Conference & Journal Publications

* indicates equal contributions

2025

1. UFID: A Unified Framework for Input-level Backdoor Detection on Diffusion Models [PDF] [Code]
   Zihan Guan*, Mengxuan Hu*, Sheng Li, Anil Vullikanti
   AAAI Conference on Artificial Intelligence (AAAI 2025)

2. Evaluating accuracy and reproducibility of large language model performance on critical care assessments in pharmacy education
   Huibo Yang , Mengxuan Hu , Amoreena Most , Brian Murray , Susan Smith , Anthony Hawkins , Sheng Li and Andrea Sikora
   Frontiers in Artificial Intelligence, section Medicine and Public Health

3. Mind Control through Causal Inference: Predicting Clean Images from Poisoned Data
   Mengxuan Hu*, Zihan Guan*, Yi Zeng, Junfeng Guo, Zhongliang Zhou, Jielu Zhang, Ruoxi Jia, Anil Kumar Vullikanti, Sheng Li
   International Conference on Learning Representations (ICLR 2025)

4. Revisiting Source-Free Domain Adaptation: Insights into Representativeness, Generalization, and Diversity.
   Ronghang Zhu, Mengxuan Hu, Weiming Zhuang, Lingjuan Lyu, Xiang Yu, and Sheng Li.
   IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2025)

5. Benign Samples Matter! Fine-tuning On Outlier Benign Samples Severely Breaks Safety [PDF] [Code]
   Zihan Guan*, Mengxuan Hu*, Ronghang Zhu, Sheng Li, Anil Vullikanti
   International Conference on Machine Learning (ICML 2025 spotlight 🏆 top 2.6%) ⭐

6. BalancEdit: Dynamically Balancing the Generality-Locality Trade-off in Multi-modal Model Editing [PDF]
   Dongliang Guo, Mengxuan Hu, Zihan Guan, Thomas Hartvigsen, Sheng Li
   International Conference on Machine Learning (ICML 2025)

7. Distort Time to Improve Video Temporal Reasoning.
   Daiqing Qi, Dongliang Guo, Hanzhang Yuan, Handong Zhao, Mengxuan Hu, Lehan Yang, and Sheng Li
   The Thirty-ninth Annual Conference on Neural Information Processing Systems (NeurIPS 2025)

8. No free lunch: Retrieval-augmented generation undermines fairness in LLMs, even for vigilant users [PDF]
   Mengxuan Hu*, Hongyi Wu*, Zihan Guan, Ronghang Zhu, Dongliang Guo, Daiqing Qi, and Sheng Li
   Findings of The Conference on Empirical Methods in Natural Language Processing (EMNLP Findings 2025)

9. Backdoor in Seconds: Unlocking Vulnerabilities in Large Pre-trained Models via Model Editing [PDF]
   Dongliang Guo, Mengxuan Hu, Zihan Guan, Junfeng Guo, Thomas Hartvigsen, and Sheng Li
   The 34th ACM International Conference on Information and Knowledge Management (ACM CIKM 2025)

2024

1. BBCaL: Black-box Backdoor Detection under the Causality Lens [PDF] [Code]
   Mengxuan Hu*, Zihan Guan*, Junfeng Guo, Zhongliang Zhou, Jielu Zhang, Sheng Li
   Transactions on Machine Learning (TMLR) (🏆 Featured Certificate)

2. DTRNet: Precisely correcting selection bias in individual-level continuous treatment estimation through reweighted disentangled representation.[PDF] [Code]
   Mengxuan Hu, Zhixuan Chu, Sheng Li.
   Transactions on Machine Learning (TMLR)

3. Img2Loc: Revisiting Image Geolocalization using Multi-modality Foundation Models and Image-based Retrieval-Augmented Generation [PDF] [Code]
   Zhongliang Zhou, Jielu Zhang, Zihan Guan, Mengxuan Hu, Ni Lao, Lan Mu, Sheng Li, Gengchen Mai
   ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR 2024)

4. Text2seg: Remote sensing image semantic segmentation via text-guided visual foundation models [PDF] [Code]
   Jielu Zhang, Zhongliang Zhou, Gengchen Mai, Mengxuan Hu, Zihan Guan, Sheng Li, Lan Mu
   ACM SIGSPATIAL International Workshop on AI for Geographic Knowledge Discovery (GeoAI 2024)

5. Causal inference with latent variables: Recent advances and future prospectives.[PDF]
   Yaochen Zhu, Yinhan He, Jing Ma, Mengxuan Hu, Sheng Li, and Jundong Li.
   Special Interest Group on Knowledge Discovery and Data Mining (SIGKDD 2024) (Tutorial Track & Survey Paper)

6. BadSAM: Exploring Security Vulnerabilities of SAM via Backdoor Attacks.[PDF] [Code]
   Zihan Guan*, Mengxuan Hu*, Zhongliang Zhou*, Jielu Zhang*, Sheng Li, and Ninghao Liu
   AAAI Conference on Artificial Intelligence (AAAI 2024) (Student Abstract)

7. Task-driven causal feature distillation: Towards trustworthy risk prediction.[PDF]
   Zhixuan Chu, Mengxuan Hu, Qing Cui, Longfei Li, and Sheng Li
   AAAI Conference on Artificial Intelligence (AAAI 2024)

2023

1. Graph-Based Counterfactual Causal Inference Modeling for Neuroimaging Analysis[PDF]
Haixing Dai, Mengxuan Hu, Qing Li*, Lu Zhang, Lin Zhao, Dajiang Zhu, Ibai Diez, Jorge Sepulcre, Fan Zhang, Xingyu Gao, Manhua Liu, Quanzheng Li, Sheng Li, Tianming Liu, Xiang Li
Workshop in International Conference on Medical Image Computing and Computer-Assisted Intervention(MICCAI 2023)