Backpropagation-Free Test-Time Adaptation via Probabilistic Gaussian Alignment

2025-08-21

AI Summary

Test-time adaptation (TTA) commonly relies on backpropagation, lacks explicit modeling of class-conditional feature distributions, and struggles with zero-shot deployment. Method: We propose Gaussian Probabilistic Alignment (GPA), the first TTA framework reformulated as a closed-form Gaussian probabilistic inference task—requiring neither gradient-based updates nor access to source data. GPA models class-conditional features via Gaussian likelihoods, employs shared covariance estimation, incorporates lightweight regularization guided by CLIP semantic priors, and maintains a history knowledge bank with dynamic updating for both online and transductive learning. Contribution/Results: GPA achieves state-of-the-art performance across diverse distribution shift benchmarks, significantly enhancing zero-shot robustness and deployment scalability without compromising computational efficiency or requiring source-domain supervision.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Security and Privacy: Data transparency and provenance
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Test-time adaptation (TTA) enhances the zero-shot robustness under distribution shifts by leveraging unlabeled test data during inference. Despite notable advances, several challenges still limit its broader applicability. First, most methods rely on backpropagation or iterative optimization, which limits scalability and hinders real-time deployment. Second, they lack explicit modeling of class-conditional feature distributions. This modeling is crucial for producing reliable decision boundaries and calibrated predictions, but it remains underexplored due to the lack of both source data and supervision at test time. In this paper, we propose ADAPT, an Advanced Distribution-Aware and backPropagation-free Test-time adaptation method. We reframe TTA as a Gaussian probabilistic inference task by modeling class-conditional likelihoods using gradually updated class means and a shared covariance matrix. This enables closed-form, training-free inference. To correct potential likelihood bias, we introduce lightweight regularization guided by CLIP priors and a historical knowledge bank. ADAPT requires no source data, no gradient updates, and no full access to target data, supporting both online and transductive settings. Extensive experiments across diverse benchmarks demonstrate that our method achieves state-of-the-art performance under a wide range of distribution shifts with superior scalability and robustness.

Problem

• Addresses test-time adaptation without backpropagation for efficiency
• Models class-conditional feature distributions for reliable predictions
• Eliminates need for source data and gradient updates

Innovation

• Backpropagation-free probabilistic Gaussian alignment
• Closed-form training-free inference method
• Lightweight regularization with CLIP priors