报告题目：Pairwise Learning Problems with Regularization Networks and Nystrom Subsampling Approach

报告时间：2024年7月1日  10:30-11:30

报告地点：综合楼644

报告人：胡婷

报告人简介：胡婷，西安交通大学管理学院教授、博士生导师，主要从事机器学习领域理论研究。目前，在Applied and Computational Harmonic Analysis，Journal of Machine Learning Research，IEEE Transactions on Signal Processing，Inverse Problems，Constructive Approximation, Neural Networks, Journal of Multivariate Analysis等杂志发表的系列论文，得到同行的广泛关注。

报告摘要：Pairwise learning usually refers to the learning problem that works with pairs of training samples, such as ranking, similarity and metric learning, and AUC maximization. To overcome the challenge of pairwise learning in the large scale computation, we introduces Nystrom sampling approach to the coefficient-based regularized pairwise algorithm in the context of kernel networks. Our theorems establish that the obtained Nystrom estimator achieves the minimax error over all estimators using the whole data provided that the subsampling level is not too small. We derive the function relation between the subsampling level and regularization parameter that guarantees computation cost reduction and asymptotic behaviors’ optimality simultaneously. The Nystrom coefficient-based pairwise learning method does not require the kernel to be symmetric or positive semi-definite, which provides more flexibility and adaptivity in the learning process. We apply the method to the bipartite ranking problem, which improves the state-of-the-art theoretical results in previous works.