Research map / Learning theory and algorithms / Fairness
Beyond Parity: Fairness Objectives for Collaborative Filtering
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Add value-unfairness penalty (ablation) | Matrix factorization · Synthetic course recommendation | value unfairness: 0.649 $\pm$ 1.8e-02 -> 0.245 $\pm$ 1.4e-02 | Helped |
| Add value-unfairness penalty (ablation) | Matrix factorization · MovieLens Million Dataset | value unfairness: 0.234 $\pm$ 6.3e-03 -> 0.223 $\pm$ 6.9e-03 | Helped |
| Add underestimation-unfairness penalty (ablation) | Matrix factorization · Synthetic course recommendation | underestimation unfairness: 0.107 $\pm$ 6.5e-03 -> 0.060 $\pm$ 3.5e-03 | Helped |
| Add absolute-unfairness penalty (ablation) | Matrix factorization · Synthetic course recommendation | underestimation unfairness: 0.107 $\pm$ 6.5e-03 -> 0.135 $\pm$ 6.2e-03 | Hurt |
| Add non-parity penalty (ablation) | Matrix factorization · Synthetic course recommendation | value unfairness: 0.649 $\pm$ 1.8e-02 -> 0.697 $\pm$ 1.8e-02 | Hurt |
| Add non-parity penalty (ablation) | Matrix factorization · MovieLens Million Dataset | non-parity unfairness: 0.036 $\pm$ 1.3e-03 -> 0.010 $\pm$ 1.5e-03 | Helped |
| Add fairness penalties, including value, absolute, underestimation, or overestimation objectives (ablation) | Matrix factorization · MovieLens Million Dataset | squared rating error: 0.887 $\pm$ 1.9e-03; 0.886 $\pm$ 2.2e-03; 0.887 $\pm$ 2.0e-03; 0.888 $\pm$ 2.2e-03; 0.885 $\pm$ 1.9e-03 | No clear effect |
Closest papers by meaning
- User-item fairness tradeoffs in recommendations (NeurIPS 2024 · similarity 0.90)
- Interpolating Item and User Fairness in Multi-Sided Recommendations (NeurIPS 2024 · similarity 0.89)
- Intra- and Inter-group Optimal Transport for User-Oriented Fairness in Recommender Systems (AAAI 2024 · similarity 0.89)
- Reconciling Predictive and Statistical Parity: A Causal Approach (AAAI 2024 · similarity 0.87)
- FairJob: A Real-World Dataset for Fairness in Online Systems (NeurIPS 2024 · similarity 0.87)
- Bridging Jensen Gap for Max-Min Group Fairness Optimization in Recommendation (ICLR 2025 · similarity 0.86)
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.