Contrastive Test-Time Adaptation

2022/04/21 by Dian Chen, Chen, Dian, Dequan Wang +5 · 29 citations
Computer Science · #Computer Vision and Pattern Recognition (cs.CV) #Domain Adaptation and Few-Shot Learning #FOS: Computer and information sciences #Multimodal Machine Learning Applications #Speech Recognition and Synthesis

paper · pdf · doi:10.48550/arxiv.2204.10377

openalex publication_date 2022/04/21 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Test-time adaptation is a special setting of unsupervised domain adaptation where a trained model on the source domain has to adapt to the target domain without accessing source data. We propose a novel way to leverage self-supervised contrastive learning to facilitate target feature learning, along with an online pseudo labeling scheme with refinement that significantly denoises pseudo labels. The contrastive learning task is applied jointly with pseudo labeling, contrasting positive and negative pairs constructed similarly as MoCo but with source-initialized encoder, and excluding same-class negative pairs indicated by pseudo labels. Meanwhile, we produce pseudo labels online and refine them via soft voting among their nearest neighbors in the target feature space, enabled by maintaining a memory queue. Our method, AdaContrast, achieves state-of-the-art performance on major benchmarks while having several desirable properties compared to existing works, including memory efficiency, insensitivity to hyper-parameters, and better model calibration. Project page: sites.google.com/view/adacontrast.

Cited by

• Progressive Conditioned Scale-Shift Recalibration of Self-Attention for Online Test-time Adaptation
• Towards Reliable Test-Time Adaptation: Style Invariance as a Correctness Likelihood
• Collaborative Learning with Multiple Foundation Models for Source-Free Domain Adaptation
• Mitigating Long-Tail Bias in HOI Detection via Adaptive Diversity Cache
• CoTBox-TTT: Grounding Medical VQA with Visual Chain-of-Thought Boxes During Test-time Training
• Groupwise Registration with Physics-Informed Test-Time Adaptation on Multi-parametric Cardiac MRI
• Robust and Generalizable Atrial Fibrillation Detection from ECG Using Time-Frequency Fusion and Supervised Contrastive Learning
• EdgeSync: Accelerating Edge-Model Updates for Data Drift through Adaptive Continuous Learning
• Diffusion-Driven Progressive Target Manipulation for Source-Free Domain Adaptation
• Attention Residual Fusion Network with Contrast for Source-free Domain Adaptation
• Buffer layers for Test-Time Adaptation
• DP-TTA: Test-time Adaptation for Transient Electromagnetic Signal Denoising via Dictionary-driven Prior Regularization
• Class-aware Domain Knowledge Fusion and Fission for Continual Test-Time Adaptation
• Test-Time Adaptation by Causal Trimming
• Self-Training with Dynamic Weighting for Robust Gradual Domain Adaptation
• Source-Free Cross-Domain Continual Learning
• Test-Time Matching: Unlocking Compositional Reasoning in Multimodal Models
• Consistent Assistant Domains Transformer for Source-free Domain Adaptation
• TRUST: Test-Time Refinement using Uncertainty-Guided SSM Traverses
• A Tale of Two Experts: Cooperative Learning for Source-Free Unsupervised Domain Adaptation
• POEM: Explore Unexplored Reliable Samples to Enhance Test-Time Adaptation
• COLA: Context-aware Language-driven Test-time Adaptation
• HyperTTA: Test-Time Adaptation for Hyperspectral Image Classification under Distribution Shifts
• Adapt in the Wild: Test-Time Entropy Minimization with Sharpness and Feature Regularization
• ADVMEM: Adversarial Memory Initialization for Realistic Test-Time Adaptation via Tracklet-Based Benchmarking
• Stabilizing Open-Set Test-Time Adaptation via Primary-Auxiliary Filtering and Knowledge-Integrated Prediction
• Learn Faster and Remember More: Balancing Exploration and Exploitation for Continual Test-time Adaptation
• ETA: Energy-based Test-time Adaptation for Depth Completion
• Adapting Vision-Language Models Without Labels: A Comprehensive Survey

Related

• Transferrable Contrastive Learning for Visual Domain Adaptation
• Domain-Specific Block Selection and Paired-View Pseudo-Labeling for Online Test-Time Adaptation
• Low-confidence Samples Matter for Domain Adaptation
• When Test-Time Adaptation Meets Self-Supervised Models
• Divide and Contrast: Source-free Domain Adaptation via Adaptive Contrastive Learning