Logit-Origin Centering for Singleton Test-Time Adaptation

2026-08-02

AI Summary

This work addresses the performance degradation of deep models under distribution shift in streaming singleton test scenarios by proposing a lightweight test-time adaptation method that requires neither labels, parameter updates, nor prior estimation. The key innovation lies in formulating singleton test-time adaptation as an identifiability problem for the first time and introducing the Prequential Logit-Origin Centering (PLOC) strategy, which adaptively centers logits using only a running mean in logit space. By applying static or dynamic logit shifts while freezing the source model and preserving rank consistency, PLOC is compatible with diverse architectures such as MLP, FT-Transformer, and TabTransformer. Extensive experiments across five benchmark datasets, three model types, and multiple source checkpoints demonstrate that PLOC significantly outperforms existing baselines for tabular data and entropy-based adaptation methods, with its static variant rigorously maintaining the source model’s AUROC.

Abstract

Tabular data is used extensively in many real-world use cases. Deep learning models have been developed to deal with tabular data, but generally perform poorly when the test data distribution differs from that of the training data. Researchers have proposed test-time adaptation approaches to deal with this problem. The fully test-time adaptation (FTTA) setting involves adapting deployed classifiers to shifted target distributions using only unlabeled test data. Leading FTTA methods inherit a batch-dependent approach from computer vision literature. This paper demonstrates for the first time that such approaches degrade sharply in strict streaming regimes where examples arrive and must be classified one at a time. This occurs because at a batch size of one, batch-level statistics become unavailable or poorly estimated. We argue that singleton tabular FTTA is not merely a small-batch variant of ordinary FTTA, but a distinct identifiability problem where only the location of the model's score stream remains directly observable. To address this, we propose Prequential Logit-Origin Centering (PLOC), a lightweight approach that keeps the source model frozen and shifts the logit space at each step. PLOC stores only a single running number (the mean of past logits), requires no labels, estimates no priors, and bypasses weight updates entirely. A deferred variant applies a static shift that preserves the source ranking, and thus the AUROC, exactly. Evaluated across five tabular benchmarks, three architectures (MLP, FT-Transformer, and TabTransformer), and five independent source checkpoints, PLOC significantly outperforms strong tabular and entropy-based baselines.

Problem

• test-time adaptation
• singleton
• tabular data
• distribution shift
• streaming

Innovation

• test-time adaptation
• tabular data
• singleton streaming
• logit centering
• frozen model