On Learning Invariant Representations for Domain Adaptation

By Han Zhao, Rémi Tachet des Combes, Kun Zhang, Geoffrey J. Gordon. International Conference on Machine Learning, 2019

558 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/acb9017f0d55ab0e8c8863b0b5dbdb373e4aee7b

How researchers see this paper

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

Citations in the corpus

Environment Inference for Invariant Learning (2020, background): “However, this approach has limitations in settings where distribution shift affects the marginal distribution over labels (Zhao et al., 2019).”
Domain Aggregation Networks for Multi-Source Domain Adaptation (2019, background): “Recent analysis (Zhao et al., 2019; Johansson et al., 2019) show that solely focusing on learning domain invariant features can be insufﬁcient when the marginal label distributions are signiﬁcantly different.”
A Review of Single-Source Deep Unsupervised Visual Domain Adaptation (2020, background): “In this line of work, [90–94] proposed generalization bounds for this scenario and focused”
Understanding Self-Training for Gradual Domain Adaptation (2020, background): “The traditional solution to adapt to the distribution shift is unsupervised domain adaptation, but existing methods are only guaranteed to work when the distribution shift is small—when the source and target distributions cannot be easily distinguished from each other (Zhao et al., 2019).”
A Survey of Unsupervised Deep Domain Adaptation (2018, background): “It is an open problem as to when the label distributions can be aligned without target labels [280].”
SelfReg: Self-supervised Contrastive Regularization for Domain Generalization (2021, methodology): “It is known that finding a flatter minima guarantees better generalization performance [19], and thus it has been used in domain adaptation and generalization fields that require high generalization performance [48, 6].”
Adversarial target-invariant representation learning for domain generalization (2019, background): “Using the triangle inequality for the H -distance (Zhao et al., 2019) and the fact that ) .”
How Neural Networks Extrapolate: From Feedforward to Graph Neural Networks (2020, methodology): “Using target domain data during training may induce useful non-linearities and may mitigate extrapolation by matching the target and source distributions, though the correctness of the learned mapping depends on the label distribution (Zhao et al., 2019).”
Unsupervised Domain Adaptation via Structurally Regularized Deep Clustering (2020, background): “The recent work of [9] improves adversarial feature adaptation, where the discriminative structures of target data may be deteriorated [69].”
Cluster Alignment With a Teacher for Unsupervised Domain Adaptation (2019, background): “Concurrently, several works [51, 48] analyze the conditional distribution shift and label distribution shift issues in UDA theoretically, but only propose limited solutions.”
The Risks of Invariant Risk Minimization (2020, background): “Further, there is evidence that this condition may not be suﬃcient for eﬀective domain adaptation or OOD generalization (Zhao et al., 2019; Johansson et al., 2019).”
Continuously Indexed Domain Adaptation (2020, background): “Adversarial domain adaptation is particularly popular due to its relatively strong theoretical insights (Goodfellow et al., 2014; Zhao et al., 2018; Zhang et al., 2019; Zhao et al., 2019) and its compatibility with neural networks.”
Maximum Density Divergence for Domain Adaptation (2020, background): “Recently, there are several studies [30], [31] performed on adversarial training which aim to address the divergence in the conditional distribution.”