We are pleased to announce the papers accepted to the 2nd Workshop on Human-Centered Recommender Systems (HCRS@TheWebConf 2026). The workshop will be held as a half-day session on Tuesday, June 30, 2026, 10:00–13:30 (GST, UTC+4), online (meeting link to be announced). Congratulations to all authors!
Submissions appear in the OpenReview venue: HCRS@TheWebConf 2026.
Oral PresentationsOral
-
Award Citation. This paper presents RecoWorld, a simulated environment framework that enables multi-turn interactions between simulated users and agentic recommender systems. The core contribution lies in its dual-view architecture where users provide natural language instructions and the recommender learns to follow them iteratively, shifting the evaluation paradigm from static offline metrics toward dynamic, instruction-driven feedback. This design philosophy resonates strongly with the spirit of HCRS, as it treats user agency and long-term satisfaction as first-class objectives rather than byproducts of engagement optimization. By formulating the problem in a Gym-like RL framework with retention-based reward signals, the work lays important groundwork for the community to build recommender systems that are genuinely responsive to human intent.
Poster PresentationsPoster
🏆 Best Paper Award:
Awarded to “RecoWorld:
Building Simulated Environments for Agentic Recommender Systems”
by Fei Liu, Xinyu Lin, Hanchao Yu, Mingyuan Wu, Jianyu Wang, Qiang Zhang,
Zhuokai Zhao, Yinglong Xia, Yao Zhang, Weiwei Li, Mingze Gao, Qifan Wang,
Lizhu Zhang, Benyu Zhang, and Xiangjun Fan. Congratulations to the
authors!