Universal Source-Free Domain Adaptation

By Jogendra Nath Kundu, Naveen Venkat, V. RahulM., R. Venkatesh Babu. Computer Vision and Pattern Recognition, 2020

404 citations recorded by the source corpus.

There is a strong incentive to develop versatile learning techniques that can transfer the knowledge of class-separability from a labeled source domain to an unlabeled target domain in the presence of a domain-shift. Existing domain adaptation (DA) approaches are not equipped for practical DA scenarios as a result of their reliance on the knowledge of source-target label-set relationship (e.g. Closed-set, Open-set or Partial DA). Furthermore, almost all prior unsupervised DA works require coexistence of source and target samples even during deployment, making them unsuitable for real-time adaptation. Devoid of such impractical assumptions, we propose a novel two-stage learning process. 1) In the Procurement stage, we aim to equip the model for future source-free deployment, assuming no prior knowledge of the upcoming category-gap and domain-shift. To achieve this, we enhance the model’s ability to reject out-of-source distribution samples by leveraging the available source data, in a novel generative classifier framework. 2) In the Deployment stage, the goal is to design a unified adaptation algorithm capable of operating across a wide range of category-gaps, with no access to the previously seen source samples. To this end, in contrast to the usage of complex adversarial training regimes, we define a simple yet effective source-free adaptation objective by utilizing a novel instance-level weighting mechanism, named as Source Similarity Metric (SSM). A thorough evaluation shows the practical usability of the proposed learning framework with superior DA performance even over state-of-the-art source-dependent approaches.

Subfields: Artificial Intelligence, Epidemiology, Infectious Diseases

View Paper: https://www.semanticscholar.org/paper/611b5f9f868c9e820d2c4cd8e8a9bb4274eb3a8f

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

In Epidemiology

In Infectious Diseases

Citations in the corpus

Generalized Source-free Domain Adaptation (2021, background): “USFDA [14] and FS [15] explore the source-free universal DA [48] and open-set DA [32], DECISION [2] is for multi-source DA.”
Continual Test-Time Domain Adaptation (2022, background): “Test-time adaptation is also referred to as source-free domain adaptation in some references [28,66].”
ViDA: Homeostatic Visual Domain Adapter for Continual Test Time Adaptation (2023, background): “Test-time adaptation (TTA), also referred to as source-free domain adaptation [6, 35, 41, 62], aims to adapt a source model to an unknown target domain distribution without relying on any source domain data.”
Tent: Fully Test-Time Adaptation by Entropy Minimization (2021, methodology): “Li et al. (2020); Kundu et al. (2020) rely on generative modeling and optimize multiple models with multiple losses.”
Class-Incremental Domain Adaptation (2020, background): “in a self-driving car), or simply unavailable during model upgrade [7,23,28].”
Towards Stable Test-Time Adaptation in Dynamic Wild World (2023, methodology): “…UDA methods have been proposed to resolve the adaptation problem when source data are absent, such as generative-based methods that generate source images or prototypes from the model (Li et al., 2020; Kundu et al., 2020; Qiu et al., 2021), and information maximization (Liang et al., 2020).”
Test-Time Classifier Adjustment Module for Model-Agnostic Domain Generalization (2021, background): “Among them, recent “source-free” setups are particularly similar to our setting [30, 25, 34].”
Efficient Test-Time Model Adaptation without Forgetting (2022, methodology): “Moreover, we propose an anti-forgetting regularizer to prevent the model parameters Θ changing too much from Θ o . modeling (Li et al., 2020; Kundu et al., 2020; Qiu et al., 2021) or information maximization (Liang et al., 2020).”
Exploiting the Intrinsic Neighborhood Structure for Source-free Domain Adaptation (2021, methodology): “Among these methods, USFDA [16] addresses universal DA [57] and SF [17] addresses open-set DA [36].”
TTT++: When Does Self-Supervised Test-Time Training Fail or Thrive? (2021, background): “Adapting machine learning models based on test samples has garnered growing interests in both generative problems such as super-resolution [13], image synthesis [14] and image manipulation [15], and discriminative problems like image classiﬁcation [16].”
Give Me Your Trained Model: Domain Adaptive Semantic Segmentation without Source Data (2021, background): “[31] proposes a universal source data-free domain adaptation when the knowledge of the label space in the target domain is not available.”
Non-Transferable Learning: A New Approach for Model Ownership Verification and Applicability Authorization (2021, background): “In addition, recently, there are domain adaptation (DA) works on adapting the domain-shared knowledge within the source model to the target one without access to the source data (Liang et al., 2020; Ahmed et al., 2021; Kundu et al., 2020; Wang et al., 2021).”
Domain Generalization: A Survey (2021, background): “One one hand, TTT (mostly) bears a resemblance with source-free DA [160]— both assume source data is inaccessible after model training.”
Source-Free Domain Adaptation for Semantic Segmentation (2021, methodology): “Recently, a tiny number of source-free UDA methods [25, 24, 27, 38, 22, 26] have been developed to tackle a similar issue on image classification.”
Contrastive Test-Time Adaptation (2022, background): “In this work we focus on the problem of test-time [50, 56] or source-free [23, 28, 58] domain adaptation where the source data is no longer available during adapting to unlabeled test data.”