作者位次 #4/8SIGIR 2026Poster· Information Retrieval (Industry)
SCOPE: Scalable Cross-Task Orthogonal Progressive Experts for Multi-Task Learning in Recommendations
Zixian Yang · Wei Xu · Li Li · Zhaokai Huang · You Li · Jianbin Lin · Wenliang Zhong · Can Ye
摘要
Online short-video recommendation systems, such as Alipay's Tab3, must model heterogeneous user feedback ranging from implicit playback signals to explicit interactions and domain-specific actions. This diversity poses significant challenges for Multi-Task Learning (MTL). Existing Mixture-of-Expe...