关联多模态大语言模型中的感知、置信度与准确率
Linking Perception, Confidence and Accuracy in MLLMs
对做多模态大模型视觉感知的人来说，这篇给出了一个可复用的思路：用原图-加噪图配对构造置信度奖励做强化学习，让模型在视觉不确定时给出校准的置信度；校准后的置信度还能直接驱动测试时扩展（自一致性/自反思/视觉自检的调度），是一种无需重训即可提升准确率的推理期手段。
原文依据 we propose Confidence-Driven Reinforcement Learning (CDRL), which uses original-noise image pairs and a novel confidence-based reward
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论近期多模态大语言模型（MLLMs）的进展主要集中于增强视觉感知以提升准确率。然而，一个关键问题尚未被探索：模型是否知道自己何时不知道？通过一项探测实验，我们揭示了 MLLMs 中严重的置信度失准问题。为解决该问题，我们提出置信度驱动强化学习（CDRL），它使用原图-噪声图配对以及一种新颖的基于置信度的奖励，以增强感知敏感性并稳健地校准模型的置信度。除训练收益外，校准后的置信度还能作为一种免费收益，使测试时扩展更为有效。我们进一步提出置信度感知测试时扩展（CA-TTS），它在置信度信号的引导下动态协调自一致性、自反思与视觉自检模块。一个专家模型扮演多重角色（如规划者、评论者、投票者）来调度这些模块并提供外部验证。我们的集成框架在四个基准上取得一致 8.8% 的提升，确立了新的最优结果。更多消融研究证明了各模块的有效性与扩展优越性。我们的代码将在论文被接收后发布。
英文原文摘要 1236 字符
Recent advances in Multi-modal Large Language Models (MLLMs) have predominantly focused on enhancing visual \perception to improve \accuracy. However, a critical question remains unexplored: Do models know when they do not know? Through a probing experiment, we reveal a severe \confidence miscalibration problem in MLLMs. To address this, we propose Confidence-Driven Reinforcement Learning (CDRL), which uses original-noise image pairs and a novel confidence-based reward to enhance perceptual sensitivity and robustly calibrate the model's confidence. Beyond training benefits, calibrated confidence enables more effective test-time scaling as a free lunch. We further propose Confidence-Aware Test-Time Scaling (CA-TTS), which dynamically coordinates Self-Consistency, Self-Reflection, and Visual Self-Check modules guided by confidence signals. An Expert Model acts in multiple roles (e.g., Planner, Critic, Voter) to schedule these modules and provide external verification. Our integrated framework establishes new state-of-the-art results with consistent 8.8% gains across four benchmarks. More ablation studies demonstrate the effectiveness of each module and scaling superiority. Our code will be released after the acception.