Fishr: Invariant Gradient Variances for Out-of-distribution Generalization

• URL: http://arxiv.org/abs/2109.02934v1
• Date: Tue, 7 Sep 2021 08:36:09 GMT
• Title: Fishr: Invariant Gradient Variances for Out-of-distribution Generalization
• Authors: Alexandre Rame, Corentin Dancette, Matthieu Cord
• Abstract summary: Fishr is a learning scheme to enforce domain invariance in the space of the gradients of the loss function. Fishr exhibits close relations with the Fisher Information and the Hessian of the loss. In particular, Fishr improves the state of the art on the DomainBed benchmark and performs significantly better than Empirical Risk Minimization.
• Score: 98.40583494166314
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Learning robust models that generalize well under changes in the data distribution is critical for real-world applications. To this end, there has been a growing surge of interest to learn simultaneously from multiple training domains - while enforcing different types of invariance across those domains. Yet, all existing approaches fail to show systematic benefits under fair evaluation protocols. In this paper, we propose a new learning scheme to enforce domain invariance in the space of the gradients of the loss function: specifically, we introduce a regularization term that matches the domain-level variances of gradients across training domains. Critically, our strategy, named Fishr, exhibits close relations with the Fisher Information and the Hessian of the loss. We show that forcing domain-level gradient covariances to be similar during the learning procedure eventually aligns the domain-level loss landscapes locally around the final weights. Extensive experiments demonstrate the effectiveness of Fishr for out-of-distribution generalization. In particular, Fishr improves the state of the art on the DomainBed benchmark and performs significantly better than Empirical Risk Minimization. The code is released at https://github.com/alexrame/fishr.

Related papers

• Normalization Perturbation: A Simple Domain Generalization Method for Real-World Domain Shifts [133.99270341855728]
  Real-world domain styles can vary substantially due to environment changes and sensor noises. Deep models only know the training domain style. We propose Normalization Perturbation to overcome this domain style overfitting problem.
  arXiv  Detail & Related papers  (2022-11-08T17:36:49Z)
• Forget Less, Count Better: A Domain-Incremental Self-Distillation Learning Benchmark for Lifelong Crowd Counting [51.44987756859706]
  Off-the-shelf methods have some drawbacks to handle multiple domains. Lifelong Crowd Counting aims at alleviating the catastrophic forgetting and improving the generalization ability.
  arXiv  Detail & Related papers  (2022-05-06T15:37:56Z)
• Self-balanced Learning For Domain Generalization [64.99791119112503]
  Domain generalization aims to learn a prediction model on multi-domain source data such that the model can generalize to a target domain with unknown statistics. Most existing approaches have been developed under the assumption that the source data is well-balanced in terms of both domain and class. We propose a self-balanced domain generalization framework that adaptively learns the weights of losses to alleviate the bias caused by different distributions of the multi-domain source data.
  arXiv  Detail & Related papers  (2021-08-31T03:17:54Z)
• Domain Generalization via Gradient Surgery [5.38147998080533]
  In real-life applications, machine learning models often face scenarios where there is a change in data distribution between training and test domains. In this work, we characterize the conflicting gradients emerging in domain shift scenarios and devise novel gradient agreement strategies.
  arXiv  Detail & Related papers  (2021-08-03T16:49:25Z)
• SAND-mask: An Enhanced Gradient Masking Strategy for the Discovery of Invariances in Domain Generalization [7.253255826783766]
  We propose a masking strategy, which determines a continuous weight based on the agreement of gradients that flow in each edge of network. SAND-mask is validated over the Domainbed benchmark for domain generalization.
  arXiv  Detail & Related papers  (2021-06-04T05:20:54Z)
• Gradient Matching for Domain Generalization [93.04545793814486]
  A critical requirement of machine learning systems is their ability to generalize to unseen domains. We propose an inter-domain gradient matching objective that targets domain generalization. We derive a simpler first-order algorithm named Fish that approximates its optimization.
  arXiv  Detail & Related papers  (2021-04-20T12:55:37Z)
• Learning Invariant Representations and Risks for Semi-supervised Domain Adaptation [109.73983088432364]
  We propose the first method that aims to simultaneously learn invariant representations and risks under the setting of semi-supervised domain adaptation (Semi-DA) We introduce the LIRR algorithm for jointly textbfLearning textbfInvariant textbfRepresentations and textbfRisks.
  arXiv  Detail & Related papers  (2020-10-09T15:42:35Z)
• Fisher Deep Domain Adaptation [41.50519723389471]
  Deep domain adaptation models learn a neural network in an unlabeled target domain by leveraging the knowledge from a labeled source domain. A Fisher loss is proposed to learn discriminative representations which are within-class compact and between-class separable.
  arXiv  Detail & Related papers  (2020-03-12T06:17:48Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.