New R2QRec method generates search queries from user logs for better recommendations

Researchers have developed R2QRec, a novel method for query-based recommendation systems that learns to generate effective search queries from user interaction logs. This approach uses a fixed retrieval system to create query supervision and evaluate recommendation performance. The system first verifies generated queries by checking if they can retrieve the observed items, then refines the query generation policy based on the rank of these items in a fused recommendation list. R2QRec achieves state-of-the-art results across three public datasets without requiring query annotations or a learned reward model, as only the query generator is trained. AI

IMPACT This method could improve the accuracy and personalization of recommendation systems by better understanding user intent from interaction data.

RANK_REASON Research paper detailing a new method for recommendation systems. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.IR (Information Retrieval) →

• alphaXiv
• arXiv
• CatalyzeX
• Connected Papers
• CORE Recommender
• DagsHub
• Gotit.pub
• Hugging Face
• Litmaps
• R2QRec
• ScienceCast
• scite Smart Citations

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →