Research Projects

There has been a large push from the Computer Systems and Hardware communities for a move away from the typical Von Neumann architecture for processing. This push primarily comes from a desire to save energy on computation, considering how widespread computer usage has become.

One initiative for moving away from the Von Neumann architecture, especially with regard to Artificial Intelligence & Machine Learning, is called Spiking Neural Networks.

The primary goal of this project is to bring the fundamental building blocks of Natural Language Processing (NLP) over to this new computing paradigm, where we aim to achieve comparable performance to the state-of-the-art approaches while saving on the amount of energy expended doing so.

This project, in accordance with the taxonomy outlined in this paper, aims to implement the first fully-automated, end-to-end Automatic Machine Learning pipeline. This pipeline interacts with users as a virtual assistant would, guiding them through formulating their problem on the data they provide. Once the problem is formulated, the agent curates the data and sends it through a pre-existing AutoML pipeline to get results.

More details on this project can be found here.