PopResume: Causal Fairness Evaluation of LLM/VLM Resume Screeners with Population-Representative Dataset

1. 1Collect population-representative resume data
2. 2Decompose the effect of protected attributes on screening outcomes using path-specific effect (PSE)-based fairness evaluation
3. 3Evaluate LLM/VLM resume screeners for causal fairness using PopResume
4. 4Analyze and address any disparities found in the screening process

Data scientists and AI engineers on a team can benefit from PopResume to evaluate the fairness of their resume screening systems, ensuring unbiased hiring practices

💡 PopResume enables fairness evaluation of resume screeners using population-representative data and PSE-based analysis

📊 Introducing PopResume: a dataset for causal fairness evaluation of LLM/VLM resume screeners 🚀