通过强化学习发现视觉-语言模型的失败模式
Discovering Failure Modes in Vision-Language Models using RL
AI总结：
本文提出基于强化学习的框架，自动发现视觉-语言模型在特定数据分布上的失败模式，通过生成问题来识别模型在细粒度视觉细节和技能组合上的不足。
摘要：
视觉-语言模型（VLMs）尽管在多模态基准测试中表现强劲，但往往在人类轻易识别的视觉概念上出现误解，如计数、空间推理和视角理解。先前研究通过人工识别这些弱点，发现它们通常源于特定技能的缺陷。然而，这种人工努力成本高、难以扩展且易受人类偏见影响，常忽视细微细节而关注显著对象，导致对模型脆弱性的不完全理解。为解决这些限制，我们提出基于强化学习的框架，无需人工干预，自动发现任何“候选VLM”在给定数据分布上的失败模式或盲点。我们的框架训练了一个问题生成代理，根据候选VLM的响应自适应生成问题以引发错误答案。通过在训练过程中聚焦细粒度视觉细节和不同的技能组合，我们的方法提高了问题的复杂性，从而识别出VLMs在新失败模式上的挣扎。我们通过展示其在各种模型组合上的通用性，证明了该框架的广泛适用性。
英文摘要：
Vision-language Models (VLMs), despite achieving strong performance on multimodal benchmarks, often misinterpret straightforward visual concepts that humans identify effortlessly, such as counting, spatial reasoning, and viewpoint understanding. Previous studies manually identified these weaknesses and found that they often stem from deficits in specific skills. However, such manual efforts are costly, unscalable, and subject to human bias, which often overlooks subtle details in favour of salient objects, resulting in an incomplete understanding of a model's vulnerabilities. To address these limitations, we propose a Reinforcement Learning (RL)-based framework to automatically discover the failure modes or blind spots of any ``candidate VLM'' on a given data distribution without human intervention. Our framework trains a questioner agent that adaptively generates queries based on the candidate VLM's responses to elicit incorrect answers. Our approach increases question complexity by focusing on fine-grained visual details and distinct skill compositions as training progresses, consequently identifying novel failure modes in which VLMs struggle. We demonstrate the broad applicability of our framework by showcasing its generalizability across various model combinations.
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。
