Maximum Classifier Discrepancy for Unsupervised Domain Adaptation

Abstract

In this work, we present a method for unsupervised domain adaptation. Many adversarial learning methods train domain classifier networks to distinguish the features as either a source or target and train a feature generator network to mimic the discriminator. Two problems exist with these methods. First, the domain classifier only tries to distinguish the features as a source or target and thus does not consider task-specific decision boundaries between classes. Therefore, a trained generator can generate ambiguous features near class boundaries. Second, these methods aim to completely match the feature distributions between different domains, which is difficult because of each domain's characteristics.To solve these problems, we introduce a new approach that attempts to align distributions of source and target by utilizing the task-specific decision boundaries. We propose to maximize the discrepancy between two classifiers' outputs to detect target samples that are far from the support of the source. A feature generator learns to generate target features near the support to minimize the discrepancy. Our method outperforms other methods on several datasets of image classification and semantic segmentation. The codes are available at https

Multi-Modal Domain Adaptation for Fine-Grained Action Recognition

“…• We test our approach on three domains from EPIC-Kitchens [8], trained end-to-end using I3D [6], and provide the first benchmark of UDA for fine-grained action recognition. Our results show that MM-SADA outperforms source-only generalisation as well as alternative domain adaptation strategies such as batch-based normalisation [29], distribution discrepancy minimisation [32] and classifier discrepancy [45].…”

“…We show that the self-supervision task of predicting the correspondence of multiple modalities is an effective domain adaptation method. On its own, this can outperform domain alignment methods [32,45], by jointly optimising for the self-supervised task over both domains. Together with adversarial training, the proposed approach outperforms nonadapated models by 4.8%.…”

Drop to Adapt: Learning Discriminative Features for Unsupervised Domain Adaptation

“…Several works [26,9,41] exploit semi-supervised learning for domain adaptation. Besides, MCD [38] and ADR [37] use a minimax training method to push target feature distributions away from the decision boundary, where both methods are composed of the feature extractor and the classifiers. More precisely, in [37], two different classifiers are sampled via stochastic dropout.…”

“…Unlike the prior arts [38,37], the proposed algorithm leverages a unified objective function to optimize all network parameters. The overall loss function is defined as a weighted sum of four objective functions:…”

“…The main assumption in domain adversarial training is that if the feature representation is domaininvariant, a classifier trained on the source domain's features will operate on the target domain as well. However, the weaknesses of domain adversarial methods have been pointed out in [38,37,41]. Since the domain discriminator simply aligns source and target features without considering the class labels, it is likely that the resulting features will not only be domain-invariant, but also non-discriminative with respect to class labels.…”