Adversarially Adaptive Normalization for Single Domain Generalization

Abstract

Single domain generalization aims to learn a model that performs well on many unseen domains with only one domain data for training. Existing works focus on studying the adversarial domain augmentation (ADA) to improve the model's generalization capability. The impact on domain generalization of the statistics of normalization layers is still underinvestigated. In this paper, we propose a generic normalization approach, adaptive standardization and rescaling normalization (ASR-Norm), to complement the missing part in previous works. ASR-Norm learns both the standardization and rescaling statistics via neural networks. This new form of normalization can be viewed as a generic form of the traditional normalizations. When trained with ADA, the statistics in ASR-Norm are learned to be adaptive to the data coming from different domains, and hence improves the model generalization performance across domains, especially on the target domain with large discrepancy from the source domain. The experimental results show that ASR-Norm can bring consistent improvement to the state-of-the-art ADA approaches by 1.6%, 2.7%, and 6.3% averagely on the Digits, CIFAR-10-C, and PACS benchmarks, respectively. As a generic tool, the improvement introduced by ASR-Norm is agnostic to the choice of ADA methods. * The main work was done during an internship at Google Research. Figure 1: Illustration of single domain generalization with the PACS [30] benchmark. The dataset contains 4 domains: art paint, cartoon, sketch, and photo domains, which share the same categories that include dog, elephant, giraffe, guitar, house, horse, and person. Single domain generalization aims at training a model on one source domain data (art paint domain in the shown case), while generalizing well to other domains with very different visual presentations.

Domain Generalization for Road Scene Object Detection under Different Weather Conditions

“…They improved the model's generalization by using four steps: local feature decomposition, local feature addressing, elemental feature synthesis, and elemental feature learning. Fan et al from the University of Texas proposed a universal normalization method [15] that enhances generalization ability through adaptive standardization and re-scaling normalization. Although these methods are effective for image classification, they are not directly applicable to domain generalization for road scene object detection under different weather conditions, as object detection involves both localization and classification tasks.…”