7 papers
Context-Constrained Transfer Learning for Tabular Foundation Models via Data Distillation
Yijun Lin, Sai Li
Tabular Foundation Models (TFMs) have demonstrated strong empirical performance as black-box inference engines through in-context learning. However, their use in transfer learning…
Routing-Aware Expert Calibration for Machine Unlearning in Mixture-of-Experts Language Models
Jingyi Xie, Yijun Lin, Yinjiang Xiong +2
Machine unlearning is increasingly important for large language models, yet unlearning in Mixture-of-Experts (MoE) architectures remains underexplored. Unlike dense models, MoE arc…
Efficient machine unlearning with minimax optimality
Jingyi Xie, Linjun Zhang, Sai Li
There is a growing demand for efficient data removal to comply with regulations like the GDPR and to mitigate the influence of biased or corrupted data. This has motivated the fiel…
Personalizing black-box models for nonparametric regression with minimax optimality
Sai Li, Linjun Zhang
Recent advances in large-scale models, including deep neural networks and large language models, have substantially improved performance across a wide range of learning tasks. The…
FAIRM: Learning invariant representations for algorithmic fairness and domain generalization with minimax optimality
Sai Li, Linjun Zhang
Machine learning methods often assume that the test data have the same distribution as the training data. However, this assumption may not hold due to multiple levels of heterogene…
Differentially Private Estimation and Inference in High-Dimensional Regression with FDR Control
Zhanrui Cai, Sai Li, Xintao Xia +1
This paper proposes new methodologies for conducting practical differentially private (DP) estimation and inference in high-dimensional linear regression. We first introduce a DP B…
Multi-dimensional domain generalization with low-rank structures
Sai Li, Linjun Zhang
In conventional statistical and machine learning methods, it is typically assumed that the test data are identically distributed with the training data. However, this assumption do…