonstrate that our approach outperforms previous attention-based pruning methods with minimal latency. |
| title | MI-Pruner: Crossmodal Mutual Information-guided Token Pruner for Efficient MLLMs |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2604.03072 |