Expand globally, shrink locally: Discriminant multi-label learning with missing labels

作者：

Highlights：

• We develop a simple yet effective discriminant multi-label learning (DM2L) method for multi-label learning with missing labels.

• We impose both local and global rank structures to model label structures, and also to provide more label discriminability.

• We apply the kernel trick to provide a nonlinear extension to enhance nonlinear ability of our model.

• Although our method involves the fewest assumptions and only one hyper-parameter, it still outperforms the state-of-the-art methods.

摘要

•We develop a simple yet effective discriminant multi-label learning (DM2L) method for multi-label learning with missing labels.•We impose both local and global rank structures to model label structures, and also to provide more label discriminability.•We apply the kernel trick to provide a nonlinear extension to enhance nonlinear ability of our model.•Although our method involves the fewest assumptions and only one hyper-parameter, it still outperforms the state-of-the-art methods.

论文关键词：Multi-label learning,Missing labels,Local low-rank label structure,Global low-rank label structure,Label discriminant information

论文评审过程：Received 13 April 2020, Revised 10 August 2020, Accepted 22 September 2020, Available online 25 September 2020, Version of Record 1 October 2020.

论文官网地址：https://doi.org/10.1016/j.patcog.2020.107675