Test-Time Training with Self-Supervision for Generalization under Distribution Shifts

Abstract

In this paper, we propose Test-Time Training, a general approach for improving the performance of predictive models when training and test data come from different distributions. We turn a single unlabeled test sample into a self-supervised learning problem, on which we update the model parameters before making a prediction. This also extends naturally to data in an online stream. Our simple approach leads to improvements on diverse image classification benchmarks aimed at evaluating robustness to distribution shifts.

arXiv commentICML 2020

References in corpus (32)

• Overcoming catastrophic forgetting in neural networks
• Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
• Prototypical Networks for Few-shot Learning
• Learning Transferable Features with Deep Adaptation Networks
• On Calibration of Modern Neural Networks
• Unsupervised Domain Adaptation with Residual Transfer Networks
• Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examples
• Rethinking the Value of Network Pruning
• Theoretically Principled Trade-off between Robustness and Accuracy
• Meta-SGD: Learning to Learn Quickly for Few-Shot Learning
• Domain Generalization via Invariant Feature Representation
• Certified Defenses against Adversarial Examples
• Countering Adversarial Images using Input Transformations
• Using Pre-Training Can Improve Model Robustness and Uncertainty
• Generalisation in humans and deep neural networks
• Robustness May Be at Odds with Accuracy
• Generalizing Across Domains via Cross-Gradient Training
• Certifying Some Distributional Robustness with Principled Adversarial Training
• Domain Adaptation for Visual Applications: A Comprehensive Survey
• PixelDefend: Leveraging Generative Models to Understand and Defend against Adversarial Examples
• Provably Robust Deep Learning via Adversarially Trained Smoothed Classifiers
• Unsupervised Learning by Predicting Noise
• Unsupervised Domain Adaptation through Self-Supervision
• advertorch v0.1: An Adversarial Robustness Toolbox based on PyTorch
• Do CIFAR-10 Classifiers Generalize to CIFAR-10?
• Algorithms and Theory for Multiple-Source Adaptation
• Rethinking ImageNet Pre-training
• Colorization as a Proxy Task for Visual Understanding
• Episodic Training for Domain Generalization
• Transfer of Adversarial Robustness Between Perturbation Types
• Boosting Supervision with Self-Supervision for Few-shot Learning
• Neural Multi-Scale Self-Supervised Registration for Echocardiogram Dense Tracking

Cited by in corpus (18)

• A Comprehensive Survey on Test-Time Adaptation under Distribution Shifts
• Test-Time Adaptable Neural Networks for Robust Medical Image Segmentation
• Evaluating Prediction-Time Batch Normalization for Robustness under Covariate Shift
• Learning Improvement Heuristics for Solving Routing Problems
• From Google Gemini to OpenAI Q* (Q-Star): A Survey of Reshaping the Generative Artificial Intelligence (AI) Research Landscape
• Robust Policies via Mid-Level Visual Representations: An Experimental Study in Manipulation and Navigation
• Vision-Language Navigation with Self-Supervised Auxiliary Reasoning Tasks
• Robust Image Classification Using A Low-Pass Activation Function and DCT Augmentation
• Multi-Parameter Molecular MRI Quantification using Physics-Informed Self-Supervised Learning
• Unsupervised Domain Adaptation in the Absence of Source Data
• Distributional Anchor Regression
• Domain-Aware Dynamic Networks
• Re-using Adversarial Mask Discriminators for Test-time Training under Distribution Shifts
• Learning to See Through Obstructions
• Uncover and Unlearn Nuisances: Agnostic Fully Test-Time Adaptation
• Self-Supervised Dynamic Networks for Covariate Shift Robustness
• Achieving Adversarial Robustness Requires An Active Teacher
• On-target Adaptation