Researchers have developed a novel framework to personalize policies in two-sided marketplaces, specifically addressing the challenge of conflicting objectives between different user groups. This framework was applied to a job marketplace to optimize free-value thresholds for job listings, aiming to improve target metrics while adhering to engagement guardrails. The system integrates ensemble-based hybrid ranking models for multi-objective optimization and a treatment effect extrapolation method to extend experimental findings to untested policy levels, demonstrating effective personalization even with constrained experiments. AI
IMPACT This research offers a new methodology for optimizing marketplace policies under constraints, potentially improving user experience and efficiency in online platforms.
RANK_REASON The cluster contains a research paper detailing a new methodology for personalization in marketplaces. [lever_c_demoted from research: ic=1 ai=0.7]
- arXiv
- ensemble-based hybrid ranking models
- Hugging Face
- machine learning
- Personalizing Marketplace Policies with Competing Objectives and Constrained Experiments: Evidence from a Job Marketplace
- treatment effect extrapolation
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
