Talk at INRIA Grenoble… check!
I spent a whole day last week visiting the INRIA Grenoble - Rhône-Alpes institute. I had the chance to present my work on machine learning applied to high-contrast imaging and direct detection of exoplanets. It was the first time this talk has been presented to a group of machine learning specialists (Thoth team – Modeling visual knowledge from large-scale data). Was I scared? Sure, but mostly, I was excited.
Since I started the position as
junior chair in data science a data scientist in academia, one of my main goals has been to spark new collaborations between labs/teams that do not interact with each other very often. But how to catalyze collaborations in academia when there is no protocol to follow and everyone is so focused on their own projects? Well, I’m trying to find that out.
Overall, it was a great experience for me, explaining the computer vision problems we encounter in astronomical high-contrast imaging to CS and ML researchers. We had interesting conversations and they showed me the cool problems they are solving with deep learning. Of course, by the end of the day I was saturated with information about neural network architectures, metrics, frameworks and so on. Hopefully, some of these ideas could be applied to finding exoplanets ;) and this visit was just the start of a fruitful collaboration.
Find below the slides of my presentation, in case you are interested.