Universal Domain Adaptation

By Kaichao You, Mingsheng Long, Zhangjie Cao, Jianmin Wang, Michael I. Jordan. 2019

576 citations recorded by the source corpus.

Domain adaptation aims to transfer knowledge in the presence of the domain gap. Existing domain adaptation methods rely on rich prior knowledge about the relationship between the label sets of source and target domains, which greatly limits their application in the wild. This paper introduces Universal Domain Adaptation (UDA) that requires no prior knowledge on the label sets. For a given source label set and a target label set, they may contain a common label set and hold a private label set respectively, bringing up an additional category gap. UDA requires a model to either (1) classify the target sample correctly if it is associated with a label in the common label set, or (2) mark it as ``unknown'' otherwise. More importantly, a UDA model should work stably against a wide spectrum of commonness (the proportion of the common label set over the complete label set) so that it can handle real-world problems with unknown target label sets. To solve the universal domain adaptation problem, we propose Universal Adaptation Network (UAN). It quantifies sample-level transferability to discover the common label set and the label sets private to each domain, thereby promoting the adaptation in the automatically discovered common label set and recognizing the ``unknown'' samples successfully. A thorough evaluation shows that UAN outperforms the state of the art closed set, partial and open set domain adaptation methods in the novel UDA setting.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

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 Computer Vision and Pattern Recognition

Citations in the corpus

Prototype-Guided Continual Adaptation for Class-Incremental Unsupervised Domain Adaptation (2022, methodology): “Uni-DA [53] assumes that the target label space is not limited to the source label space and may contain target private (unknown) classes.”
Class-Incremental Domain Adaptation (2020, methodology): “We compare against prior UDA methods DANN [11], OSBP [45], UAN [54], STA [29], and CI methods E2E [4], LETR [33], iCaRL [41], LwF-MC [7], LwM [7].”
Supplementary : Learning to Detect Open Classes for Universal Domain Adaptation (2020, methodology): “Following UAN [13], in alphabet order, we use the ﬁrst 10 classes as C , the next 5 classes as C s and the rest as C t .”
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.”
A Comprehensive Survey on Test-Time Adaptation Under Distribution Shifts (2023, unknown): “To allow more flexibility, open-partial-set domain adaptation [351] (Cs \Ct 6= ∅, Ct \Cs 6= ∅,) is studied in SFDA methods [11], [220], [309], [333], [352].”
Domain Consensus Clustering for Universal Domain Adaptation (2021, methodology): “Unlike previous works [62, 17] that perform sample-level identification on the common samples, this paper aims to mine both common classes and individual private classes simultaneously with discriminative clusters.”
Universal Source-Free Domain Adaptation (2020, methodology): “Further, we aim to exploit the learned generative classifier from the Procurement stage to complement for the purpose of separate ad-hoc networks (critic or discriminator) as utilized by the prior works [52, 6].”
A Brief Review of Domain Adaptation (2020, background): “• Universal domain adaptation (UDA) [24], generalizes the above scenarios.”
Heterogeneous Domain Generalization Via Domain Mixup (2020, methodology): “In UDA, labels are divided into two parts, common label set and private label set.”
Boundary-aware Backward-Compatible Representation via Adversarial Learning in Image Retrieval (2023, methodology): “The adversarial learning is applied to domain adaptation [21,26,29] to guide the embedding model to generate domain-insensitive embeddings.”
Applications of Unsupervised Deep Transfer Learning to Intelligent Fault Diagnosis: A Survey and Comparative Study (2019, methodology): “[148] proposed UAN and designed an instance-level transferability criterion, exploiting the domain similarity and prediction uncertainty.”
Domain Generalization: A Survey (2021, unknown): “DG, such as single- [150] and multi-source [60] DA, and heterogeneous DA [17], [156], [157], [158].”
Overcoming Negative Transfer: A Survey (2020, methodology): “You et al. [52] assumed a more realistic scenario that the target label space has some common classes with the source, but they also have their own private classes.”
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].”