Federated Adversarial Domain Adaptation

By Xingchao Peng, Zijun Huang, Yizhe Zhu, Kate Saenko. 2019

319 citations recorded by the source corpus.

Federated learning improves data privacy and efficiency in machine learning performed over networks of distributed devices, such as mobile phones, IoT and wearable devices, etc. Yet models trained with federated learning can still fail to generalize to new devices due to the problem of domain shift. Domain shift occurs when the labeled data collected by source nodes statistically differs from the target node's unlabeled data. In this work, we present a principled approach to the problem of federated domain adaptation, which aims to align the representations learned among the different nodes with the data distribution of the target node. Our approach extends adversarial adaptation techniques to the constraints of the federated setting. In addition, we devise a dynamic attention mechanism and leverage feature disentanglement to enhance knowledge transfer. Empirically, we perform extensive experiments on several image and text classification tasks and show promising results under unsupervised federated domain adaptation setting.

Subfields: Artificial Intelligence, Radiology, Nuclear Medicine and Imaging

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Jian Liang
2. Qi-Yan Dou
3. P. Heng
4. Jing Qin
5. Hangyu Zhu
6. Dapeng Hu
7. Faisal Mahmood
8. Ran He

In Radiology, Nuclear Medicine and Imaging

1. P. Heng
2. Jing Qin
3. Qi-Yan Dou
4. Jian Liang
5. Cheng Chen
6. Zhen Fang
7. Quande Liu
8. Ran He

Citations in the corpus

• Federated Class-Incremental Learning (2022, background): “[34] integrates unsupervised domain adaptation [5, 9, 27, 28, 57] into federated learning framework [14, 31].”
• Multi-site fMRI Analysis Using Privacy-preserving Federated Learning and Domain Adaptation: ABIDE Results (2020, methodology): “To address this issue, we employed federated adversarial alignment [29] that introduces two modules (a domain-specific local feature extractor, and a global discriminator) in the classification networks and divides optimization into two independent steps.”
• A Comprehensive Survey on Test-Time Adaptation Under Distribution Shifts (2023, unknown): “Inspired by federated learning [45], several studies [46], [47], [48] on federated domain adaptation offer an alternative solution for test-time domain adaptation that acquires the feedback (e.”
• Federated Learning on Non-IID Data: A Survey (2021, methodology): “A federated adversarial domain adaptation (FADA) algorithm is presented in [105], which uses adversarial adaptation techniques to solve the domain shift problem in FL systems.”
• Federated Adversarial Debiasing for Fair and Transferable Representations (2021, methodology): “A recent effort is done by [38] where embeddings of different groups are shared (see Fig.”
• Learn from Others and Be Yourself in Heterogeneous Federated Learning (2022, background): “Latest researches have studied related problems of unsupervised domain adaptation for target domain [45, 65] and domain generalization on unseen domains [52].”
• FedDG: Federated Domain Generalization on Medical Image Segmentation via Episodic Learning in Continuous Frequency Space (2021, background): “Latest literature has studied a related problem of unsupervised domain adaptation in FL paradigm [28, 41], whereas these methods typically require data from the target domain to adapt the model.”
• Source Data-Absent Unsupervised Domain Adaptation Through Hypothesis Transfer and Labeling Transfer (2020, background): “There are also two concurrent deep UDA methods [64], [65] that attempt not to access the source data during the adaptation process.”
• Data-Free Knowledge Distillation for Heterogeneous Federated Learning (2021, background): “Specifically, LT (h) is usually considered as an ideal upperbound for the global model in FL (Peng et al., 2019; Lin et al., 2020).”
• Do We Really Need to Access the Source Data? Source Hypothesis Transfer for Unsupervised Domain Adaptation (2020, background): “Note that FADA (Peng et al., 2020) does not access to the source domain but utilizes the gradient update information like federated learning, but it performs much worse than our SHOT for the multi-source setting.”
• Algorithm fairness in artificial intelligence for medicine and healthcare (2023, methodology): “As a result, many federated-learning approaches developed for learning in non-i.i.d. scenarios inherently adapt dataset shift and bias-mitigation techniques to resolve local and global drift in tandem with evaluating fairness 236,237,239,245,246,253,254 .”