Universal Domain Adaptation through Self Supervision

By Kuniaki Saito, Stan Sclaroff, Kate Saenko. 2020

391 citations recorded by the source corpus.

Unsupervised domain adaptation methods traditionally assume that all source categories are present in the target domain. In practice, little may be known about the category overlap between the two domains. While some methods address target settings with either partial or open-set categories, they assume that the particular setting is known a priori. We propose a more universally applicable domain adaptation approach that can handle arbitrary category shift, called Domain Adaptative Neighborhood Clustering via Entropy optimization (DANCE). DANCE combines two novel ideas: First, as we cannot fully rely on source categories to learn features discriminative for the target, we propose a novel neighborhood clustering technique to learn the structure of the target domain in a self-supervised way. Second, we use entropy-based feature alignment and rejection to align target features with the source, or reject them as unknown categories based on their entropy. We show through extensive experiments that DANCE outperforms baselines across open-set, open-partial and partial domain adaptation settings.

Subfields: Artificial Intelligence, Cancer Research, 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 Cancer Research

In Computer Vision and Pattern Recognition

Citations in the corpus

Prototype-Guided Continual Adaptation for Class-Incremental Unsupervised Domain Adaptation (2022, background): “C→I Step 1 [0, 2, 7, 9, 11, 27, 28, 29, 30, 33] Step 2 [37, 39, 40, 44, 45, 47, 50, 60, 62, 68] Step 3 [71, 75, 76, 82, 85, 86, 87, 88, 89, 90] Step 4 [92, 94, 96, 97, 106, 107, 108, 109, 110, 112] Step 5 [114, 115, 116, 123, 126, 128, 133, 134, 141, 145] Step 6 [146, 150, 151, 157, 160, 163, 165, 1”
Prototypical Pseudo Label Denoising and Target Structure Learning for Domain Adaptive Semantic Segmentation (2021, background): “Early efforts dedicate to designing pretext tasks [16, 18, 29, 70], which are proven beneficial for UDA when utilized as auxiliary tasks on target data [47, 53, 66].”
A Comprehensive Survey on Test-Time Adaptation Under Distribution Shifts (2023, positive): “Inspired by neighborhood clustering [300], CPGA [144] and StickerDA [145] calculate the normalized instance-to-instance similarity with a memory bank {ḡ(xi)} i=1 and thus minimize the following entropy,”
Semi-Supervised and Unsupervised Deep Visual Learning: A Survey (2022, methodology): “For instance, DANCE [170] performs neighborhood clustering by assigning the target samples to a “known” class prototype in the source domain or its neighbor in the target domain.”
Unsupervised Domain Adaptation in Semantic Segmentation via Orthogonal and Clustered Embeddings (2020, methodology): “Avoiding the need for target pseudo-labels, [43, 44] propose a self-supervised clustering technique to discover target modes without any form of supervision.”
A Survey on Negative Transfer (2020, background): “[109] proposed a universal DA network, which quantifies instance-level transferability to discover the common and private label sets, and promotes adaptation in the common label set.”
Domain Consensus Clustering for Universal Domain Adaptation (2021, background): “To achieve this goal, many efforts have been devoted to performing common sample discovery from different perspectives, such as designing new criteria [4, 3, 62, 17, 53] or introducing extra discriminator [63, 5, 38, 9].”
Contrastive Test-Time Adaptation (2022, background): “[46, 49] tackle the unsupervised domain adaptation (UDA) setting where concurrent access to source and target data is allowed.”
A Survey on Self-Supervised Learning: Algorithms, Applications, and Future Trends (2023, methodology): “It has been combined with clustering [68], semi-supervised learning [145], multi-task learning [157], [158], transfer learning [159], [160], [161], graph NNs [147], [162], [163], reinforcement learning [164], [165], [166], few-shot learning [167], [168], neural architecture search [169], robust…”
Exploiting the Intrinsic Neighborhood Structure for Source-free Domain Adaptation (2021, methodology): “To retrieve the nearest neighbors for batch training, similar to [33, 50, 62], we build two memory banks: F stores all target features, and S stores corresponding prediction scores:”