A Stitch in Time Saves Nine: A Train-Time Regularizing Loss for Improved Neural Network
Calibration
Abstract
Deep Neural Networks ( DNNs) 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.
Understanding Calibration
Highlights
-
[Novelty]
We propose an auxiliary loss to overcome miscalibration
-
[Multi-class Calibration]
Entire probability vector (all K classes) taken into account
-
[Powerful Regularizer]
Models trained using our method are relatively well calibrated even under domain/dataset drift
-
[Superior Calibration]
Outperforms SOTA methods on various datasets and models
-
[Beyond Image Classification]
Promising results in semantic segmentation in images and NL classification tasks
