Academic Seminar: Compression and Generation

Event Description:

The explosion of generative AI over the past few years is part of a long history of building AI systems that are able to not just make predictions or decisions, but are able to learn the underlying structure of the data to generate new data not present in the training set. Likewise, this technology is used to learn useful compressed representations that can be used for a wide variety of tasks, building the backbone behind all modern AI systems. Learn some of the techniques used by generative AI systems today, from Boltzmann machines to variational auto-encoders to auto-regressive models to diffusion models