Ruihan Wu (吴睿涵)

I am currently a postdoctoral researcher at the University of California, San Diego. I am fortunately advised by Kamalika Chaudhuri and also working closely with Yu-Xiang Wang.

I received my Ph.D. from Cornell University, where I was very fortunate to be advised by Kilian Q. Weinberger. Prior to that, I received my B.E. in computer science from Yao Class at Tsinghua University.

Email: ruw076 [at] ucsd [dot] edu or rw565 [at] cornell [dot] edu

[Google Scholar][LinkedIn][Github]

My research centers on the theory and applications of privacy-preserving machine learning, with a current emphasis on large language models (LLMs). I am particularly interested in:

Identifying data privacy and confidentiality risks
Designing algorithms that ensure privacy under formal guarantees
Defining practical objectives for real-world privacy protection

Beyond privacy, I am broadly interested in topics related to: machine learning safety, and online learning and multi-armed bandits, particularly their practical deployment and theoretical foundations.

Pre-prints

Beyond Per-Question Privacy: Multi-Query Differential Privacy for RAG Systems[pdf]

Ruihan Wu*, Erchi Wang*, Yu-Xiang Wang. Manuscripts

Learning-Time Encoding Shapes Unlearning in LLMs[pdf]

Ruihan Wu*, Konstantin Garov*, Kamalika Chaudhuri. Arxiv 2025.

Can We Infer Confidential Properties of Training Data from LLMs?[pdf]

Pengrun Huang, Chhavi Yadav, Ruihan Wu*, Kamalika Chaudhuri*. Arxiv 2025.

Do LLMs Really Forget? Evaluating Unlearning with Knowledge Correlation and Confidence Awareness.[pdf]

Rongzhe Wei, Peizhi Niu, Hans Hao-Hsun Hsu, Ruihan Wu, Haoteng Yin, Mohsen Ghassemi, Yifan Li, Vamsi K Potluru, Eli Chien, Kamalika Chaudhuri, Olgica Milenkovic, Pan Li. Arxiv 2025.

Improved Regret in Stochastic Decision-Theoretic Online Learning under Differential Privacy [pdf]

Ruihan Wu, Yu-Xiang Wang. Arxiv 2025.

Evaluating Deep Unlearning in Large Language Models [pdf][code]

Ruihan Wu, Chhavi Yadav, Ruslan Salakhutdinov, Kamalika Chaudhuri. ArXiv 2024.

Privacy-Preserving Retrieval Augmented Generation with Differential Privacy [pdf]

Tatsuki Koga, Ruihan Wu, Kamalika Chaudhuri. ArXiv 2024.

Influence-based Attributions can be Manipulated [pdf]

Chhavi Yadav*, Ruihan Wu*, Kamalika Chaudhuri. ArXiv 2024.

Conference Papers

On Speeding Up Language Model Evaluation [pdf]

Jin Peng Zhou*, Christian K. Belardi*, Ruihan Wu*, Travis Zhang, Carla P. Gomes, Wen Sun, Kilian Q. Weinberger. International Conference on Learning Representations (ICLR), 2025.

Large Scale Knowledge Washing [pdf]

Yu Wang, Ruihan Wu, Zexue He, Xiusi Chen, Julian McAuley. International Conference on Learning Representations (ICLR), 2025.

Online Feature Updates Improve Online (Generalized) Label Shift Adaptation [pdf][code]

Ruihan Wu*, Siddhartha Datta*, Yi Su, Dheeraj Baby, Yu-Xiang Wang, Kilian Q Weinberger. Advances of Neural Information Processing (NeurIPS), 2024.

Learning To Invert: Simple Adaptive Attacks for Gradient Inversion in Federated Learning [pdf ] [code]

Ruihan Wu*, Xiangyu Chen*, Chuan Guo, Kilian Q Weinberger. Conference on Uncertainty in Artificial Intelligence (UAI), 2023.

Does Label Differential Privacy Prevent Label Inference Attacks? [pdf ] [code]

Ruihan Wu*, Jin Peng Zhou*, Kilian Q Weinberger, Chuan Guo. International Conference on Artificial Intelligence and Statistics (AISTATS), 2023.

Differentially Private Multi-Party Data Release for Linear Regression [pdf ] [code][video]

Ruihan Wu, Xin Yang, Yuanshun Yao, Jiankai Sun, Tianyi Liu, Kilian Q Weinberger, Chong Wang. Conference on Uncertainty in Artificial Intelligence (UAI), 2022.

Online Adaptation to Label Distribution Shift [pdf ] [code][video]

Ruihan Wu, Chuan Guo, Yi Su, Kilian Q. Weinberger. Advances of Neural Information Processing (NeurIPS), 2021.

Fixes That Fail: Self-Defeating Improvements in Machine-Learning Systems [pdf ] [code]

Ruihan Wu, Chuan Guo, Awni Hannun, Laurens van der Maaten. Advances of Neural Information Processing (NeurIPS), 2021.

Making Paper Reviewing Robust to Bid Manipulation Attacks [pdf ] [code][video]

Ruihan Wu*, Chuan Guo*, Felix Wu, Rahul Kidambi, Laurens van der Maaten, Kilian Q. Weinberger. International Conference on Machine Learning (ICML), 2021.

Product Kernel Interpolation for Scalable Gaussian Processes [pdf]

Jake Gardner, Geoff Peiss, Ruihan Wu, Kilian Q. Weinberger, Andrew G. Wilson. International Conference on Artificial Intelligence and Statistics (AISTATS), 2018.

Quadratic Upper Bound for Recursive Teaching Dimension of Finite VC Classes [pdf]

Lunjia Hu, Ruihan Wu, Tianhong Li, Liwei Wang. Conference on Learning Theory (COLT), 2017.

Journal Papers

Correlator Convolutional Neural Networks as an Interpretable Architecture for Image-like Quantum Matter Data [pdf ]

Cole Miles, Annabelle Bohrdt, Ruihan Wu, Christie Chiu, Muqing Xu, Geoffrey Ji, Markus Greiner, Kilian Q. Weinberger, Eugene Demler, Eun-Ah Kim. Nature Communications, 2021.

Scalable Lattice Influence Maximization [pdf]

Wei Chen*, Ruihan Wu*, Zheng Yu*. IEEE Transactions on Computational Social Systems (TCSS), 2020.

Manuscripts

Better Membership Inference Privacy Measurement through Discrepancy [pdf]

Ruihan Wu*, Pengrun Huang*, Kamalika Chaudhuri. ArXiv 2024.

Large-Scale Public Data Improves Differentially Private Image Generation Quality [pdf ]

Ruihan Wu, Chuan Guo, Kamalika Chaudhuri. ArXiv 2023.

On Hiding Neural Networks Inside Neural Networks [pdf] [code]

Chuan Guo*, Ruihan Wu*, Kilian Q. Weinberger. ArXiv, 2020.

Talks

Google Algorithms Seminar, March 2025.

Joint Statistical Meetings, August 2024.

Math Machine Learning seminar at MPI MIS and UCLA, July 2024.

Research Seminar at Hong Kong Polytechnic University, July 2024.

Women in Data Science, February 2024.

Google Differential Privacy for ML Seminar, September 2022. [link]

TrustML Young Scientist Seminar, October 2022. [link]

Selected Awards

LinkedIn PhD Award, 2022

National Scholarship (top 2% in Tsinghua University), 2017

Gold Medal in Chinese Mathematical Olympiad , 2013

Page updated

Google Sites

Report abuse