Original article was published by MIT Open Learning on Deep Learning on Medium
Chalk Radio — Episode 3, “Making Deep Learning Human with Prof. Gilbert Strang”
In this episode, we talk with Professor Gilbert Strang about humanizing the teaching and learning process in his courses about linear algebra and matrix methods in data analysis, signal processing, and machine learning.
Mathematics Professor Gilbert Strang is one of MIT’s most revered instructors; his courses, especially the perennially popular linear algebra course 18.06, have received millions of visits on OpenCourseWare, and his lecture videos have won him a devoted following on YouTube as well. (A sample YouTube comment on one of his lectures: “This is not lecture, this is art.”) A few years ago, Professor Strang began teaching a new course (18.065) focusing on the application of mathematical matrices to deep learning and AI. This new course is very unlike a typical undergraduate math course. For one thing, there’s no final exam — in fact, there are no exams at all! Instead, Professor Strang asks each student to spend the semester developing a project that applies the techniques they’re studying to some topic or problem they personally find interesting. In this episode, we hear from Professor Strang about his efforts to humanize math teaching, the value of thinking through problems in real time during lectures — even if it means getting stuck and having to backtrack! — and the importance of staying continually conscious of your students.
This episode was originally released on March 18, 2020.