​​
Foundation Models and LLMs
[Mechanism Design for LLMs, Uncertainty Quantification with ML Models]
(Note: students under my supervision; † equal contribution; * corresponding author)​​​​​
​
Mechanism Design for LLMs​​
(Overview)​
Baiting Chen, Tong Zhu, Xuanang Li, Yichi Zhang, and Xiaowu Dai*.
Preprint, 2026.​
(Hallucination)​
-
Incentivizing truthful language models via peer elicitation games. [proceedings​][preprint][code]
Baiting Chen†, Tong Zhu†, Jiale Han, Lexin Li, Gang Li, and Xiaowu Dai*.
Advances in Neural Information Processing Systems (NeurIPS), 2025.​
​​
(Reasoning)​​
-
ALIGN: Aligned delegation with performance guarantees for multi-agent LLM reasoning. [pdf​][code​]
Tong Zhu†, Baiting Chen†, Jin Zhou, Hua Zhou, Sriram Sankararaman, and Xiaowu Dai*.
Preprint, 2026.
​​​
(Alignment)
Baiting Chen†, Tong Zhu†, Rui Yu†, and Xiaowu Dai*.
Preprint, 2026.
​
​
Uncertainty Quantification with ML Models​​​
(Prediction-Powered Inference)
-
Surrogate-powered inference: Regularization and adaptivity. [pdf​]
Jianmin Chen, Huiyuan Wang, Thomas Lumley, Xiaowu Dai, and Yong Chen.
Under Moderate Revisions at Journal of the American Statistical Association (JASA), 2025.
​​
Yang Sui, Jin Zhou, Hua Zhou, and Xiaowu Dai*.
Preprint, 2026.
​​​
(LLM Evaluation)
-
Prompt-dependent ranking of large language models with uncertainty quantification. [pdf​][code​]
Angel Rodrigo Avelar Menendez, Yufeng Liu, and Xiaowu Dai*.
Preprint, 2026.
​​
-
Sweeping failures out of the pooled rug: Simultaneous PPI for weakest-stratum confidence in routed LLM judges. [pdf​]
Jiahang Sha, Xiaowu Dai, and Wei Wang.
Preprint, 2026.
​
​
​
​
​
​