SIGIR

Continuous Input Embedding Size Search For Recommender Systems

连续输入嵌入大小搜索用于推荐系统

Yunke Qu, Tong Chen, Xiangyu Zhao, Lizhen Cui, Kai Zheng, Hongzhi Yin

AI总结 本文提出CIESS，一种基于强化学习的连续输入嵌入大小搜索方法，通过在连续搜索空间中选择任意嵌入大小提升推荐效果。

Comments Accepted to SIGIR'23. Code is available at https://github.com/qykcq/Continuous-Input-Embedding-Size-Search-For-Recommender-Systems