Divergence Optimization for Noisy Universal Domain Adaptation

TL;DR

This paper introduces a divergence-based method for Noisy Universal Domain Adaptation, effectively handling noisy source data and unknown target classes, outperforming existing approaches across various settings.

Contribution

It proposes a novel two-head neural network framework that detects noisy samples, identifies unknown classes, and aligns domain distributions in Noisy UniDA.

Findings

Outperforms existing methods in most adaptation settings

Effectively detects noisy source samples

Identifies unknown classes in target domain

Abstract

Universal domain adaptation (UniDA) has been proposed to transfer knowledge learned from a label-rich source domain to a label-scarce target domain without any constraints on the label sets. In practice, however, it is difficult to obtain a large amount of perfectly clean labeled data in a source domain with limited resources. Existing UniDA methods rely on source samples with correct annotations, which greatly limits their application in the real world. Hence, we consider a new realistic setting called Noisy UniDA, in which classifiers are trained with noisy labeled data from the source domain and unlabeled data with an unknown class distribution from the target domain. This paper introduces a two-head convolutional neural network framework to solve all problems simultaneously. Our network consists of one common feature generator and two classifiers with different decision boundaries. By optimizing the divergence between the two classifiers' outputs, we can detect noisy source samples, find "unknown" classes in the target domain, and align the distribution of the source and target domains. In an extensive evaluation of different domain adaptation settings, the proposed method outperformed existing methods by a large margin in most settings.