用训练动态生成带不确定性的伪标签，提升医学影像诊断模型的鲁棒性。
Pseudo-D: Informing Multi-View Uncertainty Estimation with Calibrated Neural Training Dynamics
- 利用神经网络训练过程中的预测波动生成不确定性伪标签。
- 在超声心动图分类任务中，校准度与多视角融合效果优于现有方法。
- 不依赖特定架构，可嵌入任意监督学习流程，适合医疗场景复杂数据。
医学影像辅助诊断系统需在噪声大、模糊或矛盾的图像上做出关键决策，但当前模型使用过于简化的独热标签，忽略了诊断不确定性。独热标签抹除了医生间差异，使模型对不完整或含伪影输入过度自信。本文提出新框架，将不确定性重新引入标签空间。通过聚合和校准训练过程中的模型预测，利用神经网络训练动态（NNTD）评估每个样本的内在难度，生成反映学习过程中模糊性的不确定性感知伪标签。该标签增强方法与架构无关，可应用于任意监督学习流程，提升不确定性估计与鲁棒性。在具有挑战性的超声心动图分类基准上验证，相比专用基线，在校准、选择性分类和多视角融合方面表现更优。
原文摘要 · Abstract (English)
Computer-aided diagnosis systems must make critical decisions from medical images that are often noisy, ambiguous, or conflicting, yet today's models are trained on overly simplistic labels that ignore diagnostic uncertainty. One-hot labels erase inter-rater variability and force models to make overconfident predictions, especially when faced with incomplete or artifact-laden inputs. We address this gap by introducing a novel framework that brings uncertainty back into the label space. Our method leverages neural network training dynamics (NNTD) to assess the inherent difficulty of each training sample. By aggregating and calibrating model predictions during training, we generate uncertainty-aware pseudo-labels that reflect the ambiguity encountered during learning. This label augmentation approach is architecture-agnostic and can be applied to any supervised learning pipeline to enhance uncertainty estimation and robustness. We validate our approach on a challenging echocardiography classification benchmark, demonstrating superior performance over specialized baselines in calibration, selective classification, and multi-view fusion.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。