狗熊会学术报告 | 朱宇教授：图像到图像回归的条件得分匹配

英文摘要

Image to image regression is an essential machine learning problem that is commonly encountered in computer vision tasks. Previous research works have been focused on task-dependent end-to-end regression models. Although such a discriminative approach has shown its proficiency in producing precise predictions, its architecture is usually overly complicated for interpretation and unable to provide uncertainty quantification. We propose a general image to image regression framework named CSMNet, based on conditional score matching and Langevin dynamics. The proposed framework is generative and emphasizes on the probabilistic perspective of general image to image regression. Discriminative models are also employed as sufficient feature extractors to complement the generative framework. The proposed method is trained by multi-level denoising score matching. Subsequently, the learned model is used for sampling from the conditional distribution via annealed Langevin dynamics. Comprehensive experiments are conducted on diverse image to image regression problems. The proposed method surpasses other generative models and achieves comparable performance to some end-to-end regression methods. The talk is based on a joint work with Hao Xin.

嘉宾介绍

朱宇博士是普渡大学统计系教授。他于2000年从密歇根大学获统计学博士学位。研究兴趣涉及工业统计、数据挖掘、降维、生物信息学和大数据分布式计算等。目前专注于数据科学的统计学习方法和理论。