Abstract:Multi-label metric learning serve as an effective strategy to facilitate multi-label classification, aiming to learn better similarity metrics from multi-label examples. Existing multi-label metric learning approaches learn consistent metrics across all multi-label instances in the label space. However, such consistent metric learning approaches are suboptimal as they neglect the nonlinear distribution characteristics of multi-label instances. In this paper, we present LSMM, a label-specific multi-metric learning framework for multi-label classification, where nonlinear distribution characteristics of multi-label examples are considered by learning label-specific multiple local metrics for different instances on the shoulder of a global one. Specifically, multi-label instances within each label space can be divided into several disjoint partitions through either semantic-based or cluster-based partition strategies, in each of which a local metric is trained to separate the instances locally. Besides, a global metric is introduced to implicitly exploit high-order label correlations across all labels. The combination of the global metric and label-specific local metrics is utilized to measure the semantic similarities between multi-label instances in each label space, under which similar intra-class instances are pushed closer and inter-class instances are pulled apart. Comprehensive experiments on benchmark multi-label data sets validate the superiority of LSMM in learning effective similarity metrics for multi-label classification.

Local Rademacher Complexity for Multi-label Learning.

Learning Vector-valued Functions with Local Rademacher Complexity and Unlabeled Data

Robust Extreme Multi-label Learning

Multi-Class Learning Using Unlabeled Samples: Theory and Algorithm.

Learning Label-Specific Multiple Local Metrics for Multi-Label Classification

Multi-label Linear Discriminant Analysis with Locality Consistency.

Learned Random Label Predictions as a Neural Network Complexity Metric

Multi-Label Learning With Local Similarity Of Samples

Learning from Multi-Dimensional Partial Labels.

Multi-Label Learning by Exploiting Label Correlations Locally

Generalization Analysis for Multi-Label Learning

The Linear Geometry Structure of Label Matrix for Multi-label Learning

A Consistent Lebesgue Measure for Multi-label Learning

Labeling Information Enhancement For Multi-Label Learning With Low-Rank Subspace

Transductive Rademacher Complexity and its Applications

Low Rank Multi-Label Classification with Missing Labels

Penalized Partial Least Square Discriminant Analysis with ℓ 1 - Norm for Multi-Label Data

Ml-rbf: RBF Neural Networks for Multi-Label Learning

Unbiased Risk Estimator to Multi-Labeled Complementary Label Learning

Label Specific Multi-Semantics Metric Learning for Multi-Label Classification: Global Consideration Helps

Exponential Tail Local Rademacher Complexity Risk Bounds Without the Bernstein Condition