Reproduce “RecRec: Algorithmic Recourse for Recommender Systems” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Sahil Verma, Ashudeep Singh, Varich Boonsanong, John P. Dickerson, Chirag Shah
The RecRec framework provides content providers with actionable feedback to improve their items' visibility in recommender systems. It achieves near 100% success rates in elevating item rankings for target user groups across diverse datasets, while requiring minimal feature changes and preserving overall recommendation quality.
Key insights
- Algorithmic recourse for recommender systems needs to focus on content providers' ability to improve their item's ranking for specific user groups.
- Effective recourse should be valid (improve ranking), sparse (minimal feature changes), and have minimal side-effects on other recommendations.
- The RecRec framework works in a black-box manner with existing recommender systems, making it practical for real-world proprietary systems.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.