Live data from Hacker News

Montreal Is Leading the AI World Takeover

cloudraker.com

21–30 of 82 posts

Re: Montreal Is Leading the AI World Takeover

#21
post #14

Earlier quoted context omitted.

Here's one data point: There are about 150 academic AI researchers in in an area of 1.4 square kilometre radius in downtown Montreal.

That's quite impressive. If you go by sheer numer in a city, Pittsburgh could have more because of Carnegie Mellon, affiliated institutes, Uber research center, and startups stemming from them. The number of AI researchers in the Bay area might be even larger. Many researchers associated with large companies publish papers at a rate comparable to academics. Boston, Redmond, and Beijing are some other contenders. Edit…

If you have some free time ( ;) ) I would be very interested in seeing a proper ranking. Maybe they're limiting themselves to "Deep Learning AI" when they say Montreal has one of the biggest concentrations of researchers in the world.

Re: Montreal Is Leading the AI World Takeover

#22
post #21

Earlier quoted context omitted.

That's quite impressive. If you go by sheer numer in a city, Pittsburgh could have more because of Carnegie Mellon, affiliated institutes, Uber research center, and startups stemming from them. The number of AI researchers in the Bay area might be even larger. Many researchers associated with large companies publish papers at a rate comparable to academics. Boston, Redmond, and Beijing are some other contenders. Edit…

If you have some free time ( ;) ) I would be very interested in seeing a proper ranking. Maybe they're limiting themselves to "Deep Learning AI" when they say Montreal has one of the biggest concentrations of researchers in the world.

I guess you are right regarding Deep Learning. Even then, it might be hard to beat the Bay Area. Unfortunately it would take a lot of work to do a proper ranking (and many companies might not make the info public). Maybe we need a better AI for that. ;)

Re: Montreal Is Leading the AI World Takeover

#24

Earlier quoted context omitted.

Could you name some DL companies in Berlin please? I'm very interested in this and based here, so I'd love to learn more about interesting projects here. So far my impression was that the "deep tech" scene in Berlin is minuscule compared to e.g. London and SV, and that most companies here work in ad tech and e-commerce (two fields which I have no interest in at all).

Do you live in Berlin? Am also looking for like minded people. Munich seems to have more to be honest, and has a new ML incubator.

Do you know the name of the incubator?

Re: Montreal Is Leading the AI World Takeover

#25

"Leading" Self-delusions of grandeur of Quebeckers are funny. Too bad they can't attract talent and can't use the ones they attracted there Meanwhile some top experts are in Silicon Valley, some others in London, there's a lot of ML investment in Berlin and other cities Montreal is a great city but gets dragged down by the French Quebec BS

Could you name some DL companies in Berlin please? I'm very interested in this and based here, so I'd love to learn more about interesting projects here. So far my impression was that the "deep tech" scene in Berlin is minuscule compared to e.g. London and SV, and that most companies here work in ad tech and e-commerce (two fields which I have no interest in at all).

By pure chance I know one exception to this ad-tech/e-commerce bias in Berlin: I'm at a Berlin-based company[0] working on DL and deep RL for industrial applications (think robots, optimisation and control).

[0] micropsi-industries.com/join_us

Re: Montreal Is Leading the AI World Takeover

#26
post #14
post #5

No aspiring or self-respecting AI researcher could take an article like this at face value. "The largest concentration of independent AI researchers in the world"? Citation needed. Better yet, data.

Here's one data point: There are about 150 academic AI researchers in in an area of 1.4 square kilometre radius in downtown Montreal.

Montreal has more students/capita than anywhere in North America, due to a clustering of big Unis downtown.

So, in a way, it has a 'geography' advantage to such numbers :).

That said, I'm highly doubtful that they will be able to effectively leverage the knowledge.

Bombardier etc. are slow, monolithic entities, who will no doubt take AI up, but I wonder how effectively and profitably for local entities.

Air Canada is an airline operator - much like Verizon etc. - they 'operate' they don't 'innovate'.

On the startup side, Montreal is weak - no as bad as other places unknown for startups (say, Houston?), but it does not have the 'acquiring entities' to take up the funnel. Toronto has this disadvantage as well, but because it's bigger, English, and a little more well-known, I think it's less of a problem.

The language issue cannot be overlooked - a little less than 1/2 of Montreal is primarily English speaking, and though basically every Francophone speaks English, it does affect how networks are oriented. Francophone networks are weak. They are very local, and poor, and for a variety of reasons a little 'self oriented'. This is a Quebec thing, but also a French thing, and naturally a thing that happens when you're an enclave of people speaking something other than English in North America.

Finally - there is a lack of industrial spirit in Montreal. They are academically, socially and creatively inclined, but they don't want to build 'products and services' and aren't keen on making 'big bucks'.

The series of 'big business' magnates that came up in Quebec over the last 40 years have been mostly locally oriented: taking up TV/Media, Energy and retail banking sectors during the 'quiet revolution'. The big 'startup' here is a Circus!

In my very Francophone neighbourhood, 100% of the waiters and waitresses will tell you they are 'in school' or 'actors' or 'photographers' - yet many of them are late 20's early 30's - and it's not exactly hollywood around here. An acquaintance calls herself a 'playwright' and she's written 2 tiny productions for the local theatre with 2 actors each - no props or anything. She's late in her 'career' already. This attitude is due to a bunch of historical factors and it runs deep.

I'm not suggesting people can't build great companies here, and surely some will pop up - and there are some natural advantages such as lots of students and ridiculously low cost of living ... but I'm not sure if anything big will come of it.

To grow a startup you need all sorts of motivated people with special skills: recruiting, supply chain, sometimes niche stuff like ASICS etc. - I never, ever meet anyone here remotely in that vein.

It's a great city, but I'm not sure it will ever be the 'leader' in anything applied or industrious. I predict a few wins here and there.

Re: Montreal Is Leading the AI World Takeover

#27
post #14

Earlier quoted context omitted.

Here's one data point: There are about 150 academic AI researchers in in an area of 1.4 square kilometre radius in downtown Montreal.

Montreal has more students/capita than anywhere in North America, due to a clustering of big Unis downtown. So, in a way, it has a 'geography' advantage to such numbers :). That said, I'm highly doubtful that they will be able to effectively leverage the knowledge. Bombardier etc. are slow, monolithic entities, who will no doubt take AI up, but I wonder how effectively and profitably for local entities. Air Canada is…

Boston

Re: Montreal Is Leading the AI World Takeover

#28
post #7
post #6

Has AI advanced from the 1990s? Or has there been a change in the environment it can be applied to, with the rise of big data, big infrastructure - big everything?

They learned how to train deep nets ten years ago (and are getting better fast as new techniques are invented). As a result, neural nets started taking over. Big Everything is important, but without the new techniques for deep nets, AI would not be so sexy right now.

This should not be downvoted, he is essentially correct.

'AI' has always been around, but the 'big leap' we've seen recently is entirely due to researchers finally able to make Neural Nets actually work :).

'Deep Learning' really refers to a specific kind of Neural Network.

The first application was image recognition. It's also used a lot now in voice recognition - and they're trying to jam it into everything possible.

I think it's basically fair to say that at least in 2016 AI pretty much boils down to Deep Learning / Neural Nets to move something more classical along in terms of capability.

What 'AI' means changes over time, but I think it's safe to say the renewal and hype is based on the 'big eureka' in Neural Nets and their relevant applications.

Re: Montreal Is Leading the AI World Takeover

#29
post #25

Earlier quoted context omitted.

Could you name some DL companies in Berlin please? I'm very interested in this and based here, so I'd love to learn more about interesting projects here. So far my impression was that the "deep tech" scene in Berlin is minuscule compared to e.g. London and SV, and that most companies here work in ad tech and e-commerce (two fields which I have no interest in at all).

By pure chance I know one exception to this ad-tech/e-commerce bias in Berlin: I'm at a Berlin-based company[0] working on DL and deep RL for industrial applications (think robots, optimisation and control). [0] micropsi-industries.com/join_us

That's way cool. I'm a cybernetics graduate, and I've long thought that the current ways of regulating processes could probably be improved upon with ML. Tough to get any one to pay attention because PID is usually good enough, and the field generally value mathematical analyzability, which ANNs are not particularly known for.

Is there any way I can contact you directly?

Re: Montreal Is Leading the AI World Takeover

#30
post #21

Earlier quoted context omitted.

If you have some free time ( ;) ) I would be very interested in seeing a proper ranking. Maybe they're limiting themselves to "Deep Learning AI" when they say Montreal has one of the biggest concentrations of researchers in the world.

I guess you are right regarding Deep Learning. Even then, it might be hard to beat the Bay Area. Unfortunately it would take a lot of work to do a proper ranking (and many companies might not make the info public). Maybe we need a better AI for that. ;)

If you strictly looked at published papers on the subject matter and looked at the location of the authors you could do this relatively simple.
Post reply on HN