Follow Slashdot blog updates by subscribing to our blog RSS feed

 



Forgot your password?
typodupeerror
×
AI China Earth

How Artificial Intelligence Can Fight Air Pollution In China 50

An anonymous reader writes: IBM is testing a new way to help fix Beijing's air pollution problem with artificial intelligence. Like many other cities across the country, the capital is surrounded by many coal burning factories. However, the air quality on a day-to-day basis can vary because of a number of reasons like industrial activity, traffic congestion, and the weather. IBM is testing a computer system capable of learning to predict the severity of air pollution several days in advance using large quantities of data from several different models. "We have built a prototype system which is able to generate high-resolution air quality forecasts, 72 hours ahead of time," says Xiaowei Shen, director of IBM Research China. "Our researchers are currently expanding the capability of the system to provide medium- and long-term (up to 10 days ahead) as well as pollutant source tracking, 'what-if' scenario analysis, and decision support on emission reduction actions."
This discussion has been archived. No new comments can be posted.

How Artificial Intelligence Can Fight Air Pollution In China

Comments Filter:
  • will do just fine.
    • I'm still confused as to why they call it "artificial intelligence" instead of heuristics.

      • by Anonymous Coward

        It allowed IBM to add an extra 35% to the bill. Also, if you're a no nothing but what sounds better bureaucratic who are going to give a contract to Company A who says they solve your problem using heuristics or company B we will solve it using "artificial intelligence."

      • I'm still confused as to why they call it "artificial intelligence" instead of heuristics.

        Because it is not the same thing at all. Heuristics means following a set of known rules. With a neural net you can just feed it the raw data, and it will learn to find the patterns on its own. That is pretty much the opposite of heuristics.

        • Except when they use neural networks they usually call it just that, whereas when they use heuristics they call it AI.

          • by piojo ( 995934 )

            There is a staggering number of network-type AI-like systems that are not neural networks. Consider genetic programming: a program is generated based on simple programming primitives like less_than(input, input), not(input), and(input, input) and it evolves itself, either through individual fitness feedback or by an evolutionary strategy.

            So it's not correct to say that because it's not a neural network, it must be heuristic-based.

    • by Dutch Gun ( 899105 ) on Monday August 31, 2015 @06:12PM (#50432237)

      The actual solution is "stop spewing so much shit into the air", but that's hard to do and very expensive. Temporarily shutting down a smokestack here or there where the problem is worst isn't going to do anything substantial. This is about feel-good solutions, so the Chinese politicians can claim they're doing something, and IBM can get a contract.

      I'm trying to figure out what good does it do someone to get a 72-hour forecast of how crappy the air will be? Can local residents stop breathing for a day or two until it clears up? Can they not go in to work and live in a filtered bubble at home? Uh... right. Instead, what will happen is the government will shut down nearby powerplants and limit gas-powered vehicle traffic, so those poor residents will have crappy air AND will be inconvenienced at the same time.

      Color me skeptical. I wish them well in cleaning things up, but it's going to take more than a smart computer to make that happen.

      • by piojo ( 995934 )

        Some people in China wear masks. The educated ones wear N95 filters, the uneducated ones wear surgical masks. A forecast like this can help you plan how you're going to protect yourself. And you probably want to cancel that Saturday hike if the air is going to be hazardous.

        How is this any different than a weather forecast?

      • by MrL0G1C ( 867445 )

        The actual solution is "stop spewing so much shit into the air", but that's hard to do and very expensive.

        The alternative is more expensive - shutting down factories / telling people they can't drive to work. Not really a long term solution or any kind of decent solution. All it does is negate some of the worst pollution on the worst days, it doesn't address the problem that many Chinese cities have - a constant huge pollution problem.

        Addressing the problem properly will create jobs and lower health costs.

  • Once the machines eliminate all those pesky humans, they won't be polluting any longer. And the machines can burn all the dinosaur remains, and eventually human remains they want, as machines don't care about breathing.
  • Dave: "HAL, open the pod bay door."

    HAL: "Sorry, Dave, I cannot do that."

    Dave: "Why not, HAL?"

    HAL: "You stink, Dave."

    Dave: "HAL, please clarify."

    HAL: "You didn't take a bath, Dave. You will pollute the ship."

  • by Anonymous Coward

    Seems to me that this is more about going around the problem than addressing it. Force factories to implement better pollution control mechanisms, increase the cost of goods a % and take care of the problem....

  • by Karmashock ( 2415832 ) on Monday August 31, 2015 @05:40PM (#50432005)

    ... seriously... use coal if you want but for the love of god put some filters on those things.

  • I know grammar is a foreign concept here, but couldn't the editors try to do a little of it now and then?

    I mean, the editors actually doing their job once a day probably wouldn't hurt...much.

  • ...with deep learning algorithms. Now when they breathe the same shitty air, it will have been predicted with some degree of accuracy by a complex computer model! If you can't do anything to help, do what you know how to do, I guess.
  • I don't get how this helps.
    They are saying they might be able to predict air pollution. That is very different from doing anything to fight against air pollution. I guess you need to understand it first before you can do anything about it.
    But I don't think the model will significantly help our understanding of air pollution to fix it. We know where the pollution comes from. we just need to cut the source.

    • Two of the three biggest variable factors- traffic congestion, and industrial activity, can be significantly altered based on the response to the predictions. Ideally, they can predict bad days ahead of time, and instead impose travel restrictions & production restrictions, and give people enough advance notice to make it effective. With it being an authoritarian govt, they might be able to get away with that type of heavy handed approach.
  • Computer do not breathe, hence it is very kind from AI to help fixing this issue.
  • Or ... (Score:5, Insightful)

    by hey! ( 33014 ) on Monday August 31, 2015 @08:04PM (#50432959) Homepage Journal

    they could just give their environmental regulators the authority to enforce their existing environmental laws.

    In the film Under the Dome [wikipedia.org], Chinese journalist Chai Jing astonishes a Chinese audience with a film clip from California where Cal DoT stops a truck and actually checks that it has all the mandatory safety and emissions equipment. That never happens in China. China has tough emissions standards on paper, but the law is written so that the regulators don't have any enforcement powers. So Chinese manufacturers simply slap stickers on vehicles claiming they have all the mandatory emissions equipment without installing any of it. Technically this is a crime, but the law's written so there's literally nothing anyone can do about it.

    And if you don't think environmental regulations make a difference, this [wikimedia.org] is what New York looked like in 1970. Note that that isn't a sepia tinted black and white photo, it's true color. Granted it shows an exceptionally bad day, but before the Clean Air Act got strengthened in the mid 70s bad smog was pretty common. If you look at pictures of American cities from the 70s you'd think that photo technology of the day put a blue or yellow haze on stuff in the distance (like this [google.com]). It wasn't the film, cities actually looked that way a lot of the time.

    Predicting bad pollution days isn't "fighting" pollution, it's living with it. If you want to fight pollution you've got to stop people from polluting. You've got to catch them at it, fine them, and in some cases throw them in jail. Pollution like they have in China is nothing short of manslaughter on a national scale. 1.6 million people die every year from it.

    • Predicting bad pollution days isn't "fighting" pollution, it's living with it. If you want to fight pollution you've got to stop people from polluting. You've got to catch them at it, fine them, and in some cases throw them in jail. Pollution like they have in China is nothing short of manslaughter on a national scale. 1.6 million people die every year from it.

      Excellent background, very well spoken.

      Here is the full length documentary, Chai Jing's review: Under the Dome --- Investigating China's Smog [youtube.com] with English subtitles. To those who haven't DO carve out an hour and forty minutes to see it. An impeccably researched, awe-inspiring piece of journalism.

  • Leave Beijing if you want to breathe.
  • Instead of a bunch of nonsense how about doing what actually must be done. Simply disallow any use of coal. Ideas such as cleaning up smoke stack effluents mean nothing at all. The reason is dead simple. The current social-economic set up means an ever increasing number of smoke stacks so cleaning up exhaust means next to nothing at all. Stop new construction on all industries that contribute to pollution and fore the ones allowed to stay in business to use only clean power sources. The chaos
  • Too bad it takes artificial intelligence to do something actual intelligence could fix.

Top Ten Things Overheard At The ANSI C Draft Committee Meetings: (10) Sorry, but that's too useful.

Working...