Slashdot Mirror
How Canada Ended Up As An AI Superpower
pacopico writes: Neural nets and deep learning are all the rage these days, but their rise was anything but sudden. A handful of determined researchers scattered around the globe spent decades developing neural nets while most of their peers thought they were mad. An unusually large number of these academics -- including Geoff Hinton, Yoshua Bengio, Yann LeCun and Richard Sutton -- were working at universities in Canada. Bloomberg Businessweek has put together an oral history of how Canada brought them all together, why they kept chasing neural nets in the face of so much failure, and why their ideas suddenly started to take off. There's also a documentary featuring the researchers and Prime Minster Justin Trudeau that tells more of the story and looks at where AI technology is heading -- both the good and the bad. Overall, it's a solid primer for people wanting to know about AI and the weird story of where the technology came from, but might be kinda basic for hardcore AI folks.
when it wasn't profitable on a 5 year timescale like they do here in the states.
Hi! I make Firefox Plug-ins. Check 'em out @ https://addons.mozilla.org/en-US/firefox/addon/youtube-mp3-podcaster/
Lots of really smart people attracted to doing advanced things in Canada does not result in a product people need later.
https://en.wikipedia.org/wiki/...
But its all fun in Canada doing AI until then.
Could it be like the MCM/70 https://en.wikipedia.org/wiki/... with Canada doing pioneering AI work?
Domestic spying is now "Benign Information Gathering"
Plus, Yann LeCun never went to Canada long term. He was at NYU and now is at Facebook heading up their AI division. Hinton is split between Google and UOFT. The headline and summary are a bit overblown.
It is funny to see the industry still trying to flog neural networks. Pathetic.
Lots of prime time computing doesn't run on your hardware. Can everyone with internet access connect to an AI and get it to help with daily tasks and issues? I think we do. Frequently, the daily tasks and issues with which these AIs help are not our own.
"Forcing everyone out" is what happens when you tried so hard to keep out foreigners by restricting new housing and now there isn't enough housing for both the current residents and the newcomers. In other words, you did it to yourself!
Any sufficiently unpopular but cohesive argument is indistinguishable from trolling.
You're an idiot.
"Old man yells at systemd"
Hi,
I lived in Canada for over 30 years and this isn't exactly the case. Yes they have courses on it at UofT and some great minds have come out of there - but not so fast. Myself and many of my peers in the same area moved to the US simply because lack of work... Now working at Microsoft in the US and giving speeches on ML topics.
Canadian artificial intelligence is no match for American natural stupidity!
Oh, you mean Eh Aye, eh?
#DeleteFacebook
Sure I can.
I just need to redefine the word "planet" to "city".
#DeleteFacebook
This is probably the biggest AI hype in the long running AI hype-->trough-->hype cycle we've had going on since the 50s-60s.
1) the only reason AI and specifically NN are meeting the success they have, such as they are, is brute force. These are the same algos and designs that were around in the 80s and 90s, really, not much "progress" has happenend sionce then however, CPUs have become much more powerful and memory is much bigger and faster and ways of connecting those CPUs together are now well understood and thanks to the internet, theer is a treasure trove of data sutiable to fewed into these designs, for example, VoIP feeding into speech recognition.
2) the "big wins" of AI like world champioinship at Go and Deep Blue in chess are in fields that are dleiberately constructed by humans to flummox human information processing strengths and play to human information processing weaknesses. That's what a game is. Games generally are designed to go beuyond human capabilities in look-ahead- the ability to reason through a long series of if-then-what cycles . We actually have a short horizon for this kind of thing, along with terrible short-term memory. Computers OTOH have astronomical capabilites in these areas. This gives them a preternatural advantage in games.
If Kasparov had the on one billioneth of the look-ahead , reason-ahead ability that Depp Blue do you think he still would have lost? Because I ampretty sure he would be unbeatable.
3) We have zero idea of how the human brains works. Zero. We are at about the same level of understandin our brains as the ancients were when they realized that, hey, the heart... it's a *pump*.....
As an example, to just grab at another field completely randomly, we still dont understand how dust collection cyclones behave as they fine extract particles from swirling air. People get their PhDs studying this, advancing our understanding. Another example. There are organs in the body which have just recently been recognized as organs. Google these things; I am not making them up.
But people like Kurzweil (sp?) want people to believe that we're just about ready to upload our consciousness into a computer. Meanwhile self-criving cars are on the road and actually, we havce zero idea how thye do what they do. what their actual representation , if you cna call it that, of what they see, if you can call it that, is. So of course they're killing people. Of course they are. But hey, we're the species who ignited an atomic bomb not yet certain in the knowledge that it wouldn't ignite the entire atomsphere, so hey, what did you expect.
AI as it is now isn't bullshit, but it's not anythwere near HAL-9000 level AI and it's so far away from it we can't even see the road to it. Everything from the computers it's based on to our understanding of basic physics will have to improve 1000 fold before we're anywhere near where the hype-money-machine says we are now.
You can also thank the stupid breathless media who take everything their told at face value for contributing to this current mass dleusion.
Can you suggest a source to try?
Nice try but I'm looking at Toronto's zoning code and it's filled with laws that limit how many housing units can be built within the city. Minimum parking requirements, minimum setbacks, maximum floor area ratios, height limits, and on and on.
Any sufficiently unpopular but cohesive argument is indistinguishable from trolling.
The kinds of questions I have are a little bit more, cost orientated. Currently I am very interested in topics like Health Care, Auto topics, and Taxes. I guess maybe Wolfram has already solved those singular issues. But it's the questions I will have that A.I. needs to be able to solve for that I would measure usefulness by.
NASA actually put american men on the moon, onboard a wildling-designed spaceship, propelled by a nazi-designed rocket.
Well yeah, the US has well been known as the beneficiary of brain drains for a very long time now. This is not news at all. Wernher Von Braun's team of rocket scientists, and Avro's top engineers inclusive. Why be a wildling when going south of the wall gives you a much larger selection of companies to work for (and better weather...why would you want to live in a frozen wasteland? You already know most of them would prefer to be south, given 95% of the wildlings live within 150 miles of the wall.) Each of them will pay more than anything you could get north of the wall, plus you'll have access to much better facilities and resources than you would have had there.
It's even better these days: Rather than just NASA, there are three private sector companies developing and/or launching rockets, with one of them creating the most technologically advanced rockets in the world; even better than every government.
[snipped complaints about laws not being enforced against Muslims]
it's filled with laws that limit
Yeah, and laws that target rape didn't get enforced when it was muslims breaking them:
A neighbour had called the police after hearing the [13 year old] girl scream. The girl was arrested for being drunk and disorderly, but the [muslim] men were not questioned
ecause most of the perpetrators were of Pakistani heritage, several council staff described themselves as being nervous about identifying the ethnic origins of perpetrators for fear of being thought racist; others, the report noted, "remembered clear direction from their managers" not to make such identification.
There's more, you can read all about it.
Like you said, "Nice try" ... but your assertion that this could not be happening because there were laws against it is clearly stupid, and you should feel stupid for making that assertion.
I'm a minority race. Save your vitriol for white people.
Y'houser!
Sorry.
Confucius say, "Find worm in apple - bad. Find half a worm - worse."
Aye. Well, actually Nay, but you set me up there.
Not that this wasn't entirely predictable.
Do you want polite Terminators? Because this is how you get polite Terminators.
I don't think I communicated correctly. I don't know of anyone renting general access to their AI for it to perform various tasks like an AI MTurk. However, all the usual network services that people use every day are using prime time AIs to help with the service they perform for you. Google has AIs to decide what you're thinking of when you write search terms and what to show you when you visit Google News. Facebook has AIs to decide what advertising to show you. When you talk to a virtual assistant, there's AIs helping to try and get the assistant to do what you say, and helping to decide what advertisements your requests suggest would be best to throw your way when you're at a site with advertisers that pay for the AIs' suggestions. I feel confident that they're using AIs to help make catchpas that other AIs can't defeat. IBM bought the digital part of The Weather Channel so that they could use its data for selling weather predicting services of its AI. Here's an article telling how more than 100 web services are using Watson instances to power apps and other online business. Here's another article telling how Ross Intelligence is using a Watson to help lawyers act like they've read all the recent decisions. Eviebot and Cleverbot will chat with you. None of these are at the level of science fiction AIs, but they are providing actual value to their owners, and often to the customers of their owners.
until you get them going about either not speaking French or speaking French, they hassle you about the alcoholic beverages and sausages in a cooler buried under camping gear, go on-and-on about how much nicer Toronto is than any U.S. city, torture you to boredom with endless gossip about ice dancers or hockey players, and warn you "I'll be back" with the speech accent of Jordan Peterson before they crash a Bombardier Ski-Doo through the glass doors?
This is really cool to consider various A.I.'s that are in use today by businesses; for commerce. The subtle application I am in search of is an A.I. that does not represent some mercantile oriented transaction; but as a 3rd party observer in nature. As any observer of elementary economics will agree, "Customers" are the other half of the economic cycle. My search is for a Customer positioned solutoin.
Poor A/C, these are solved issues. A.I. is a method, that solves questions, not just issues. Nor would this A.I. be closed source.
AI, eh?
Is Watson beyond your budget?