Welcome to Northern Frontier, our new video interview series that showcases in-depth and engaging conversations with some of the brightest academic minds in AI research today.
With explosive buzz around the field, and AI talent in North America reaching a critical mass, it’s a perfect time to slow down, grab a coffee, and focus on the hard science instead of the hype. In this Q&A series, we’ve asked top machine learning researchers at leading universities to really delve into their respective areas of research interest and answer questions that get to the heart of what’s going on in ML academic circles right now.
Our inaugural Frontier conversation is with Prof. Tamara Broderick of the Massachusetts Institute of Technology. Prof. Broderick talks about tradeoffs in variational inference, the need for understanding the underlying assumptions in ML algorithms and models, and improving ease of use of this machinery for non-experts. She emphasizes the need for interpretability of prediction results and discusses how researchers from different academic backgrounds can help with building new algorithms to improve these factors overall.
Borealis AI launches fellowship program to support and grow emerging AI talent in Canada
Prof. Greg Mori on deep structured models for human activity