CSDA teaches Colombian students about trustworthy machine learning tools

Researchers from the Centre for Social Data Analytics recently developed and delivered a three-week intensive course about trustworthy machine learning tools for master’s students at the University of Los Andes (Colombia).

The elective course for students studying computer science, engineering, and data analytics showed them how to design accurate, trustworthy, and transparent machine learning tools to support high stakes decisions in the public sector.

CSDA Director Professor Rhema Vaithianathan and Senior Research Fellow Dr Diana Benavides-Prado developed the course - Machine Learning for Decision-Making Support in Public Sector High-Stakes Problems - and delivered it remotely via Zoom.

Students learned about the challenges that face public sector organisations working to deploy machine learning tools, including social licence, ethics, human-algorithm interaction, algorithmic fairness, and interpretability before exploring how to construct, evaluate and validate machine learning tools that are accurate, trustworthy, ethical, and transparent.

“It was great to see such a strong interest in this subject with 35 students who all had a genuine interest in learning how to use machine learning for social good,” says Dr Benavides-Prado. “The public sector use of machine learning is the focus of our work at CSDA, and this was a great opportunity to share what we have learned and how we do our work. It was also great to use case studies of some of our translational data analytics projects like the Allegheny Family Screening Tool to demonstrate the challenges and opportunities in this space.