新类识别/领域自适应

对以下文章的记录：https://blog.csdn.net/weixin_42001089/article/details/120416890

Discovering New Intents via Constrained Deep Adaptive Clustering with Cluster Refinement
基于聚类细化的约束深度自适应聚类发现新意图

论文：https://ojs.aaai.org/index.php/AAAI/article/view/6353
论文源码：https://github.com/thuiar/CDAC-plus
解读：https://mp.weixin.qq.com/s/wHKWvP_qzyOyXPHXCHEmBQ

Discovering New Intents with Deep Aligned Clustering
利用深度对齐聚类发现新意图

论文链接: https://arxiv.org/pdf/2012.08987.pdf
代码链接: https://github.com/thuiar/DeepAligned-Clustering
解读：https://blog.csdn.net/anxiaomai520/article/details/123638953
https://www.jianshu.com/p/a6e832432539

Supporting Clustering with Contrastive Learning
对比学习横扫文本聚类任务

论文：https://arxiv.org/abs/2103.12953
代码：https://github.com/amazon-research/sccl
解读：https://mp.weixin.qq.com/s/tY_ZJyvEAXpEdnd5VlBO7g

A Unified Objective for Novel Class Discovery
新类发现的统一目标

论文：https://arxiv.org/abs/2108.08536
项目：https://ncd-uno.github.io
代码：https://github.com/DonkeyShot21/UNO
解读：https://blog.csdn.net/qq_25480605/article/details/120326075

Deep Open Intent Classification with Adaptive Decision Boundary
具有自适应决策边界的深度开放意图分类

论文链接: https://arxiv.org/pdf/2012.10209.pdf
代码链接:https://github.com/thuiar/Adaptive-Decision-Boundary
解读：https://mp.weixin.qq.com/s/N5ITCjpFhmd3xBN7S8hvaw

Generalized Source-free Domain Adaptation
广义无源域适应

论文：https://arxiv.org/abs/2108.01614
代码：https://github.com/Albert0147/G-SFDA
解读：https://mp.weixin.qq.com/s/lPwCUlcGASZTLoVQT1oC4w