论文标题
及时的针迹节省了九个：火车时间正规化损失以改善神经网络校准
A Stitch in Time Saves Nine: A Train-Time Regularizing Loss for Improved Neural Network Calibration
论文作者
论文摘要
众所周知，深神经网络（DNN S）会造成过度自信的错误，这使他们在安全至关重要的应用中使用了问题。最先进的（SOTA）校准技术仅根据预测标签的置信度提高了，并使非校准（例如TOP-2，TOP-5）的置信度未经校准。这种校准不适用于使用后处理的标签。此外，大多数SOTA技术在事后学习了一些超参数，遗漏了图像的范围或像素特定的校准。这使得它们不适合在域移动下或诸如语义分割之类的密集预测任务下进行校准。在本文中，我们主张在火车时间本身中进行介入，以便直接产生校准的DNN模型。我们提出了一种新颖的辅助损失函数：置信度和准确性的多级差异（MDCA），以实现相同的MDCA，可以与其他应用程序/特定于特定于任务的损失函数一起使用。我们表明，使用MDCA进行的训练会导致预期校准误差（ECE）以及图像分类和分割任务的静态校准误差（SCE）的培训。我们报告CIFAR 100数据集的ECE（SCE）得分为0.72（1.60），而SOTA则为1.90（1.71）。在域移动下，使用MDCA在PACS数据集上训练的RESNET-18模型的平均ECE（SCE）得分为所有域的平均ECE（SCE）分数为19.7（9.7），而SOTA为24.2（11.8）。对于分割任务，我们报告了与焦点损失相比，Pascal -VOC数据集的校准误差减少了2倍。最后，即使数据不平衡，MDCA培训也可以改善校准，以及自然语言分类任务。我们在此处发布了代码：代码可在https://github.com/mdca-loss上找到
Deep Neural Networks ( DNN s) are known to make overconfident mistakes, which makes their use problematic in safety-critical applications. State-of-the-art ( SOTA ) calibration techniques improve on the confidence of predicted labels alone and leave the confidence of non-max classes (e.g. top-2, top-5) uncalibrated. Such calibration is not suitable for label refinement using post-processing. Further, most SOTA techniques learn a few hyper-parameters post-hoc, leaving out the scope for image, or pixel specific calibration. This makes them unsuitable for calibration under domain shift, or for dense prediction tasks like semantic segmentation. In this paper, we argue for intervening at the train time itself, so as to directly produce calibrated DNN models. We propose a novel auxiliary loss function: Multi-class Difference in Confidence and Accuracy ( MDCA ), to achieve the same MDCA can be used in conjunction with other application/task-specific loss functions. We show that training with MDCA leads to better-calibrated models in terms of Expected Calibration Error ( ECE ), and Static Calibration Error ( SCE ) on image classification, and segmentation tasks. We report ECE ( SCE ) score of 0.72 (1.60) on the CIFAR 100 dataset, in comparison to 1.90 (1.71) by the SOTA. Under domain shift, a ResNet-18 model trained on PACS dataset using MDCA gives an average ECE ( SCE ) score of 19.7 (9.7) across all domains, compared to 24.2 (11.8) by the SOTA. For the segmentation task, we report a 2X reduction in calibration error on PASCAL - VOC dataset in comparison to Focal Loss. Finally, MDCA training improves calibration even on imbalanced data, and for natural language classification tasks. We have released the code here: code is available at https://github.com/mdca-loss
