Publicationsedit
Peer-reviewed and accepted papersedit
| Publication | Venue |
|---|---|
| Illusory Pattern Perception Drives Spurious Inference in Large Language Models. Peihua Mai*, Zhuoyan Shao*, Xinbao Qiao*, Meng Zhang, Xinyue Zhou, Yan Pang. Compares LLM and human illusory pattern perception across three psychological tasks, with SAE analysis of associated internal features. OpenReview · arXiv · Code | NeurIPS 2026, 6-12 December 2026, Sydney, Australia (main site). |
| When Sample Selection Bias Precipitates Model Collapse. Xinbao Qiao, Xianglong Du, Wei Liu, Jingqi Zhang, Peihua Mai, Meng Zhang, Yan Pang. Low-resource verification regimes can turn local sample selection into persistent tail pruning. OpenReview · arXiv · Code | ICML 2026, 6-11 July 2026, Seoul, South Korea. |
| Beyond Binary Erasure: Soft-Weighted Unlearning for Fairness and Robustness. Xinbao Qiao, Ningning Ding, Yushi Cheng, Meng Zhang. Soft-weighted corrective unlearning for non-binary fairness and robustness interventions. OpenReview · arXiv · Publisher | AAAI 2026, 20-27 January 2026, Singapore. |
| Hessian-Free Online Certified Unlearning. Xinbao Qiao, Meng Zhang, Ming Tang, Ermin Wei. Certified unlearning without explicit Hessian inversion; efficient removal through recollected statistics. OpenReview · arXiv · Code | ICLR 2025, 24-28 April 2025, Singapore. |
| DynFrs: An Efficient Framework for Machine Unlearning in Random Forest. Shurong Wang, Zhuoyang Shen, Xinbao Qiao, Tongning Zhang, Meng Zhang. Exact and efficient random-forest unlearning in dynamic environments. OpenReview · arXiv · Code | ICLR 2025, 24-28 April 2025, Singapore. |
Manuscripts under reviewedit
- Decentralized Free-Support Wasserstein Barycenter.
Xinbao Qiao, Bokai Hou, Peihua Mai, Wenqian Li, Wenjing Yan, Ying-Jun Angela Zhang.
Under review.
Movable supports build geometry-aware shared barycenters under network communication constraints.