Research map / Learning theory and algorithms / Fairness
An External Fairness Evaluation of LinkedIn Talent Search
Tina Behzad, Siddartha Devic, Vatsal Sharan, Aleksandra Korolova, David Kempe
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 |
|---|---|---|---|
| apply DetGreedy post-processing (LinkedIn self-report) | LinkedIn Talent Search · MinSkew@100 | MinSkew@100: -0.259 -> -0.011 | Improved |
| external audit measurement vs LinkedIn reported value | LinkedIn Talent Search · MinSkew@25 | MinSkew@25: -0.011 -> -0.360 | Hurt |
| external audit measurement vs LinkedIn reported value | LinkedIn Talent Search · MinSkew@100 | MinSkew@100: -0.011 -> -0.213 | Hurt |
| increase cutoff k from 25 to 100 | LinkedIn Talent Search · MinSkew@k | MinSkew@k: -0.360 -> -0.213 | Improved |
| correct skew for integrality of small k (ablation) | LinkedIn Talent Search · Skew@k women | Skew@k | No clear effect |
| restrict race labels to White vs Other instead of four classes (ablation) | Ethnicolr labeling · race label validation | agreement with annotators (%): 65.5 -> 72 | Improved |
| restrict race labels to White vs Other instead of four classes (ablation) | annotation · race label validation | inter-annotator agreement (%): 82 -> 88 | Improved |
| compare women vs men at top ranks | LinkedIn Talent Search · churn k=25,50 | churn rate: women churn approximately 0.07 units more than men | Hurt |
| use BLS proportions instead of observed top-200 proportions as baseline | audit · general queries Q2 | deviation from group proportion | Mixed |
Closest papers by meaning
- Matchings Under Biased and Correlated Evaluations (NeurIPS 2025 · similarity 0.85)
- FairJob: A Real-World Dataset for Fairness in Online Systems (NeurIPS 2024 · similarity 0.85)
- Probing Social Bias in Labor Market Text Generation by ChatGPT: A Masked Language Model Approach (NeurIPS 2024 · similarity 0.83)
- Networked Inequality: Preferential Attachment Bias in Graph Neural Network Link Prediction (ICML 2024 · similarity 0.83)
- Understanding challenges to the interpretation of disaggregated evaluations of algorithmic fairness (NeurIPS 2025 · similarity 0.83)
- Identifying Reasons for Bias: An Argumentation-Based Approach (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.
