Thread by @jm_alexia

Panelists of the #montreal #AI #ecoystem @froginthevalley @marktmaclean @MILAMontreal @rllabmcgill @NarjesBoufaden @PhilBeaudoin of @SymposiumAi, we need to talk. I’m bringing other AI researchers (from Montreal or not) into the discussion because I feel it’s very important. 1/15

It relates to more than just Montreal, but the general AI ecosystem: @atg_abhishek, @goodfellow_ian, @fchollet, @hardmaru, @element_ai, @Montreal_IA, @johnolafenwa, @MTLINTL, @math_rachel, @mtlaiethics, @AndrewYNg. 2/15

I didn't get the chance to make a comment yesterday as time was over. There was a lot of talk about how great the ecosystem is great except for some minor communications flaws, but what nobody even mentioned is how incredibly closed this ecosystem is to outsiders. 3/15

At the panel, I only heard once the word “NGO” by Yoshua Bengio, but otherwise, it was always Academia vs Industry as if it was a binary thing. The reality is very different, there are: NGO, people who got out academia because of the abuse/exploitation they endured 4/15

or simply because they couldn’t deal with the politics and environment (there are thousands of articles about this). Then, there are people like me who attempted many times to get enrolled in a graduate program but were rejected. 5/15

If someone like me with 10 papers in statistics (8 published, 2 first-authors) and now 2 papers in AI (unpublished) got rejected repeatedly, it tells you that there is a large amount of extremely talented people who are prevented from participating in the ecosystem. 6/15

By not mentioning them and not considering them in the panel, it’s like pretending they don’t even exist. Someone mentioned the broken pipeline, saying that it’s very difficult to get in a PhD program and Joelle Pineau response was that it would be fixed soon. 7/15

She mentioned 20-40 extra spots for PhDs. I’m sorry but this doesn’t fix the problem. She also mentioned that they had 300 applicants for GANs research and she said maybe people should instead focus on other more interesting problems. 8/15

This shows that she would not accept people working in GANs. This is completely valid, it’s her rights to decide who she wants to accept as students, but this shows that there is immense gatekeeping and that just adding few places is not fixing the problem. 9/15

Out of these 300 applicants interested in GANs, there might 50 that are extremely talented and will have to stop doing research and go into some unrelated job or the industry. 10/15

We lose so many talents that would love to do science (regardless of money) outside of academia. And then, we do not even mention their existence in these panels, which worsen the sense of the ecosystem is fully closed to “others”. This needs to be at least acknowledged! 11/15

I strongly believe in a future with basic income so that researchers can work even if they are not part of academia; in the meantime, there is still plenty of AI hobbyists working on AI on the free time, we shouldn’t ignore them just because they don’t have a fancy title. 12/15

My suggestion is to create some kind of possibly international AI lab outside of academia with NO HIERARCHY, where anyone can propose a project and/or be a PI (without needing any fancy title). If people are interested in this idea, I’d love to talk more about it. 13/15

I don’t have the fund to start such an endeavor but I strongly believe that this is the future, we need to stop pretending that everything is fine and that people outside of academia and the industry don’t exist 14/15

or are nor “real researchers” (not from the panel, but I have heard this so many times). 15/15