VISion On Request: Enhanced VLLM efficiency with sparse, dynamically selected, vision-language interactions

Read full text Summarise with VIRAT
Authors: Adrian Bulat, Alberto Baldrati, Ioannis Maniadis Metaxas, Yassine Ouali, Georgios Tzimiropoulos
Cited by: 0
References: 50

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Existing approaches for improving the efficiency of Large Vision-Language Models (LVLMs) are largely based on the concept of visual token reduction. This approach, however, creates an information bottleneck that impairs performance, especially on challenging tasks that require fine-grained understanding and reasoning. In this work, we challenge this paradigm by introducing VISion On Request (VISOR), a method that reduces inference cost without discarding visual information. Instead of compressing the image, VISOR improves efficiency by sparsifying the interaction between image and text tokens. Specifically, the language model attends to the full set of high-resolution visual tokens through a small, strategically placed set of attention layers: general visual context is provided by efficient cross-attention between text-image, while a few well-placed and dynamically selected self-attention layers refine the visual representations themselves, enabling complex, high-resolution reasoning when needed. Based on this principle, we first train a single universal network on a range of computational budgets by varying the number of self-attention layers, and then introduce a lightweight policy mechanism that dynamically allocates visual computation based on per-sample complexity. Extensive experiments show that VISOR drastically reduces computational cost while matching or exceeding state-of-the-art results across a diverse suite of benchmarks, and excels in challenging tasks that require detailed visual understanding.

Fields of Study & Subjects

Multimodal Machine Learning Applications Advanced Neural Network Applications Visual Attention and Saliency Detection

Publication details

Journal / source
arXiv (Cornell University)
Publication date
24 March 2026
Language
EN
Document type
Preprint