This article is written for people outside the field? Baidu is alright but a bit of a shambles. All the did was build a dense GPU rack for RNN training. Their imagenet results were from cheating. The theoretically interesting work isn't really happening there.
The Microsoft paper was funny. It was PR BS that runs 10x real time.
Speech is nice, and iflytek does fine, but speech is becoming a commodity now.
The interesting work isn't happening in either of those companies and it's not happening on those problems.
This article was written by someone who didn't know the field interviewing someone who is trying to get his budget increased via fear mongering.
Also, hilariously, the article totally neglects that more research comes from Canada and Europe (deepmind) than the US....
In fact, given both Hinton and Bengio's being Canadian and LeCun being French, I would say US really in the game in terms of producing "AI grandmasters" (assuming that's their definition of grandmaster). Krizhevsky was studying in Canada. DeepMind is European.
From both the research and industrial perspectives, China's education system has well prepared students for the field of deep learning - the rigorous / analytic nature has proven a massive benefit when dealing with such large models and datasets.
This context is the proper use of the H1-B visa. Providing technical talent for which Americans can not fill the role.
But almost all of H1-B is abused where immigrants are used to do jobs that can be done by Americans as a means of lowering wages.
This reminds me of a famous story of an early AI system, which competed against humans in a table top war game focused around ship design.
>After weeks of experimentation, and some 10,000 two-to-thirty-minute battles, Eurisko came up with what would be the winning fleet. To the humans in the tournament, the program's solution to Traveller must have seemed bizarre. Most of the contestants squandered their trillion-credit budgets on fancy weaponry, designing agile fleets of about twenty lightly armored ships, each armed with one enormous gun and numerous beam weapons.
>Eurisko, however, had judged that defense was more important than offense, that many cheap, invulnerable ships would outlast fleets consisting of a few high-priced, sophisticated vessels. There were ninety-six ships in Eurisko's fleet, most of which were slow and clumsy because of their heavy armor. Rather than arming them with a few big, expensive guns, Eurisko chose to use many small weapons.
>In any single exchange of gunfire, Eurisko would lose more ships than it destroyed, but it had plenty to spare. The first battle in the tournament was typical. During four rounds of fire, the opponent sank fifty of Eurisko's ships, but it lost nineteen -all but one-of its own. With forty-six ships left over, Eurisko won.
>Even if an enemy managed to sink all Eurisko's sitting ducks, the program had a secret weapon -a tiny, unarmed extremely agile vessel that was, Lenat wrote, "literally unhittable by any reasonable enemy ship." The usefulness of such a ship was discovered during a simulated battle in which a lifeboat remained afloat round after round, even though the rest of the ships in the fleet had been destroyed. To counter opponents using the same strategy, Eurisko designed another ship equipped with sophisticated guidance computer and a giant accelerator weapon. Its only purpose was killing enemy lifeboats.
I've always wondered if a similar strategy might work in real life.
People in pretty much every country are doing some pretty amazing stuff. A lot of it you won't hear about if you limit yourself to press releases from venture capitalists.
It's always disappointing to see science described as if it's a sportsball game between nations.
Whoever develops the tech first, we all win. You know, as long as the new sentient AI doesn't slaughter us all in our sleep that is.
>People in pretty much every country are doing some pretty amazing stuff. A lot of it you won't hear about if you limit yourself
I agree, and this is why I am unworried about the recent moves to cut government research funding in the US. There are plenty of other countries with great research institutions.
Look what happened when Bush pushed out most of the stem cell research -- research moved to Singapore and China. But the world still benefited, just not US companies.
Yeah, I thought a lot of the deep-learning revolution had come out of Canadian universities.
1 example: https://en.m.wikipedia.org/wiki/Cougaar
However, I wouldn't be surprised if the U.S. intelligence agencies want to do that just as much to Americans and the whole planet, but do it in secret.
I can state without hesitation that some of the greatest minds in Supercomputing and computational science are permanent members of the staff at these DOE research labs. These staff are selected much like the top AI researchers: academic pedigree and publications in top tier research journals.
Therefore, I think it is more a problem of application. Until government has a killer AI app, it will be hard to justify a massive investment in AI tech at scale. Contrast that with large companies who have already deployed AI tech in production...
https://motherboard.vice.com/en_us/article/the-fbi-cant-find...
Here's some insight: http://io9.gizmodo.com/academic-fraud-in-china-is-getting-ou...
http://www.thelancet.com/journals/lancet/article/PIIS0140-67...
http://www.scidev.net/global/networks/editorials/china-must-...
AI/machine learning has slightly more accountability than many other scientific fields, in that unless you provide an implementation for people to inspect or your results are easy to replicate, people generally aren't going to be particularly interested for long. There are exceptions to that principle, like AlphaGo, but the tests of AlphaGo speak for themselves. It's a very results-driven field, where you're not going to be able to get away with specious claims for long.
... You guys need to get out of deep nets and go back to your classics
“Without followers, evil cannot spread.” - Spock
The biggest proponents of Chinese power is department of defense.
The exact same language would be used if the US were competing against Russia on this, and previously was used when the USSR was the primary political competitor of the US. It has also been used to refer to competition with Japan, now and in the past.
NYTimes, 2005, vs the Japanese (and recently ascendant China):
"A global race is under way to reach the next milestone in supercomputer performance, many times the speed of today's most powerful machines. And beyond the customary rivalry in the field between the United States and Japan, there is a new entrant -- China -- eager to showcase its arrival as an economic powerhouse."
http://www.nytimes.com/2005/08/19/technology/a-new-arms-race...
---
"A Global Arms Race to Create a Superintelligent AI is Looming"
https://motherboard.vice.com/en_us/article/a-global-arms-rac...
---
"Elon Musk and Stephen Hawking Warn of Artificial Intelligence Arms Race"
http://www.newsweek.com/ai-asilomar-principles-artificial-in...
---
"How Amazon Triggered a Robot Arms Race"
https://www.bloomberg.com/news/articles/2016-06-29/how-amazo...
---
"Amazon and Google Continue Cloud Arms Race With New Data Centers"
http://fortune.com/2016/09/30/amazon-google-add-data-centers...
---
2009: "Smartphone vs. feature phone arms race heats up"
http://www.zdnet.com/article/smartphone-vs-feature-phone-arm...
---
"The Adultery Arms Race"
https://www.theatlantic.com/magazine/archive/2014/11/the-adu...
Into foreign tech economies....
China has also progressed quite handily economically and technologically, areas that don't admit to too much reliance on things that don't work. Cautious skepticism isn't the same as out-of-hand dismissal.