Publications

Compute-efficient LLM Training via Online Batch Selection

NeurIPS'24 (Spotlight)

Rethinking Data Shapley for Data Selection Tasks: Misleads and Merits

ICML'24 (Oral Presentation)
arXiv

Efficient Data Valuation for Weighted Nearest Neighbor Algorithms

AISTATS'24 (Oral Presentation)
arXiv

Privacy-Preserving In-Context Learning for Large Language Models

ICLR'24
arXiv

DP-OPT: Make Large Language Model Your Privacy-Preserving Prompt Engineer

ICLR'24 (Spotlight)
arXiv

Threshold KNN-Shapley: A Linear-Time and Privacy-Friendly Approach to Data Valuation

NeurIPS'23 (Spotlight)
arXiv

A Randomized Approach for Tight Privacy Accounting

NeurIPS'23
arXiv

Data Banzhaf: A Robust Data Valuation Framework for Machine Learning

AISTATS'23 (Oral Presentation)
arXiv Code

LAVA: Data Valuation without Pre-Specified Learning Algorithms

ICLR'23 (Spotlight)
OpenReview

ModelPred: A Framework for Predicting Trained Model from Training Data

SaTML'23
Arxiv OpenReview

Renyi Differential Privacy of Propose-Test-Release and Applications to Private and Robust Machine Learning

NeurIPS'22
ArXiv

Improving Cooperative Game Theory-based Data Valuation via Data Utility Learning

ICLR 2022 Workshop on Socially Responsible Machine Learning
Arxiv Workshop Version Poster

Concurrent Composition of Differential Privacy

TCC'21, TPDP'21
TCC Version Workshop Version Poster Slides

DPlis: Boosting Utility of Differentially Private Deep Learning via Randomized Smoothing

PETS'21 (acceptance rate: 19%)
Arxiv

RIGA: Covert and Robust White-Box Watermarking of Deep Neural Networks

WWW'21 (acceptance rate: 20.6%)
Paper arXiv Code Talk

Improving Robustness to Model Inversion Attacks via Mutual Information Regularization

AAAI'21 (acceptance rate: 21%)
Paper arXiv Slides Poster Code

A Principled Approach to Data Valuation for Federated Learning

Federated Learning: Privacy and Incentives (2020)
Paper arXiv