R2QRec: Learning Recommendation Queries from Retrieval Feedback
What it is
R2QRec is a method for query-based recommendation that learns to generate effective retrieval queries from user interaction logs without explicit query annotations. It uses a two-stage process: retrieval verification to select supervised targets and policy refinement to optimize the ranking of generated queries against a fixed retrieval system.
The problem it solves
Query-based recommendation systems struggle to learn effective query generation from interaction logs that lack explicit query annotations.
What you could use it for
- Improving recommendation systems that rely on text-based retrieval
- Generating search queries from implicit user behavior logs
- Optimizing retrieval-augmented generation for personalized recommendations
Why it matters
Enables recommendation systems to leverage retrieval-based architectures without requiring expensive manual query labeling, potentially improving personalization and scalability.
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.