Lex Fridman Podcast #148 – Charles Isbell and Michael Littman: Machine Learning and Education
33 snips
Dec 26, 2020 Charles Isbell, Dean of the College of Computing at Georgia Tech, and Michael Littman, a computer scientist at Brown University, dive into the intersection of machine learning and education. They discuss the role of AI in sparking student curiosity and the importance of resilience in learning. Their light-hearted banter touches on the contrasting cultures of their universities, and they humorously explore how technology reshapes education post-COVID. Their deep friendship is evident as they reflect on personal growth and mutual influences.
AI Snips
Chapters
Books
Transcript
Episode notes
Machine Learning vs. Statistics
- Machine learning is more than just statistics, encompassing rules, symbols, and software engineering principles.
- The difference in perspective between computer scientists and statisticians influences research outcomes.
NeurIPS vs. ICML
- In the early 2000s, ICML showcased machine learning done by computer scientists, while NeurIPS felt like computer scientists trying to impress statisticians.
- This difference in target audience has an impact on what researchers prioritize and what they worried about.
Data-Centric Assignments
- Charles Isbell designs machine learning assignments where students analyze data sets rather than implementing algorithms from scratch.
- This focuses on understanding data characteristics and differences between algorithms.
