Terrance Liu

I am a PhD student in the Machine Learning Department at Carnegie Mellon University, where I am advised by Steven Wu. My research centers around privacy-preserving machine learning and uncertainty quantification. Previously, while completing my master's degree (also from CMU), I worked on multimodal machine learning and federated learning under Louis-Philippe Morency. Before that, I obtained my bachelor's degree in math and economics from the University of Chicago.

My work has been supported by the Bloomberg Data Science Ph.D. Fellowship (link).

If you'd like to chat, please feel free to reach out at terrancl(at)andrew(dot)cmu(dot)edu.

Publications
* denotes equal contribution

Breaking the Gold Standard: Extracting Forgotten Data under Exact Unlearning in Large Language Models [arXiv]
Xiaoyu Wu, Yifei Pang, Terrance Liu, Zhiwei Steven Wu

Generate-then-Verify: Reconstructing Data from Limited Published Statistics [arXiv]
Terrance Liu*, Eileen Xiao*, Adam Smith, Pratiksha Thaker, Zhiwei Steven Wu
TPDP 2025 (oral)

Winning the MIDST Challenge: New Membership Inference Attacks on Diffusion Models for Tabular Data Synthesis [arXiv]
Xiaoyu Wu, Yifei Pang, Terrance Liu, Zhiwei Steven Wu
Winner (of all 4 tracks) of the MIDST Challenge @ SatML 2025 (article)
TPDP 2025

Enhancing One-run Privacy Auditing with Quantile Regression-Based Membership Inference [arXiv]
Terrance Liu, Matteo Boglioni, Yiwei Fu, Shengyuan Hu, Pratiksha Thaker, Zhiwei Steven Wu
TPDP 2025

Generating Private Synthetic Data with Genetic Algorithms [arXiv][code]
Terrance Liu*, Jingwu Tang*, Giuseppe Vietri*, Zhiwei Steven Wu (alphabetical order)
ICML 2023

Confidence-Ranked Reconstruction of Census Microdata from Published Statistics [link][arXiv][code]
Travis Dick, Cynthia Dwork, Michael Kearns, Terrance Liu, Aaron Roth, Giuseppe Vietri, Zhiwei Steven Wu (alphabetical order)
Proceedings of the National Academy of Sciences (PNAS)
TPDP 2023 (oral)

Policy Impacts of Statistical Uncertainty and Privacy [link][code]
Ryan Steed, Terrance Liu, Zhiwei Steven Wu, Alessandro Acquisti
Science

Private Synthetic Data with Hierarchical Structure [arXiv]
Terrance Liu, Zhiwei Steven Wu
TPDP 2022

Iterative Methods for Private Synthetic Data: Unifying Framework and New Methods [arXiv][code]
Terrance Liu*, Giuseppi Vietri*, Zhiwei Steven Wu
NeurIPS 2021

Leveraging Public Data for Practical Private Query Release [arXiv][code]
Terrance Liu, Giuseppi Vietri, Thomas Steinke, Jonathan Ullman, Zhiwei Steven Wu
ICML 2021

Learning Language and Multimodal Privacy-Preserving Markers of Mood from Mobile Data [arXiv]
Paul Pu Liang*, Terrance Liu*, Anna Cai, Michal Muszynski, Ryo Ishii, Nicholas Allen, Randy Auerbach, David Brent, Ruslan Salakhutdinov, Louis-Philippe Morency
ACL 2021 (oral)

Think Locally, Act Globally: Federated Learning with Local and Global Representations [arXiv][code]
Paul Pu Liang*, Terrance Liu*, Liu Ziyin, Nicholas Allen, Randy Auerbach, David Brent, Ruslan Salakhutdinov, Louis-Philippe Morency
NeurIPS'19 Workshop on Federated Learning
(oral, distinguished student paper award)