In this page, Prof. Rama Ramakrishnan discusses the kinds of projects students undertook in 15.773 Hands-On Deep Learning.
There have been many memorable projects in the years that I’ve been teaching the course, and I’ve been very impressed with the creativity that students bring to these projects.
Broadly speaking, projects tend to come from a handful of different categories:
- One important category is students using deep learning to help with medical decision-making and a common theme is to identify the presence of disease from images, for example, to identify metastases from brain MRI scans. I’m always pleased to see contributions in this category, since it’s clearly a short path to human benefit if it works.
- Another big category is using deep learning to make marketing and sales operations much more efficient.
- Another category is applications in the sports analytics realm. I remember there was one project where the team built a model to look at sports videos and then tag the videos automatically with what sort of action was being displayed by a player in a particular video frame.
- Finally, using deep learning to make money in the stock market is a perennial favorite!
