Hacker News Reader

Comment by defen | Hacker News Reader

Princeton has an extremely strong mathematics department (possibly the best in the world) plus the IAS is nearby. Perhaps they are looking for close collaboration with pure mathematicians? My understanding is that there are not very strong theoretical underpinnings to most of the machine learning techniques in use today.

View on HN

Indeed, the IAS has an upcoming program on machine learning.

The IAS program is underway since 2017 http://www.math.ias.edu/theoretical_machine_learning

Yes, although the special one year program (with lots of shorter term guests) doesn't kick off until 2019.

> My understanding is that there are not very strong theoretical underpinnings to most of the machine learning techniques in use today.

I wonder how come this is the case from looking at ArXiv, it seems there are plenty of low-hanging fruit relating to the Foundations of Machine Learning.

> My understanding is that there are not very strong theoretical underpinnings to most of the machine learning techniques in use today.

Yes, particularly with regards to generalization. There's a large number of theorists racing to explain that.

as far as i can tell from perusing the literature for the last ~year very few people care about theory in ML

? I know people at most schools I pay attention to getting paid to specifically work on that. In fact this is what probably one of the most famous applied mathematicians in certain engineering crowds (Ingrid Daubechies) is working on for her big project from what i understand

Reply on news.ycombinator.com