Published in European Conference on Computer Vision (ECCV), 2026
Reveals why feature magnitude misleads OOD detectors and proposes an angular separation perspective for robust out-of-distribution detection.
Recommended citation: Hanlin Li, Jing Ma, Zehang Wei, Jiamin Yan, Xiang Xiang. (2026). "Why Feature Magnitude Deceives OOD Detectors: An Angular Separation Perspective." ECCV 2026.
Download Paper
Published in International Conference on Machine Learning (ICML), 2026
Proposes PTTA, a method that purifies malicious samples to enable robust test-time model adaptation.
Recommended citation: Jing Ma, Hanlin Li, Xiang Xiang. (2025). "PTTA: Purifying Malicious Samples for Test-Time Model Adaptation." ICML 2025.
Download Paper
Published in Advances in Neural Information Processing Systems (NeurIPS), 2026
Proposes decoupled entropy minimization for improved test-time adaptation and domain generalization.
Recommended citation: Jing Ma, Hanlin Li, Xiang Xiang. (2026). "Decoupled Entropy Minimization." NeurIPS 2026, 38: 161195–161230.
Download Paper
Published in IEEE Transactions on Image Processing (TIP), 2026
Introduces a frequency-aware approach for domain generalization by leveraging spectral domain information.
Recommended citation: Xiang Xiang, Jing Ma, Hanlin Li, et al. (2026). "Frequency-Aware Domain Generalization." IEEE TIP 2026.
Download Paper