Research map / Learning theory and algorithms / Fairness
Beyond Match Maximization and Fairness: Retention-Optimized Two-Sided Matching
Ren Kishimoto, Rikiya Takehi, Koichi Tanaka, Yoji Tomita, Masahiro Nomura, Riku Togashi, Yuta Saito
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use MRet instead of Max Match | ranking · Synthetic matching | user retention rate | Improved |
| use MRet instead of FairCo | ranking · Synthetic matching | user retention rate | Improved |
| maximize matches (Max Match) instead of uniform random ranking | ranking · Synthetic matching | user retention rate | No clear effect |
| increase popularity skew kappa for Max Match | Max Match · Synthetic matching | user retention rate | Hurt |
| add fairness-of-exposure term (FairCo) to match maximization | ranking · Synthetic matching, skewed popularity | user retention rate | No clear effect |
| use FairCo lambda=0.1 instead of default 100 (ablation) | FairCo · Synthetic matching | user retention rate | Improved |
| use FairCo with equal-exposure criterion (ablation) | FairCo · Synthetic matching | user retention rate | Improved |
| learn retention function with XGBoost instead of oracle f (MRet vs MRet best) (ablation) | MRet · Synthetic matching | user retention rate | Hurt |
| use MRet instead of fairness baselines on sparse real dating data | ranking · Real dating platform data | user retention rate | Improved |
| use FairCo/Uniform instead of Max Match on sparse real data | ranking · Real dating platform data | user retention rate | Hurt |
| increase retention-model training data to n=16000 (ablation) | MRet · Real dating platform data | user retention rate | Improved |
| apply MRet when retention function is non-concave | MRet · Real dating platform data | user retention rate | No clear effect |
Closest papers by meaning
- Leveraging Opposite Gender Interaction Ratio as a Path towards Fairness in Online Dating Recommendations Based on User Sexual Orientation (AAAI 2024 · similarity 0.85)
- User-item fairness tradeoffs in recommendations (NeurIPS 2024 · similarity 0.84)
- Putting Gale & Shapley to Work: Guaranteeing Stability Through Learning (NeurIPS 2024 · similarity 0.84)
- Promoting Fairness Among Dynamic Agents in Online-Matching Markets under Known Stationary Arrival Distributions (NeurIPS 2024 · similarity 0.83)
- Beyond the Half Approximation: Fair and Efficient Online Class Matching (NeurIPS 2026 · similarity 0.83)
- Intra- and Inter-group Optimal Transport for User-Oriented Fairness in Recommender Systems (AAAI 2024 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
