Research map / Probabilistic and scientific modeling / Molecules and proteins
Bias in Evaluation Processes: An Optimization-Based Model
L. Elisa Celis, Amit Kumar, Anay Mehrotra, Nisheeth K Vishnoi
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 |
|---|---|---|---|
| fix alpha=1 (remove skew parameter) (ablation) | bias evaluation model · JEE-2009 (Birth category) | TV distance: 0.09 -> 0.15 | Hurt |
| fix tau to entropy of f_D (remove information constraint) (ablation) | bias evaluation model · JEE-2009 (Birth category) | TV distance: 0.09 -> 0.21 | Hurt |
| fix alpha=1 (ablation) | bias evaluation model · JEE-2009 (Gender) | TV distance: 0.07 -> 0.15 | Hurt |
| fix tau to entropy of f_D (ablation) | bias evaluation model · JEE-2009 (Gender) | TV distance: 0.07 -> 0.19 | Hurt |
| fix alpha=1 (ablation) | bias evaluation model · Semantic Scholar (Gender) | TV distance: 0.03 -> 0.09 | Hurt |
| fix tau to entropy of f_D (ablation) | bias evaluation model · Semantic Scholar (Gender) | TV distance: 0.03 -> 0.23 | Hurt |
| fix alpha=1 (ablation) | bias evaluation model · Synthetic Network | TV distance: 0.03 -> 0.05 | Hurt |
| fix tau to entropy of f_D (ablation) | bias evaluation model · Synthetic Network | TV distance: 0.03 -> 10.0 | Hurt |
| use proposed model instead of implicit variance model | implicit variance model · Semantic Scholar (Gender) | TV distance: 0.13 -> 0.03 | Improved |
| use proposed model instead of implicit variance model | implicit variance model · Synthetic Network | TV distance: 0.22 -> 0.03 | Improved |
| use proposed model instead of multiplicative bias model | multiplicative bias model · JEE-2009 (Birth category) | TV distance: 0.14 -> 0.09 | Improved |
| use proposed model instead of multiplicative bias model | multiplicative bias model · JEE-2009 (Gender) | TV distance: 0.07 -> 0.07 | No clear effect |
Closest papers by meaning
- On Mitigating Affinity Bias through Bandits with Evolving Biased Feedback (ICML 2025 · similarity 0.85)
- Prompting Fairness: Integrating Causality to Debias Large Language Models (ICLR 2025 · similarity 0.84)
- Strategic Costs of Perceived Bias in Fair Selection (NeurIPS 2025 · similarity 0.84)
- Matchings Under Biased and Correlated Evaluations (NeurIPS 2025 · similarity 0.84)
- Identifying Reasons for Bias: An Argumentation-Based Approach (AAAI 2024 · similarity 0.84)
- Understanding challenges to the interpretation of disaggregated evaluations of algorithmic fairness (NeurIPS 2025 · similarity 0.84)
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.
