Safeguarding Mutual Correction in Source-Free Domain Adaptation via Cut Statistics

2026-10-02

AI Summary

This study addresses the challenges of confirmation bias hindering error correction and cross-model supervision propagating errors in source-free domain adaptation. To this end, we propose SafeCut, a method that employs cutting statistics as a reliability measure for unlabeled samples. Departing from conventional unidirectional supervision paradigms, SafeCut establishes a bidirectional mutual correction mechanism between vision-language models and pre-trained models, dynamically modulating the direction and intensity of corrections via reliability gating. We theoretically prove that this mechanism guarantees a net positive correction signal, enabling precise sample-level mutual correction. Extensive experiments demonstrate that SafeCut achieves state-of-the-art performance across multiple benchmarks by effectively suppressing erroneous corrections while amplifying genuine ones.

Abstract

Source-Free Domain Adaptation (SFDA) aims to adapt a source-pretrained model to an unlabeled target domain without access to the original source domain. While early single-model approaches rely on self-refinement, they are inherently susceptible to confirmation bias and struggle to correct their own systematic errors. To overcome this limitation, recent methods introduce Vision-Language (ViL) models as external knowledge sources. However, these approaches operate in a largely unidirectional paradigm, using the ViL model primarily to supervise the source-pretrained model. This overlooks a key structural property: the two models exhibit distinct failure modes -- where one produces an incorrect prediction, the other may produce a correct one, creating a natural opportunity for mutual correction within the target domain. Yet, without ground-truth labels, identifying which model is correct on any given sample is non-trivial, and naively exchanging predictions risks propagating errors across models. To address this challenge, we propose SafeCut, a novel approach that leverages the cut statistic as a label-free measure of prediction reliability to gate cross-model supervision. Our approach dynamically controls both the direction and strength of supervision based on relative reliability, selectively amplifying true corrections while suppressing miscorrections on a per-sample basis. We further provide theoretical justification showing that this reliability-gated mechanism guarantees a net-positive correction signal. Extensive experiments across diverse SFDA benchmarks demonstrate that SafeCut achieves state-of-the-art performance, highlighting the effectiveness of safeguarding mutual correction in SFDA via cut statistics.

Problem

• Source-Free Domain Adaptation
• Mutual Correction
• Confirmation Bias
• Vision-Language Models
• Error Propagation

Innovation

• Source-Free Domain Adaptation
• Mutual Correction
• Cut Statistic
• Vision-Language Models
• Reliability Gating