Today we are hosting the annual visiting day for PhD students accepted into our computer science department at NYU. It’s the first time this event is being held on-line (last year’s visiting day took place just before the pandemic turned everything upside down).
Much of the event consists of a faculty member describing, in about 15 minutes, a large body of research in some broad area, such as machine learning, compiler optimization, or natural language processing. Each such area has an entire group of faculty working on it, and involves collaborations with lots of students and other labs around the world.
I am impressed, as I watch these presentations, at how each presenter is able to describe what is really an enormous amount of research in a clear and economical way. There is something beautiful about seeing a large topic summarized briefly in a way that still captures the excitement of doing research in that topic.
Ideally it would be great to see these short presentations expanded out as a kind of fractal. As a student is interested in learning a bit more about any sub-topic, they should be able to go down one or more levels to learn more, and then continue to descend into further details or pop back up and move on to other sub-topics as desired.
Maybe we should be building tools to support the creation of fractal presentations. That might be a good next step in the evolution of on-line learning.