VisRes Bench: 关于评估视觉语言模型的视觉推理能力
VisRes Bench: On Evaluating the Visual Reasoning Capabilities of VLMs
AI总结：
VisRes Bench通过三个层次评估VLMs的视觉推理能力，揭示其在复杂任务中的局限性，并为多模态研究提供统一框架。
摘要：
视觉语言模型（VLMs）在视觉问答和图像描述等任务上取得了显著进展。然而，这些模型在视觉推理方面的能力与依赖语言先验的程度有多大仍不清楚。为此，我们引入了VisRes Bench，这是一个设计用于在自然环境中研究视觉推理的基准测试。通过分析模型在三个复杂性层次上的行为，我们揭示了感知和关系视觉推理能力的明显局限性。VisRes在不同层次上隔离了不同的推理能力。第一层在模糊、纹理变化、遮挡和旋转等扰动下测试感知完成和全局图像匹配；第二层测试单一属性（如颜色、计数、方向）上的规则推理；第三层则针对需要整合多个视觉属性的组合推理。在超过19000张受控任务图像上，我们发现最先进的VLMs在细微感知扰动下表现接近随机，揭示了超出模式识别的抽象能力有限。我们最后讨论了VisRes如何为多模态研究中的抽象视觉推理提供统一框架。
英文摘要：
Vision-Language Models (VLMs) have achieved remarkable progress across tasks such as visual question answering and image captioning. Yet, the extent to which these models perform visual reasoning as opposed to relying on linguistic priors remains unclear. To address this, we introduce VisRes Bench, a benchmark designed to study visual reasoning in naturalistic settings without contextual language supervision. Analyzing model behavior across three levels of complexity, we uncover clear limitations in perceptual and relational visual reasoning capacities. VisRes isolates distinct reasoning abilities across its levels. Level 1 probes perceptual completion and global image matching under perturbations such as blur, texture changes, occlusion, and rotation; Level 2 tests rule-based inference over a single attribute (e.g., color, count, orientation); and Level 3 targets compositional reasoning that requires integrating multiple visual attributes. Across more than 19,000 controlled task images, we find that state-of-the-art VLMs perform near random under subtle perceptual perturbations, revealing limited abstraction beyond pattern recognition. We conclude by discussing how VisRes provides a unified framework for advancing abstract visual reasoning in multimodal research.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
