The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence) Real world model explainability with Rayid Ghani - TWiML Talk #283
Jul 18, 2019
Rayid Ghani, the Director of the Center for Data Science and Public Policy at the University of Chicago, shares insights on applying machine learning for social good. He explores the crucial role of explainability in AI, emphasizing the need for relevant context in decision-making. Ghani discusses data-driven strategies from political campaigns and the ethical challenges in predictive modeling. He highlights the importance of trust and feedback mechanisms to improve model transparency, particularly in sensitive areas like healthcare and public safety.
AI Snips
Chapters
Transcript
Episode notes
Obama 2012 Campaign Data Science
- Rayid Ghani discusses his experience working on the Obama 2012 campaign's data science team.
- They used a closed-loop system, collecting voter data to personalize outreach across various channels.
Corporate Skills in a Campaign Setting
- Rayid Ghani's corporate experience translated well to the campaign, particularly in building complete data systems.
- The campaign also required new techniques, like integrating machine learning with behavior change methods.
Real-World Challenges in Public Policy
- Rayid Ghani emphasizes the challenge of applying machine learning to real-world public policy issues.
- He highlights the importance of scoping problems, identifying available data, and ensuring tangible impact on policy and society.