A Minimax Probability Machine for Non-Decomposable Performance Measures
报告人：张波研究员
中科院数学与系统科学研究院
报告时间：2022年03月25日（星期五）下午3点
报告地点：腾讯会议560-227-159
报告摘要：Imbalanced classification tasks are widespread in many real-world applications. For such classification tasks, in comparison with the accuracy rate, it is usually much more appropriate to use non-decomposable performance measures such as the Area Under the receiver operating characteristic Curve (AUC) and the F_\beta measure as the classification criterion since the label class is imbalanced. On the other hand, the minimax probability machine is a popular method for binary classification problems and aims at learning a linear classifier by maximizing the accuracy rate, which makes it unsuitable to deal with imbalanced classification tasks.
In this talk, we develop a new minimax probability machine for the F_\beta measure, called MPMF, which can be used to deal with imbalanced classification tasks. A brief discussion is also given on how to extend the MPMF model for several other non-decomposable performance measures listed in the talk. To solve the MPMF model effectively, we derive its equivalent form which can then be solved by an alternating descent method to learn a linear classifier. Further, the kernel trick is employed to derive a nonlinear MPMF model to learn a nonlinear classifier. Several experiments on real-world benchmark datasets demonstrate the effectiveness of our new model. This talk is based on recent joint work with Junru Luo and Hong Qiao.
J Luo, H Qiao & B Zhang, A Minimax Probability Machine for Non-Decomposable Performance Measures, IEEE Transactions on Neural Networks and Learning Systems, Published online, DOI:10.1109/TNNLS.2021.3106484, 2021.
报告人简介：张波，1983年毕业于山东大学数学系，1985年在西安交通大学获硕士学位，1992年在英国Strathclyde大学获博士学位。现任中科院数学与系统科学研究院“百人计划”研究员，应用数学研究所副所长，反问题国际联合会东亚分会副主席，中国工业与应用数学学会秘书长、大数据与人工智能专委会副主任，中国数学会常务理事。1992-1994在英国Keele大学做博士后，1995-1997任英国Brunel大学Research Fellow，1997-2007年历任英国Coventry大学高级讲师、Reader、应用数学教授，2004年通过中科院“百人计划”回国，在波传播与散射、反问题与成像、机器学习与智能数据分析的理论和算法方面进行了系统深入研究，在国际重要学术期刊发表论文130余篇。曾任2019年反问题国际联合会Calderon奖委员会成员，国际著名SCI期刊《IEEE Transactions on Cybernetics》编委（2012-2018），应邀在2012年第6届和2018年第9届反问题国际会议做1小时大会特邀报告，三次获得中科院优秀导师奖（2013, 2019, 2020），获2021年度中国科学院大学领雁奖。