Overview
Deep Learning and Reinforcement Learning (DLRL) Summer School 2020
This year, I was 1 of the 300 participants selected from around the globe for the DLRL 2020.
This is a prestigious annual event organized by the 3 major machine learning reaseach institutes in Canada: MILA (Montreal), Vector (Toronto) and Amii (Alberta), all brought together through CIFAR.
Overall impression
I was impressed by the number of women on the lecture panel and the number of women attending the event. I was also surprised at the number of new faculty members that introduced their novel areas of research and discussed their new lives as academic faculty.
DLRL in the times of COVID
Each year, the event is hosted by 1 of the 3 AI institutes and 2020 was supposed to be here at MILA in Montreal. Unfortunately, given COVID, it was held virtually (there was talks of cancelling it). It was certainly different. The organizers were quite impecceble with overall details, zoom links, Q&A sessions and discussion panels. We even had a virtual poster session via Gathertown and a virtual trivia night. Still, I think being physically there and meeting everyone in person would have been nice.
Common Talk themes
The full listing of the talks and access the slide and videos, should be made available soon.
As a common theme, there was a lot of pondering over what defines intelligence and how intelligent are AI systems we are building today. There was a real push in the direction of understanding and deconvoluting “black box” algorithms. There was also several talks about moving to more agnostic algorithms using more data and less bias by avoiding labels.
Links & resources
DLRL content is mostly shared freely after the event. For beginners, I find the 2017 edition a great resource.
For the 2015 lectures: DLRL2015
For the 2016 lectures: DLRL2016
For the 2017 lectures: DLRL2017
For the 2018 lectures: DLRL2018
For the 2019 lectures: DLRL2019
Youtube channels: Mila, vector, Amii
comments powered by Disqus