标题： 校准感知的提示学习用于医疗视觉-语言模型 显示英文标题

标题： Calibration-Aware Prompt Learning for Medical Vision-Language Models

摘要： 医学视觉-语言模型（Med-VLMs）通过利用大规模图像-文本预训练，在各种医学影像任务中表现出色。然而，它们的置信度校准研究较少，仍然是一个重大挑战。因此，校准不当的预测可能导致过度自信的错误，削弱临床信任和决策的可靠性。为了解决这个问题，我们引入了CalibPrompt，这是第一个在提示调优过程中校准Med-VLMs的框架。CalibPrompt在标签数据稀缺的情况下，通过精心设计的校准目标优化一小部分可学习的提示。首先，我们研究了一个正则化器，旨在使平滑后的准确率与预测的模型置信度对齐。其次，我们引入了一个角度分离损失，以最大化文本特征的接近性，从而提高多模态Med-VLMs在置信度估计中的可靠性。在四个公开可用的Med-VLMs和五个不同的医学影像数据集上的大量实验表明，CalibPrompt在不影响干净准确率的情况下持续改善校准。我们的代码可在https://github.com/iabh1shekbasu/CalibPrompt获取。 显示英文摘要

摘要： Medical Vision-Language Models (Med-VLMs) have demonstrated remarkable performance across diverse medical imaging tasks by leveraging large-scale image-text pretraining. However, their confidence calibration is largely unexplored, and so remains a significant challenge. As such, miscalibrated predictions can lead to overconfident errors, undermining clinical trust and decision-making reliability. To address this, we introduce CalibPrompt, the first framework to calibrate Med-VLMs during prompt tuning. CalibPrompt optimizes a small set of learnable prompts with carefully designed calibration objectives under scarce labeled data regime. First, we study a regularizer that attempts to align the smoothed accuracy with the predicted model confidences. Second, we introduce an angular separation loss to maximize textual feature proximity toward improving the reliability in confidence estimates of multimodal Med-VLMs. Extensive experiments on four publicly available Med-VLMs and five diverse medical imaging datasets reveal that CalibPrompt consistently improves calibration without drastically affecting clean accuracy. Our code is available at https://github.com/iabh1shekbasu/CalibPrompt.