PrefRec: Recommender Systems with Human Preferences for Reinforcing Long-term User Engagement

Xue, Wanqi; Cai, Qingpeng; Xue, Zhenghai; Sun, Shuo; Liu, Shuchang; Zheng, Dong; Gai, Kun; An, Bo

doi:10.1145/3580305.3599473

Cited by 7 publications

(2 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These manually generated labels, along with other feedback like comments, are then utilized in a multi-task learning framework to improve recommendations [3]. However, these rule-based methods require notable manual effort and may not consistently align with the operational metrics [34] of the recommender system, such as user engagement. For example, the proposed unbiased watch time label in [37] has been observed to reduce user engagement, evidenced by a reduction in "share" [37].…”

Section: Introductionmentioning

confidence: 99%

LabelCraft: Empowering Short Video Recommendations with Automated Label Crafting

Bai,

Zhang,

et al. 2024

Proceedings of the 17th ACM International Conference on Web Search and Data Mining

View full text Add to dashboard Cite

Short video recommendations often face limitations due to the quality of user feedback, which may not accurately depict user interests. To tackle this challenge, a new task has emerged: generating more dependable labels from original feedback. Existing label generation methods rely on manual rules, demanding substantial human effort and potentially misaligning with the desired objectives of the platform. To transcend these constraints, we introduce LabelCraft, a novel automated label generation method explicitly optimizing pivotal operational metrics for platform success. By formulating label generation as a higher-level optimization problem above recommender model optimization, LabelCraft introduces a trainable labeling model for automatic label mechanism modeling. Through meta-learning techniques, LabelCraft effectively addresses the bilevel optimization hurdle posed by the recommender and labeling models, enabling the automatic acquisition of intricate label generation mechanisms. Extensive experiments on real-world datasets corroborate LabelCraft's excellence across varied operational metrics, encompassing usage time, user engagement, and retention. Codes are available at https://github.com/baiyimeng/LabelCraft.

show abstract

Section: Introductionmentioning

confidence: 99%

LabelCraft: Empowering Short Video Recommendations with Automated Label Crafting

Bai,

Zhang,

et al. 2024

Proceedings of the 17th ACM International Conference on Web Search and Data Mining

View full text Add to dashboard Cite

show abstract

“…• We collect the first reinforcement learning from human feedback (RLHF) dataset for long-term engagement optimization problem in recommendation and propose three new tasks to evaluate the performance of recommender 1 The work in this chapter has been published as Wanqi Xue, Qingpeng Cai, Zhenghai Xue, Shuo Sun, Shuchang Liu, Dong Zheng, Peng Jiang, Kun Gai, Bo An. PrefRec: Recommender systems with human preferences for reinforcing long-term user engagement [131]. Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), 2023.…”

Section: Guidances Of Human Preferencesmentioning

confidence: 99%

Robust and adaptive decision-making: a reinforcement learning perspective

Xue

View full text Add to dashboard Cite

show abstract

A Map of Exploring Human Interaction Patterns with LLM: Insights into Collaboration and Creativity

Li,

2024

Lecture Notes in Computer Science

View full text Add to dashboard Cite

PrefRec: Recommender Systems with Human Preferences for Reinforcing Long-term User Engagement

Cited by 7 publications

References 24 publications

LabelCraft: Empowering Short Video Recommendations with Automated Label Crafting

LabelCraft: Empowering Short Video Recommendations with Automated Label Crafting

Robust and adaptive decision-making: a reinforcement learning perspective

A Map of Exploring Human Interaction Patterns with LLM: Insights into Collaboration and Creativity

Contact Info

Product

Resources

About