Workshop on Artificial Intelligence, Video Analysis, and Human Rights
June 6-7, 2017
Over the past two years, Carnegie Mellon University’s Center for Human Rights Science (CHRS) has partnered with computer scientists and human rights organizations to develop efficient methods for the preservation and analysis of large video collections (from a few dozen hours to thousands of hours) that document human rights violations.
Using the latest developments from machine learning, computer vision, and signal processing, this effort has enabled human rights organizations to: rapidly search massive video collections lacking metadata for specific objects, sounds, speech acts, and text; detect and estimate the size of human crowds; automatically detect gunshots in videos; auto-synchronize multiple videos based on sound; semi-automatically geolocate videos lacking metadata; create 3D scene and event reconstruction; and safely and securely preserving video for future use in advocacy and accountability efforts. CHRS partners have already used these tools and techniques to conduct investigations and advocacy efforts in Nigeria, Syria, and Ukraine, among others.
This workshop is part of our efforts to share tools and techniques with the broader human rights and civil liberties community, solicit feedback to improve them, and determine what new problems we should focus on over the next few years. In the process, we will introduce the fundamentals of artificial intelligence and machine learning, and demonstrate how these techniques can benefit the human rights community if used properly.