BriefGPT.xyz
Oct, 2024
基于监督对比学习的多标签分类相似性-不相似性损失
Similarity-Dissimilarity Loss with Supervised Contrastive Learning for Multi-label Classification
HTML
PDF
Guangming Huang, Yunfei Long, Cunjin Luo, Sheng Liu
TL;DR
本研究解决了多标签分类中正样本确定的挑战,提出了五种样本之间的关系以增强对比学习的效果。通过引入相似性-不相似性损失函数,该方法根据不同关系动态调整损失权重,从而显著提升了模型在多标签文本分类上的性能与鲁棒性。
Abstract
Supervised Contrastive Learning
has been explored in making use of label information for
Multi-label Classification
, but determining positive samples in multi-label scenario remains challenging. Previous studies
→