Skip to content

Stay Rational, Resist Illusions: De-biasing Source-Free Domain Adaptation with CLIP-Verified Pseudo-Source Domain

Sep 2026 · Machine-mediated learning · Vol 115 · 0 citations · 67 references

TL;DR

A novel Staying Rational and Resisting Illusions (SRRI) method for SFDA to resolve the two fundamental challenges: confirmation bias and domain shift, and innovatively leverage the CLIP (Contrastive Language-Image Pre-training) model as a source of external knowledge to mitigate feature confounding.

View source

Similar papers

#artificial intelligence Preprint Oct 2026

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

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....

Seong-Whan Lee, Changhee Lee · 0 citations
Open access Aug 2026

Towards source-free domain adaptation: a model-heterogeneous perspective

A Peer-level Heterogeneous Perception Framework is proposed that departs from such paradigms by enabling balanced collaboration between heterogeneous models by introducing an auxiliary domain that is significantly different from the target domain and employ an auxiliary model with the same architecture as the source mo...

Zhi-Ze Wu, Yu-Tao Fu, Huan-Xin Zou et al. · 0 citations
Open access Sep 2026

A novel source-free domain adaptation with high-confidence sample selection and feature disentanglement for machinery fault diagnosis

This paper proposes a novel SFDA with high-confidence sample selection and feature disentanglement for machinery fault diagnosis, which effectively alleviates the adverse influence of noisy pseudo-labels during the stage of adaptation.

Yi-Ming Yuan, Kang Wu, Xing-Xing Jiang et al. · 0 citations
Aug 2026

Adaptive Class-wise Multicentric Prototype Source-Free Domain Adaptation for Privacy-Preserving BCIs.

Recently, conventional domain adaptation (DA) methods have demonstrated promising performance in cross-subject classification in electroencephalogram (EEG)-based brain-computer interfaces (BCIs). However, these methods require direct access to labeled source subject data, potentially compromising biometric privacy. Sou...

Yi Yang, Chen-Xi Sun, Ruoning Lyu et al. · 0 citations
Conference Aug 2026

Cross-view consistent teacher for source-free object detection

Experimental results show that the proposed SFDA object detection framework for the YOLO family of single-stage detectors consistently outperforms the baseline on multiple cross-domain detection tasks, demonstrating its effectiveness and good generalization ability under the source-free setting.

Ruixue Gao, Xiaoyao Yang, Shujuan Hou · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.