Active Reasoning Vision-Language Models via Sequential Experimental Design
For researchers working on VLMs and complex visual reasoning, this work provides a principled framework to overcome the trade-off between field of view and fine-grained detail, enabling more accurate perception in high-resolution tasks.
The paper addresses the perceptual bandwidth bottleneck in Vision-Language Models (VLMs) by framing visual reasoning as a sequential decision-making process based on Bayesian optimal experimental design. Their training-free inference strategy improves state-of-the-art models on gigapixel-level benchmarks, significantly outperforming baselines and narrowing the gap to human-level performance.
Visual perception in modern Vision-Language Models (VLMs) is constrained by a fundamental perceptual bandwidth bottleneck: a broad field of view inevitably sacrifices the fine-grained details necessary for complex reasoning. Inspired by the classical paradigms of active vision and information foraging, we frame overcoming this limitation as a sequential decision-making process. We formalise this process through the lens of the sequential Bayesian optimal experimental design (S-BOED) problem. While exact Bayesian inference is intractable in continuous gigapixel spaces, we derive principled yet tractable approximations that balance spatial coverage against resolution. To validate this framework, we present a training-free inference strategy as a practical instantiation of the S-BOED objective for agents equipped with multiple vision tools. Designed as a flexible template, this strategy accommodates arbitrary optimisation algorithms, ranging from efficient greedy sampling to look-ahead planning, to approximate the optimal design. Empirical evaluations on gigapixel-level benchmarks demonstrate that our approach further boosts the performance of state-of-the-art models, significantly outperforming standard baselines and effectively narrowing the gap towards human-annotated oracles.