Slashdot Mirror
Is China Outsmarting America in AI? (nytimes.com)
An anonymous reader shares an NYTimes article: Beijing is backing its artificial intelligence push with vast sums of money. Having already spent billions on research programs, China is readying a new multibillion-dollar initiative to fund moonshot projects, start-ups and academic research (Editor's note: the link could be paywalled; alternative source), all with the aim of growing China's A.I. capabilities, according to two professors who consulted with the government on the plan. China's private companies are pushing deeply into the field as well, though the line between government and private in China sometimes blurs. Baidu -- often called the Google of China and a pioneer in artificial-intelligence-related fields, like speech recognition -- this year opened a joint company-government laboratory partly run by academics who once worked on research into Chinese military robots. China is spending more just as the United States cuts back. This past week, the Trump administration released a proposed budget that would slash funding for a variety of government agencies that have traditionally backed artificial intelligence research.
How much money is Google putting into AI research? Amazon? Apple? IBM? Others? How successful are they compared to the Chinese government's efforts?
How many products or services do people use which rely on U.S. company's AI efforts and how many which rely on Chinese created efforts?
The idea that the only comparison is between Chinese government funding and U.S. government funding is ridiculous. The private companies in the U.S. working on AI are the ones actually accomplishing things nowadays and announcing another government 5-year plan for China to win some sort of AI race isn't going to change that.
The party of stupid and the party of evil get together and do something both stupid and evil, then call it bipartisan.
The tech giants -- Microsoft, Amazon, Google, etc. -- did not depend on an infusion of cash from governments to become leaders. Although there are likely exceptions, governments tend to do poorly when picking winners and losers. My guess is that China's major gains in A.I. will occur from spycraft, in other words, stealing the intellectual property from companies in the West.
Well, how good are their AIs at Go?
Great minds think alike; fools seldom differ.
In all fairness, I believe you are conflating AI with AGI.
Artificial intelligence has been dramatically improving at a staggering pace and is focused on singular tasks. Artificial "General" Intelligence is still nowhere to be seen on the current technological horizon, and would allow a computer to be amazing at any number of tasks.
That has not stopped writers, who earned their IT chops in a movie theater, from repeatedly suggesting that any AI that can drive a car or beat a World Master Go player is just steps away from initiating a discussion about its personal dreams and ambitions.
The problem with the USA is that it now eats too much of its own dog food.
Societal pressure to have students do well in tests in small monolithic cultures, like South Korea and Finland, do not necessarily translate perfectly into larger societies. The American educational system is actually praised across much of Asia, because it focuses more on creativity rather than regurgitation. Slower students do tend to find it easier to drop out of US schools, but as compensation, the best students usually find unique opportunities that allow them to excel in ways that Asian countries don't usually allow.
China is never going to have the world's biggest economy until it ditches its totalitarian system. There is just too much you cannot do.
Yes but Chinese companies like Baidu, TenCent and Alibaba are comparable to Google, Facebook, and Amazon and are very competitive in AI research and accomplishments. China's home grown supercomputers are the fastest in the world and the Chinese are pursuing multiple avenues of development and improvement in software and hardware. Assuming that the US will continue leading in computer hardware and software without a national push to improve is naive.
The problem is that you misunderstand what "artificial intelligence" means. John McCarthy, the person who coined the term in 1956, defined it as making machines "behave in ways that would be called intelligent if a human were so behaving." It explicitly does not require machines to be sentient. It does not require the machine to follow the same "thought processes" that a human would when performing that action. When a human plays chess, or translates a document into a different language, or drives a car down a street while obeying traffic laws and not hitting anything, everyone agrees they are displaying intelligence. Therefore when a computer does the same thing, that counts as artificial intelligence. That's been the standard definition of the term for the last 60 years.
If you want a computer to be sentient, that's something completely different. We're nowhere near being able to do that. We aren't even sure how to define what that would mean. But that isn't what the term "artificial intelligence" means.
"I'm too busy to research this and form an educated opinion, but I do have time to tell everyone my uninformed opinion."
Every year we keep stupidly claiming that AI is just around the corner. Every year we are disappointed.
The truth is we have tricked ourselves. The rapid pace of Moore's law (computing power keeps doubling) has created incredible simulations. But paintings and statues do NOT spontaneously come alive, no matter how accurately they simulate a person. Neither do computer chips.
Indeed. But most people fall for cargo-cult, i.e. they cannot distinguish things that look similar on the outside. Apparently, actually understanding how something works requires advanced human intelligence, and it seems only something like 10% of the population has that. Hence the stupid claims.
There is a fundamental difference between real AI and what computer chips can do. The ability of computer chips to parse written, audio, and visual information is amazing, and keeps growing but it is NOT real AI and will never be.
While I sort-of agree at this time, there is a small, residual change that the physicalists are right and that humans are only advanced automatons. But it does indeed not look like it at all. A lot of research has not produced any credible theory how general intelligence (true/strong AI) could be created and it clearly is not a question of computing power. For example, the only thing we have that approaches strong AI in still a very limited field is automated theorem proving. But this one gets bogged down in complexity so early, that a smart human being can do things that a computer the size of the whole universe cannot do.
And there is the elephant in the room, constantly ignored by Neuro-"science": Consciousness. Observable only together with intelligence, and nobody has any idea what it is or how it works. In fact, current Physics does not allow it, as there is no mechanism for it. Saying it is an "emergent property of complexity" is just bullshit and akin to claiming it is "magic". Now, is two things are getting observed only together, a sound assumption is that they are facet of the same thing. Yet that also gets ignored by those that predict strong AI "anytime soon".
Computers will shortly be able to accept input via camera and microphone as accurately as they get it from a keyboard or mouse. That is not real AI. Nor is the amazingly complex search functions and databases we have created.
They are useful, and worth investing in, but more money has been wasted on them than is appropriate.
The term usually used these days is "weak AI". Weak AI was historically called "automation" and it is the "AI" without intelligence.
Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
Much of the current problem started with WWII.
.
The USA was lucky, not great.
Name any other large, well populated, educated, industrialised nation with large amounts of natural resources that was not bombed during WWII. During the 1940-1970s the USA was able to build on what they already had, the rest of the world was effectively rebuilding roads, rail. schools, hospitals, and all the other infrastructure required. More to the point, they were able to build and sell the things the rest of the world needed.
During the 1950s the USA account for over 50% of the entire worlds GDP, today its about 20%.
The world is no longer reliant on the US, sure it impacts all the world, but so does China and the EU.
The US is 4% of the worlds population, so 96% of the worlds population and 80% of world trade are not US based.
China can (and will) surpass the USA, so will India and Brazil, may not happen in my life time, but it will happen, and I am not so sure the US is capable to accepting that cultural shock. I think high up in some sectors of the US government they understand this which is why they are meddling in the politics of Asian countries, they don't want as Asian Trading Bloc because that is 60% of the worlds population, and the area of greatest economic growth potential. Growth potential in the USA is almost nil, its a saturated market.
And while Trump et al keeps shouting USA USA USA and USA first, the rest of the world keeps on improving, and putting the USA further and further down the ladder. For example, the world is not longer reliant on Boeing, there is Airbus, and China is getting into the act too. ARM is doing well, Its British not US. Samsung is doing well, again not US. And there are thousands of examples where non-US products are better than US ones.
Its not like the US has failed, it more like the rest of the world has grown up and is no longer dependant. And because of that, the natural progression is that the US will fall behind in many fields
The USA is 14th for Reading
The USA is 25th for Maths
If you look at the data, the difference comes down to demographics, urban vs rural and race.
The top scorers in PISA are highly urbanised and racially homogeneous.
If you look at data from American Whites and Asians in the cities, they compare very well.
There is less data for blacks and rural populations, but I think those demographics in America would also compare favourably to their fellows in other OECD countries.
So you'd add "not running a program" to the list of requirements in your personal definition, next to "not a computer chip". Why not just say "must be a human brain"? It's no less arbitrary, and no less unrelated to actual research.
Do you also define "learning" as requiring biology, since your meaning of the word apparently excludes a dozen existing fields?
Why would anyone engrave "Elbereth"?
China is already outdoing the US in a whole bunch of things, but international press do not cover this so lots of people don't really know what's going on there.
This is an understandable, often overlooked problem that not many people stop to think about.
The way press works for international coverage, for several countries, is to only publish a limited range of stories that overcomes the cultural/language barrier, when not ultimately going only for eye grabbing content.
And that's fine, because really, who's got the time and attention to know everything that's happening all around the world?
It's just naive and kinda dangerous to build an image of a country and it's industry based on the very limited information you get from main channels.
It's why even nowadays we still have so many people with this image that products coming from China are all shoddily made or clones of american/european products, when in fact not only China controls the vast majority of production for most electronics we use in a daily basis, several design decisions and technology advances also happen there.
It's nothing magical really... when you have a single country taking care of a huge percentage of worldwide production and manufacturing of tech related products for over a decade, of course they'll start developing their own products from start to finish. Think about what your own country would do in a similar situation.
People who have been paying attention for one reason or another to chinese branded smartphones, tablets, laptops and several other lines of products will know that they are fast becoming indistinguishable from high end lines of american and european brands. And particularly for their own market, there is no culural barrier to overcome. Technologies that are highly related to culture like AI (because recent advancements have been going around speech recognition and such) are bound to evolve in a different way.
Who's the leader in end-consumer quadcopters right now? DJI, indisputably, right? You know what DJI stands for? Dà-Jing Innovations Science and Technology Co. It's a Shenzhen based and born company. There's a whole bunch of tech crammed in those drones that were developed by the company... tech for obstacle avoidance, 3d tech for hand gesture recognition, radars and sensors.
Some people might not know, but Lenovo is also a chinese company. Yes, the one that now owns the staple of business laptops, the Thinkpad line. The same company that owns Motorola.
There's a whole bunch of cases like those in the tech industry.
Not to mention how chinese companies have been buying left and right a whole bunch of hotel businesses, movie studios and other companies people have no idea about:
http://fortune.com/2016/03/18/...
Sure, a whole bunch of tech that several chinese companies made in the past were straight rip offs of designs from US, europeran and japanese based companies, but this has changed in later years. And the further you go into several tech devices, the more you understand how much of the technology behind them are really not coming from a single US brand.
High end technology for all sorts of displays nowadays have a majority made in South Korea (LG and Samsung). Central parts of cameras of all shapes and sizes, including smartphone cameras, mostly comes from Sony, a japanese company. Samsung also dominates when it comes to technology related to storage (memory chips and whatnot), but that market is a bit more balanced. CPUs, GPUs and SOCs are still mostly developed by american and british companies (Qualcomm, Intel nVidia), but that doesn't mean they don't have chinese or asian competition (Samsung, Mediatek, Allwinner). More importantly though is that in several areas of technology, if a chinese company isn't already there among the top businesses involved, there's likely to be one encroaching.
So yeah, I don't know if chinese companie
China can (and will) surpass the USA, so will India and Brazil, may not happen in my life time,
China and India will, but not Brazil. The US has a larger population than Brazil, and that gap will increase (the fertility rates are the same, and more people want to immigrate to the US). The social and economic gaps between the US and Brazil will narrow, but probably not close entirely.
The US may only be 4% of the world's population, but that still makes it the third most populated country after China and India.
What governments invest in and what private businesses invest in are usually two different categories of research - as far as I know, that is the way it has always been. Private companies typically invest in monetising research results, or do research on things that can be made into a saleable product - that is after all the primary purpose of most businesses. But where does that research come from? Normally from government funded sources - certainly this is the case outside the US, where most universities are publicly funded. The point to take home is that the fundamental research, which to the un-initiated can seems like mere daydreaming and waste of time and money, is crucially important for the high-tech businesses of tomorrow, who are going to monetise the results that flow from fundamental research.
Just look back at the decades of research into how inherited traits are actually transmitted - it started around 1830 with Gregor Mendel, and it was not until the late 20th century that we got so close that businesses could smell the money, but now DNA research is one of the most important, fastest growing industries. So that's 170 years of curiosity driven, fundamental research and 20 years of private businesses investing in the same. Public funding of research is what drives this kind of research - because companies in general only think about the next fiscal year and the bottom line.
I'll do that once the EU countries give up their separate UN votes :)