ML models’ estimates may be less accurate for disadvantaged subgroups, an MIT study finds


    Explanation methods that mimic a larger model by generating simple approximations of machine learning predictions are sometimes used by researchers.Machine learning (ML) may assist human decision-makers in high-risk situations. For example, an ML model might forecast which law school applicants have the greatest chances of passing the bar exam and help an admissions officer choose which students should be admitted. Some even found out that ML systems could detect