报告题目:Imbalanced Learning with Noisy Label
报告时间:2023-12-20 14:00-16:00
报 告 人 :彭柳华 副教授 墨尔本大学
报告地点:概率统计系办公室
Abstract:Learning with imbalanced data
becomes critical and has received much attention from academia, industry, and
government funding agencies. The degree of imbalance among classes may be large
and sometimes become extreme. The classifier is naturally biased towards to
majority class which makes the minority group challenging to be identified. In
this project, we propose a novel method for handling imbalanced learning by
artificially introducing noisy labels to balance the classes. We provide both
theoretical and numerical justifications on our proposed method. Real data
applications show the outperformance of our method compared with existing
methods.