@你的昵称_ 不太了解文本聚类的形式，难道是转化成embedding后再进行聚类？那用互信息或者对比学习都可以，只要有新意就行。。
论文粗读“Dual Mutual Information Constraints for Discriminative Clustering”Li, Hongyu et al. “Dual Mutual Information Constraints for Discriminative Clustering.” ...
感觉互信息工作已经有很多人做了，多模态聚类这块的可以看看DMIM(好像是郑州大学)，COMPLETER(pengxi)，Self-supervised Information Bottleneck based Multi-view Subspace Clustering (SIB-MSC)
论文粗读“Dual Mutual Information Constraints for Discriminative Clustering”Li, Hongyu et al. “Dual Mutual Information Constraints for Discriminative Clustering.” ...
