数学科学研究所
Insitute of Mathematical Science

Seminar: Wasserstein Generative Regression

Seminar| Institute of Mathematical Sciences

Time: Saturday, December 27th, 2025, 11:00-11:30AM

Location: RS408, IMS

Speaker: Shanshan Song, Tongji University


Abstract: We propose a new and unified approach for nonparametric regression and conditional distribution learning. Our approach simultaneously estimates a regression function and a conditional generator using a generative learning framework, where a conditional generator is a function that can generate samples from a conditional distribution. The main idea is to estimate a conditional generator satisfying the constraint that it produces a good regression function estimator. We use deep neural networks to model the conditional generator. Our approach can handle problems with multivariate outcomes and covariates, and can be used to construct prediction intervals. We provide theoretical guarantees by deriving non-asymptotic error bounds and the distributional consistency of our approach under suitable assumptions. We perform numerical experiments to demonstrate the effectiveness and superiority of our approach over some existing approaches in various scenarios.



地址:上海市浦东新区华夏中路393号
邮编:201210
上海市徐汇区岳阳路319号8号楼
200031(岳阳路校区)