Below, Prof. Rama Ramakrishnan describes how 15.773 Hands-On Deep Learning was initially conceived and developed.
The course was first offered in the spring of 2022.
When I joined the Sloan faculty in 2019, I looked for a deep learning course in the Sloan course catalog and couldn’t find one. That surprised me. I’d had extensive prior experience with deep learning as an entrepreneur and was convinced it was a transformative technology—one that would be extremely useful for Sloan students, especially aspiring entrepreneurs, wanting to build intelligent new products or infuse intelligence into existing products and services. After a couple of semesters of teaching, I was also convinced that a significant fraction of Sloan MBA students wanted to roll up their sleeves and build things themselves, rather than just be consumers or managers of things built by others.
So I put together a proposal and a rough syllabus, then socialized it with a few colleagues. It resonated with my Sloan colleague Vivek Farias and we decided to create the course together. I built out the Computer Vision lecture sequence and tested it with the LGO program over the summer of 2021. It was very well received, and that encouraged us to build out the full version.
When designing the lectures and exercises, I self-imposed a critical design constraint. I wanted to make sure students without extensive mathematical training wouldn’t be blocked from understanding and building with deep learning models. So the course is intentionally “light on math but heavy on intuition” so that it is accessible to a much broader set of people.
We launched the course in Spring 2022. I have been delighted to see the tremendous interest from students. The course has been oversubscribed every year since launch.
