Lijie Hu

Email: lijie [dot] hu [at] kaust [dot] edu [dot] sa

Hello! I am Lijie Hu, a fourth-year Ph.D. student in Computer Science at King Abdullah University of Science and Technology (KAUST) since Spring 2021, and I am very fortunate to be advised by Prof. Di Wang in PRADA Lab (Provable Responsible AI and Data Analytics Lab). Before that, I received my Master's degree in Mathematics from Renmin University of China.

My research interests are Explainable AI and Privacy-preserving AI. Specifically, my research goal is to build faithful XAI systems that are easily understood by users and are robust in various environments, especially for large models (e.g., large language models and large multimodal). I am also interested in applying the XAI to real-world scenarios (e.g., healthcare, recommender systems, traffic forecasting, materials, etc.).

Pinned
I am looking for several interns. Please do not hesitate to contact me if you are interested.

News

07/2024: Our paper "SATO: Stable Text-to-Motion Framework" has been accepted at The 32nd ACM Multimedia Conference (ACM MM 2024)!
07/2024: Two papers have been accepted at The 1st Conference on Language Modeling (COLM 2024)!
05/2024: I accepted the invitation to serve as a reviewer for NeurIPS 2024.
05/2024: Our paper "Improving Interpretation Faithfulness for Vision Transformers" has been accepted at The 41st International Conference on Machine Learning (ICML 2024)!
04/2024: Our paper "Faster Rates of Differentially Private Stochastic Convex Optimization" has been accepted at Journal of Machine Learning Research (JMLR)!
03/2024: I am honored to receive the ICLR 2024 Travel Grant.
02/2024: I am honored to have been elected to the AAAI Student Committee!
02/2024: I accepted the invitation to serve as a reviewer for ICML 2024.

01/2024: Our paper "Differentially Private Natural Language Models: Recent Advances and Future Directions" has been accepted at The 18th Conference of the European Chapter of the Association for Computational Linguistics (EACL 2024)!
01/2024: Our paper "Faithful Vision-Language Interpretation via Concept Bottleneck Models" has been accepted at The 12th International Conference on Learning Representations (ICLR 2024)!
11/2023: I accepted the invitation to serve as a reviewer for IEEE Transactions on Neural Networks and Learning Systems (TNNLS) and ACM Transactions on Privacy and Security (TOPS).
10/2023: I accepted the invitation to serve as a reviewer for AISTATS 2024.
10/2023: Our paper "Nearly Optimal Rates of Privacy-preserving Sparse Generalized Eigenvalue Problem" has been accepted at IEEE Transactions on Knowledge and Data Engineering (TKDE)!
07/2023: Our paper "Finite Sample Guarantees of Differentially Private Expectation Maximization Algorithm" has been accepted at The 26th European Conference on Artificial Intelligence (ECAI 2023)!
05/2023: Our paper "Generalized Linear Models in Non-interactive Local Differential Privacy with Public Data" has been accepted by the Journal of Machine Learning Research (JMLR)!
05/2023: Our proposal "Towards Faithful Transformers and Attention Mechanisms," Co-PIs with Prof. Di Wang, has been granted by SDAIA-KAUST Center of Excellence in Data Science and AI (SDAIA-KAUST) $53,326 USD. Thanks to SDAIA-KAUST!

Selected Publications ("*" equal contribution, "†" corresponding author, "__" advised student)
Explainable Artificial Intelligence (XAI)

[ICML] Improving Interpretation Faithfulness for Vision Transformers. [Link] [ArXiv] [Code] [Video]
Lijie Hu*, Yixin Liu*, Ninghao Liu, Mengdi Huai, Lichao Sun, and Di Wang.
The 41st International Conference on Machine Learning (ICML 2024).
Selected as a Spotlight paper (3.5% acceptance rate).
[ICLR] Faithful Vision-Language Interpretation via Concept Bottleneck Models. [Link] [Code] [Video]
Songning Lai*, Lijie Hu*†, Junxiao Wang, Laure Berti-Equille, and Di Wang.
The 12th International Conference on Learning Representations (ICLR 2024).
[AAAI] SEAT: Stable and Explainable Attention. [Link] [ArXiv] [Code] [Video]
Lijie Hu*, Yixin Liu*, Ninghao Liu, Mengdi Huai, Lichao Sun, and Di Wang.
The 37th AAAI Conference on Artificial Intelligence (AAAI 2023).
Selected as an Oral paper.
[TKDE] Towards Stable and Explainable Attention Mechanisms.
Lijie Hu*, Xinhai Wang*, Yixin Liu*, Ninghao Liu, Mengdi Huai, Lichao Sun, and Di Wang.
Major Revision, IEEE Transactions on Knowledge and Data Engineering (TKDE).

Large Language Models / Large Multimodals (LLM/MLLM)

[COLM] Multi-hop Question Answering under Temporal Knowledge Editing. [ArXiv] [Code]
Keyuan Cheng*, Gang Lin*, Haoyang Fei*, Yuxuan Zhai, Lu Yu, Muhammad Asif Ali, Lijie Hu†, and Di Wang.
The 1st Conference on Language Modeling (COLM 2024).
[COLM] MONAL: Model Autophagy Analysis for Modeling Human-AI Interactions. [ArXiv] [Code]
Shu Yang*, Muhammad Asif Ali*, Lu Yu, Lijie Hu†, and Di Wang.
The 1st Conference on Language Modeling (COLM 2024).
[ACM MM] SATO: Stable Text-to-Motion Framework. [ArXiv] [Code]
Wenshuo Chen, Hongru Xiao, Erhang Zhang, Lijie Hu, Lei Wang, Mengyuan Liu, and Chen Chen.
The 32nd ACM Multimedia Conference (ACM MM 2024).

Privacy-preserving Artificial Intelligence

[EACL] Differentially Private Natural Language Models: Recent Advances and Future Directions. [Link] [ArXiv] [Video]
Lijie Hu, Ivan Habernal, Lei Shen and Di Wang.
Findings of the 2024 European Chapter of the Association for Computational Linguistics (EACL 2024 Findings).
[JMLR] Generalized Linear Models in Non-interactive Local Differential Privacy with Public Data. [Link]
Di Wang*, Lijie Hu*, Huanyu Zhang, Marco Gaboardi, and Jinhui Xu.
Journal of Machine Learning Research, Volume 24, 132 (2023), Pages 1-57 (JMLR).
[ECAI] Finite Sample Guarantees of Differentially Private Expectation Maximization Algorithm. [Link]
Di Wang*, Jiahao Ding*, Lijie Hu, Zejun Xie, Miao Pan, and Jinhui Xu.
The 26th European Conference on Artificial Intelligence (ECAI 2023).
[AISTATS] Privacy-preserving Sparse Generalized Eigenvalue Problem. [Link]
Lijie Hu*, Zihang Xiang*, Jiabin Liu, and Di Wang.
The 26th International Conference on Artificial Intelligence and Statistics (AISTATS 2023).
[TKDE] Nearly Optimal Rates of Privacy-preserving Sparse Generalized Eigenvalue Problem. [Link]
Lijie Hu*, Zihang Xiang*, Jiabin Liu, and Di Wang.
IEEE Transactions on Knowledge and Data Engineering, 2023 (01): 1-14 (TKDE).
[PODS] High Dimensional Differentially Private Stochastic Optimization with Heavy-tailed Data. [Link] [ArXiv] [Video]
Lijie Hu, Shuo Ni, Hanshen Xiao, and Di Wang.
Proceedings of the 41st ACM SIGMOD-SIGACT-SIGAI Symposium on Principles of Database Systems (PODS 2022).
Invited to The ACM Transactions on Database Systems special issue on Best of PODS 2022.
CCS Workshop on Privacy Preserving Machine Learning 2021.
[ALT] Faster Rates of Differentially Private Stochastic Convex Optimization. [Link]
Jinyan Su, Lijie Hu, and Di Wang.
Proceedings of The 33rd International Conference on Algorithmic Learning Theory (ALT 2022).
[JMLR] Faster Rates of Private Stochastic Convex Optimization. [Link]
Jinyan Su, Lijie Hu, and Di Wang.
Journal of Machine Learning Research, Volume 25, 114 (2024), Pages 1−41 (JMLR).

Selected Preprints ("*" equal contribution, "†" corresponding author, "__" advised student)

Editable Concept Bottleneck Models. [ArXiv] [Code]
Lijie Hu*, Chenyang Ren*, Zhengyu Hu, Cheng-Long Wang, and Di Wang.
A Hopfieldian View-based Interpretation for Chain-of-Thought Reasoning. [ArXiv] [Code]
Lijie Hu, Liang Liu, Shu Yang, Xin CHEN, Hongru Xiao, Mengdi Li, Pan Zhou, Muhammad Asif Ali, and Di Wang.
Semi-supervised Concept Bottleneck Models. [ArXiv] [Code]
Lijie Hu, Tianhao Huang, Huanyi Xie, Chenyang Ren, Zhengyu Hu, Lu Yu, and Di Wang.
Leveraging Logical Rules in Knowledge Editing: A Cherry on the Top. [ArXiv] [Code]
Keyuan Cheng*, Muhammad Asif Ali*, Shu Yang*, Gang Ling, Yuxuan Zhai, Haoyang Fei, Ke Xu, Lu Yu, Lijie Hu†, and Di Wang.
Dialectical Alignment: Resolving the Tension of 3H and Security Threats of LLMs. [ArXiv] [Code]
Shu Yang, Jiayuan Su, Han Jiang, Mengdi Li, Keyuan Cheng, Muhammad Asif Ali, Lijie Hu, and Di Wang.
Fair Text-to-Image Diffusion via Fair Mapping. [ArXiv] [Code]
Jia Li*, Lijie Hu*, Jingfeng Zhang, Tianhang Zheng, Hua Zhang, and Di Wang.

Services

AAAI Student Committee
Program Committee Member/Reviewer
2024: NeurIPS, ICML, AISTATS
2023: ACL, AISTATS, EuroS&P, EMNLP, ECAI
2022: EMNLP
External/Sub Reviewer
2024: WACV, KDD
2023: CVPR, FAccT, ICCV
2022: WACV, ICLR, CVPR, ECCV, NeurIPS, ESORICS
2021: CVPR, ICCV, NeurIPS
Volunteer
2023: AISTATS
Journal Reviewer
ACM Transactions on Database Systems
ACM Transactions on Privacy and Security
IEEE Transactions on Neural Networks and Learning Systems

Experience

Research Intern, Ant Group (Jun. 2023 - Oct. 2023, China)
Mentor: Dr. Lu Yu
Visiting Student, Technical University Darmstadt (Jun. 2022 - Aug. 2022, Germany)
Advisor: Dr. Ivan Habernal
Remote Visiting Student, Lehigh University (Feb. 2022 - Dec. 2022)
Advisor: Prof. Lichao Sun
Research Intern, NEC Laboratories China (Jul. 2020 - Jan. 2021, China)
Mentor: Dr. Yu Wu and Dr. Wenjuan Wei
Visiting Student, University of Salerno (Mar. 2019 - Jun. 2019, Italy)
Advisor: Prof. Giuseppe Fenza and Prof. Carmen De Maio

Selected Awards

ICLR 2024 Travel Award.
AISTATS 2023 Top Reviewer.
AAAI 2023 Travel Award.
CEMSE Dean's List Award, KAUST, 2022, 2024.
Invited to The ACM Transactions on Database Systems special issue on Best of PODS 2022.
Beijing Honored Graduates (Top 5%), 2017.
Merit Student of Beijing (Top 1%), 2016.
China National Scholarship (Top 2‰), 2014, 2015.

Fundings

2024-2025, "Towards Faithful Transformers and Attention Mechanisms"; Co-PIs: Prof. Di Wang, Lijie Hu; SDAIA-KAUST Center of Excellence in Data Science and AI (SDAIA-KAUST), $53,326 USD.

Teaching

Teaching Assistant
- CS 229: Machine Learning, Spring 2022, Spring 2023, Spring 2024 @KAUST.
- Calculus, Fall 2017, Spring 2018 @RUC.

Talks

2023.4: "Towards Faithful Explainable AI", Ant Group, Alibaba.
2023.2: "Stable and Explainable Attention", AAAI 2023.