Machine Learning in Human Rights: HURIDOCS

This lecture covers the application of machine learning in human rights using the HURIDOCS project as a case study. Topics include defining project goals, handling false positives and negatives, measuring model performance with precision and recall, data availability and quality, privacy concerns, and building transparency and trust in machine learning models. The instructor discusses the importance of understanding human rights issues, setting measurable metrics, and visualizing model confidence. The lecture emphasizes the need for inclusive datasets, consent, and clear oversight to mitigate bias. Case studies from UPR Info demonstrate the practical challenges and considerations in implementing machine learning for human rights monitoring.