Exploiting Multi-Label Correlation in Label Distribution Learning

Zhiqiang Kou jing wang yuheng jia xin geng

2023-08-03

Abstract:Label Distribution Learning (LDL) is a novel machine learning paradigm that assigns label distribution to each instance. Many LDL methods proposed to leverage label correlation in the learning process to solve the exponential-sized output space; among these, many exploited the low-rank structure of label distribution to capture label correlation. However, recent studies disclosed that label distribution matrices are typically full-rank, posing challenges to those works exploiting low-rank label correlation. Note that multi-label is generally low-rank; low-rank label correlation is widely adopted in multi-label learning (MLL) literature. Inspired by that, we introduce an auxiliary MLL process in LDL and capture low-rank label correlation on that MLL rather than LDL. In such a way, low-rank label correlation is appropriately exploited in our LDL methods. We conduct comprehensive experiments and demonstrate that our methods are superior to existing LDL methods. Besides, the ablation studies justify the advantages of exploiting low-rank label correlation in the auxiliary MLL.

Machine Learning

What problem does this paper attempt to address?

This paper aims to address the issue of the exponential output space in Label Distribution Learning (LDL). Specifically, many existing LDL methods tackle this challenge by leveraging the correlations between labels, but these methods mostly assume that the label distribution matrix has a low-rank structure. However, recent studies have shown that the label distribution matrix is often full-rank, which makes methods based on the low-rank assumption problematic. To solve the above issue, the authors propose a new approach that introduces an auxiliary Multi-Label Learning (MLL) process in LDL and utilizes low-rank label correlations in this process, rather than directly using the low-rank assumption in LDL. This method not only better captures the relationships between labels but also effectively improves the model's performance. Experimental results demonstrate the effectiveness and superiority of the proposed method.

Exploiting Multi-Label Correlation in Label Distribution Learning

Label Distribution Learning with Label Correlations Via Low-Rank Approximation

Label Distribution Learning by Exploiting Label Correlations.

Label Distribution Learning by Exploiting Label Distribution Manifold

Label Distribution Learning by Exploiting Sample Correlations Locally.

Exploiting Indirect Linear Correlation for Label Distribution Learning

Label Distribution Learning with Label Correlations on Local Samples

Label distribution learning by maintaining label ranking relation

Multi-Label Learning by Exploiting Label Correlations Locally

Incomplete Label Distribution Learning by Exploiting Global Sample Correlation

Weakly Supervised Label Distribution Learning Based on Transductive Matrix Completion with Sample Correlations

Label Distribution Learning: A Local Collaborative Mechanism

Incomplete Label Distribution Learning Via Label Correlation Decomposition

Multi-Label Learning by Exploiting Label Correlations with LDA

Label Enhancement with Sample Correlations Via Low-Rank Representation.

Label Distribution Learning Through Exploring Nonnegative Components

Inaccurate Label Distribution Learning

Learning Low-Rank Label Correlations for Multi-label Classification with Missing Labels

Incomplete Label Distribution Learning

Generalizable Label Distribution Learning

Multi-Label Learning with Global and Local Label Correlation