Please create an account to participate in the Slashdot moderation system

 


Forgot your password?
Close
typodupeerror
×
AI China

An AI Epidemiologist Sent the First Warnings of the Wuhan Virus (wired.com) 60

Posted by msmash from the marching-forward dept.
An anonymous reader shares a report: On January 9, the World Health Organization notified the public of a flu-like outbreak in China: a cluster of pneumonia cases had been reported in Wuhan, possibly from vendors' exposure to live animals at the Huanan Seafood Market. The US Centers for Disease Control and Prevention had gotten the word out a few days earlier, on January 6. But a Canadian health monitoring platform had beaten them both to the punch, sending word of the outbreak to its customers on December 31. BlueDot uses an AI-driven algorithm that scours foreign-language news reports, animal and plant disease networks, and official proclamations to give its clients advance warning to avoid danger zones like Wuhan.

Speed matters during an outbreak, and tight-lipped Chinese officials do not have a good track record of sharing information about diseases, air pollution, or natural disasters. But public health officials at WHO and the CDC have to rely on these very same health officials for their own disease monitoring. So maybe an AI can get there faster. "We know that governments may not be relied upon to provide information in a timely fashion," says Kamran Khan, BlueDot's founder and CEO. "We can pick up news of possible outbreaks, little murmurs or forums or blogs of indications of some kind of unusual events going on." Khan says the algorithm doesn't use social media postings because that data is too messy. But he does have one trick up his sleeve: access to global airline ticketing data that can help predict where and when infected residents are headed next. It correctly predicted that the virus would jump from Wuhan to Bangkok, Seoul, Taipei, and Tokyo in the days following its initial appearance.
This discussion has been archived. No new comments can be posted.

An AI Epidemiologist Sent the First Warnings of the Wuhan Virus More

An AI Epidemiologist Sent the First Warnings of the Wuhan Virus

Comments Filter:

  • The question is... how accurate is it? (Score:5, Interesting)

    by Riceballsan ( 816702 ) on Tuesday January 28, 2020 @12:51PM (#59664624)
    I always find it really annoying when the report is "Oh the AI predicted this". What they don't cover is how many predictions did the AI make? Remember the twitter account that correctly predicted the winner of every foot ball game for a full season (technically the guy posted every possible ending to all the games... set them all to hidden. and unhid the ones that actually came true... Did the AI correctly predict the danger of this virus... or did it predict thousands of virus's and this one happened to be where it predicted.

    • Yes, I was going to say the same thing. If this is reporting false positives all the time, or indicating every little flare-up of flu across the world, then it's not very useful.

      • Re:The question is... how accurate is it? (Score:4, Insightful)

        by taustin ( 171655 ) on Tuesday January 28, 2020 @01:01PM (#59664678) Homepage Journal

        Flare-ups of the flu are bigger news than this has been so far. Coronavirus (or, rather, this particular one, there are many, some of which pretty deadly) has less than 100 reported deaths so far worldwide, versus an average of 35,000 flu deaths (and 200,000 hospitalizations) in the US alone every year.

        Remember SARS? The civilization ending apocalypse that killed less than 1,000 people? This is like that.

        • That's what leads me to believe this AI may be producing a continuous stream of "warnings". The fact that it keyed on a hundred people with "flu-like symptoms" in China, would have to mean that it is also reporting every hotspot worldwide. My local community is having a flare of up the flu, and a county next to ours cancelled school for 2 days to help stop the spread in their area. So it would have to be reporting hundreds of areas daily with moderate to high cases of the flu.

          Maybe that's what people who

          • I'd be more surprised if it weren't producing a continuous stream of warnings because every year various parts of the world or different countries go through "flu season" to varying degrees. Some years it's more severe than others in terms of the symptoms or how long people stay ill and sometimes it seems like everyone catches it but it's just a mild inconvenience for a a week.

            But the AI isn't a crystal ball. It can just tell us about outbreaks that appear flu-like or that match some other pattern of how

        • Re: (Score:1)

          by rldp ( 6381096 )

          Yeah i mean people already die of Cancer so nobody should care about this AIDS thing pfft its so dumb.

          I know a guy who died of diabetes so I'm not at all concerned about getting rabies or tetanus.

          There's only enough space in my brain for one thing.

          (you sound like an anti-vaxxer)

        • >Remember SARS?

          SARS, if I remember correctly, didn't spread this fast. It took a month or two to get to these numbers. That is also assuming you trust the reporting from China that is locking information down.

          What doesn't help is that it is Flu season. So you have a respiratory disease season coupled with an outbreak of another respiratory disease.

          Not to believe the AI models or w.e. but I think at this point SARS wasn't this bad. Not that this is a apocalypse either, it's just the first outbreak during

          • Re: (Score:3)

            by taustin ( 171655 )

            >Remember SARS?

            SARS, if I remember correctly, didn't spread this fast. It took a month or two to get to these numbers.

            And? It's been a couple of weeks, or half as long as your shortest timeline, and about 1/10th the number of people have died. Either it's spreading slower, or it's less deadly (or both).

            That is also assuming you trust the reporting from China that is locking information down.

            If they're underreporting by three orders of magnitude it's still a lower death rate than the flu.

            What doesn't help is that it is Flu season. So you have a respiratory disease season coupled with an outbreak of another respiratory disease.

            Not to believe the AI models or w.e. but I think at this point SARS wasn't this bad.

            The deliberately generated hysteria, however, was.

            Not that this is a apocalypse either, it's just the first outbreak during the age of social media.

            The flu kills an average of 35,000 people a year in the US. Two years ago, it killed about 80,000, over twice as many (with four times as many hospitalizations at 800,000). Two y

            • Re: (Score:3, Insightful)

              by penandpaper ( 2463226 )

              >Either it's spreading slower, or it's less deadly (or both).

              It's not spreading slower, even with official numbers. Again, that assumes you trust the numbers China is putting out. Given their response that does not add up.

              >it's still a lower death rate than the flu.

              Ok? What does that have to do with this disease and its communicability? We haven't seen this disease before and we are still in the initial phase. If there is a 2 week gestation period in infected hosts then expect mortality to rise as the

        • Flare-ups of the flu are bigger news than this has been so far. Coronavirus (or, rather, this particular one, there are many, some of which pretty deadly) has less than 100 reported deaths so far worldwide, versus an average of 35,000 flu deaths (and 200,000 hospitalizations) in the US alone every year.

          Remember SARS? The civilization ending apocalypse that killed less than 1,000 people? This is like that.

          My wife keeps seeing these pics on the Faceberg of bodies stacked up in China, asking me if it's real. Someone has even created this video of a Chinese doctor in scrub gear and a mask (conveniently hiding his face from authorities) supposedly begging for help from Beijing while a crying nurse (also in mask and full scrub gear) desperately asks the doctor what she should do while the bodies pile up outside. It was filmed in some generic office somewhere, an obvious hoax. Not a piece of medical equipment to b

          • Re: (Score:1)

            by rldp ( 6381096 )

            Yeah they quarantine cities of 60 million all the time it's no big deal! Trust the chinese government why dont you!

            • Or it's just an excuse for an authoritarian government to crack down on the population and flex its might while giving the illusion of safety and benevolence. There's an old saying about not letting a good tragedy go to waste.

              Do I believe that this is the real explanation for what's going on? No, of course not, but it's the type of thing that is still plausible enough for me to peg it somewhere in the 2% - 5% likely range. I'd say it's more credible than the theories about it being a bioweapon that escap

          • >What's giving the power to media to inflate this stuff is the novelty of it:

            The novelty and the fact that China is cracking down on information which allows speculation to run rampant. All we can do is sit back and hope that China isn't lying too much.

        • How many cities were quarantined during the SARS epidemic? Or MERS? Or zika? Or the measles? I wasn't concerned about those. They were just one more trial run. This one is different. Hopefully it's one more trial run as well but I wouldn't be surprised if it's already 10x worse than reported. And, is it less deadly? We don't know yet. The reported data shows things are acceptably bad ... but what if the CCP is lying. This, along with the pig ebola crisis, can bring down the CCP; making it one more in a lo

        • It is way too early to be relaxed abut this.The reported death rate of 3% is calculated by dividing known infections by known deaths. Trouble is that both of those numbers are very rubbery.

          If the numbers are accurate, the situation is very worrying because the number of known infections includes people who will die from the disease but have not yet done so. The eventual death rate could be a very scary 10%.

          However, the infection number could very well be low. If most sufferers just exhibited mild symptom

        • What you're not seeing was the huge amount of work health organizations around the world to keep sicknesses like SARS (and now Wuhan coronavirus) contained. SARS didn't kill many people because of the monumental amount of effort to prevent its spread. There generally isn't the same effort with the flu because, unlike SARS, which had about a 10% mortality rate, the mortality rate for influenza is significantly below 1%. The Wuhan coronavirus is about 3% mortality, for reference -- significantly higher than t

    • I think it's common knowledge among developers that AI is not even close to a moon shot. In this case, when used together with human intelligence, it's a very powerful tool that allows a researcher (or researchers) to bring up a potential list of threats instead of having to manually sifting through a ton of data. It's a tool for practical data analysis.

      Here's a Ted Talk about this very concept.

      https://www.youtube.com/watch?... [youtube.com]

      • Kind of ironic in a way that it needs a society that leaks information like a sieve in order to work. If AI doesn't know anything about anyone then how can it help?

      • Right but what I mean is, the rating system... is obviously the current form of AI doesn't tell us the greatest threat, it highlights the top candidates, and lets humans focus on those vectors... Still though to judge the quality of an AI... I'd assume the amount narrowed down is the level of quality of it. IE turning a pile of 100,000 candidates to 10, is great... turning a pile of 100,000 candidates to 99,999 is technically the same thing, but obviously almost useless. The thing with articles is there's n

    • What they don't cover is how many predictions did the AI make?

      There you go, taking all the fun out of everything. It's Just as accurate as The Prophecies of Nostradamus or Miss Cleo, you just have to only talk weirdly enough and in-general enough and then count your hits and ignore the misses. Hell, even the MISSES are positives if you count right.

      "Ok, person in the audience that I'm talking to, is your name ... SLARTIBARTFAST? No? OK, do you know someone by that name? No? OK, is there someone AROUND you with that name? No? Is there someone is your ROW w

  • Yeah but how many times did it cry "Wolf!"? (Score:3)

    by kimgkimg ( 957949 ) on Tuesday January 28, 2020 @12:53PM (#59664634)
    How many other times has this AI "alerted" to something? This may have been one notification among thousands that it has generated which caused people to largely ignore it's output. Need more facts here.

    • The answer: EVERY time.

      In 2019, 8,200 people died from the flu, out of 140,000 confirmed cases--just in the United States.
      https://abcnews.go.com/Health/... [go.com]

      So far, Coronavirus has killed 106, worldwide, out of 4,500 confirmed cases.
      https://www.bbc.com/news/world... [bbc.com]

      Why don't we see dramatic headlines and quarantines for the flu, which is a much more deadly disease? Well, it's because we all already know about the flu, it's not a NEW scary thing.

  • My local news site has this headline: "Two local residents test negative for coronavirus".

    This is not news! Why do they do this?!?!

    • Re:Fake News (Score:4, Informative)

      by rldp ( 6381096 ) on Tuesday January 28, 2020 @01:12PM (#59664722)

      Because there was a previous story about two local resident being tested for it, and this is a follow up.

    • Re: (Score:2)

      by nagora ( 177841 )

      My local news site has this headline: "Two local residents test negative for coronavirus".

      This is not news! Why do they do this?!?!

      Well, I assume that if they were tested then there was some suspicion they might have the virus.

    • This is not news! Why do they do this?!?!

      So you'll click on it, sperg out on how stupid it is to a bunch of other people, who click on it to gape at how stupid it is, and they tell two friends, and they tell two friends, and so on, and so on..
      You didn't complete the ritual like a good little clone, though, and post the link. They won't give you your treat if you don't post the link! :p

    • Because they'd already gotten their dicks hard of the idea of having a big story and made a giant write up with all kinds of details. The test comes back negative, but they can still just dump all of those same details into the story to fill out a few minutes of air time or the rest of the page. News is a product like everything else, so it shouldn't be surprising that the same advertising techniques that we all hate are used to sell it.

  • Hmm... (Score:1)

    by Rei ( 128717 )

    Let me know if it predicts an outbreak of Lyme disease in China concurrent with the current corona viral outbreak. Cases of corona and Lyme [cloudinary.com] en masse in a short period of time can cause confusion, unresponsiveness, disorientation, shallow breathing, unconsciousness, and in severe cases, coma or even death [medicalnewstoday.com].

  • Could have been fake news but I was reading the virus came from bats, specifically carnivorous bats. It's interesting that western cultures that have mostly stuck with eating herbivores don't get these types of diseases.

  • China is still holding on to the virus and will not share it with other scientists. Of course, it no longer matters. Now, it is spreading around the world. The problem is, what is the REAL Mortality and Ro on this? Yeah, Chinese government has given up some stats, but the problem is, that their stats does not match what is happening Wuhan. In particular, when they claimed that only 40 ppl had died, multiple Wuhan hospitals had dead bodies lining their hallways. That hints that the death rate was in the 100s

    • Re: (Score:2)

      by Rei ( 128717 )

      "Holding onto it"? There are people with the disease outside of China; the Chinese government couldn't "hold onto it" if they wanted to. It's not a friggin' panda. ;)

      • But we already have Panda anti-virus.

        ba dum tish.

        I'll see myself out.

      • China is still holding on to the virus and will not share it with other scientists. Of course, it no longer matters. Now, it is spreading around the world.

        You do understand that this was spotted in China in MID DECEMBER. Almost 1.5 months ago. And that is what CHina ACKNOWLEDGES. The fact that it was so fast isolated, hints that it occurred MUCH EARLIER. And China STILL has not shared the virus directly. They did not give a vial of it to any other nation. Not to WHO. Not to CDC. Not to Russia (well, that we know of).

    • China is still holding on to the virus and will not share it with other scientists. Of course, it no longer matters. Now, it is spreading around the world. The problem is, what is the REAL Mortality and Ro on this? Yeah, Chinese government has given up some stats, but the problem is, that their stats does not match what is happening Wuhan. In particular, when they claimed that only 40 ppl had died, multiple Wuhan hospitals had dead bodies lining their hallways. That hints that the death rate was in the 100s, and possibly 1000s, when the government was claiming 40.

      Be wary of those pics and vids. Some of them are hoaxes, and some of them are misrepresented. The standard "if you found it on the Internet..." rules apply here.

    • Yet more anti-china lying nonsense from WindBourne. You know they released the genome very early on, and that multiple countries are already using that info to work on vaccines. Why do you even bother anymore? Even common sense shows you are wrong. How are other countries testing people without knowing what the virus is? You're an idiot as well as a liar.
      • Windbourne:

        China is still holding on to the virus and will not share it with other scientists. Of course, it no longer matters.

        Caffeinated Bacon i.e paid chinese troll:

        Yet more anti-china lying nonsense from WindBourne. You know they released the genome very early on, and that multiple countries are already using that info to work on vaccines.

        First off, having a SEQUENCE is not the same as releasing the genome (which would involve giving SAMPLES).
        Secondly, nobody makes vaccines from simple sequences.

        Like always, CB (also known as Crimson Tsunami/red wave) is doing his paid job of trying to defend China. As I said, China has NOT shared the virus with any other nation. Since CB has called ME that liar, show us which nations did China give sample to? Here, CB, I will help you out. Here is a l [google.com]

        • Your ignorance is unbelievable WindBourne.
          It's always you who is the liar. You have still never ever shown a lie.
          People are making vaccines from sequences without the actual virus.
          Australia isolated and grew samples in less than a week of having access to an infected patient.
          Your ignorance is expected though. You still believe medical technology is still the same as when you designed a website for the cdc and thought you were a doctor.

        • Although virus samples can also be used to validate molecular diagnostic tests, most labs have moved away from using whole viruses in favour of synthetically producing parts of the virus from partial genomes, says Mackay.

          Thanks for the link WindBourne. It's always amusing when your very own links contradict your claims.
          (LOL, you are a complete joke. Read your own links once in a while.)
          https://amp.abc.net.au/article... [abc.net.au] Educate yourself with that link too. It shows it only took a handful of days after receiving a sample to grow it...
          You are completely clueless, as always.

    • Australia is already sharing their lab grown samples anyway.
      But you probably still think medical science and technology is still the same as it was decades ago. (Like last time this topic showed up and you made a complete fool of yourself.)

  • Postmortem opportunist? (Score:3)

    by hackingbear ( 988354 ) on Tuesday January 28, 2020 @01:29PM (#59664808)

    There were news reporting of new type of lung infections back in mid-December. And by Dec 31 [wikipedia.org], the Chinese authority already publicly announced the initial outbreaks and reported to WTO. (However, nobody seemed to think it could spread this fast and nobody wanted to disrupt the massive holiday travel.)

    Rather than a master fortune teller, this "health monitoring platform" seems more like postmortem opportunist monitoring WHO announcements and attempting at PR pitch.

  • Whilst the common Slashdot rabble love to blather on about "false positives" or other irrelevant details; realistically these lower class individuals are so far out of their scope it's obvious why they can't comprehend the practical applications for such a tool.

    Consider if you were someone like an executive officer employed by a large corporation. You'd be scheduling international appointments related to conducting your business duties. Given a tool which provided such a warning, it would be your duty to us

  • Or is it just noticing a pattern and reporting it? That logic has been around for decades.

    • Re: (Score:3, Informative)

      by olfdag_kerfunke ( 6260520 )

      Yes, it's called "AI" but honestly it's just an algorithm. Whether it uses machine learning and "neural networks" or a series of branches/switch statements and dynamic weighting, the results are all referred to by this generic term. The sole aspect which might justify the application of the term is non-linear dynamical behavior that creates apparently chaotic but statistically orderly output. This type of varied "unpredictable" output resembles what you'd expect to see from an "intelligence".

      The real proble

    • It isn't even doing that. It is the humans that report it: "Once the automated data-sifting is complete, human analysis takes over, Khan says. Epidemiologists check that the conclusions make sense from a scientific standpoint, and then a report is sent to government, business, and public health clients."

      • It isn't always possible to remove outliers or bad data via autonomous algorithmic procedures. The "AI" in that sense acts as a filter which mostly lacks the genuine intelligence required to adequately assess each case. In cases where you can accurately filter away irrelevant data, the workload is reduced. This is a success from an efficiency standpoint.
  • ..so, what's your point?
    Maybe they should have an AI cook pasta for us, because it apparently is good at seeing what sticks to the wall.
  • Soon the AI will "learn" to avoid hits to its social credit score and remain as tight lipped as the Doctors.
  • It must be the time when everything is "AI". This could just as well have been built with out any "AI"....
  • Is the real purpose of early prediction on large outbreaks like this so they can short the market days ahead of time? Muahhaha.

  • US turn (Score:3)

    by sdinfoserv ( 1793266 ) on Tuesday January 28, 2020 @03:54PM (#59665416)
    That's all right, the US will more than eatup any potential benefit as Lord Trumpkin rolls back any environmental regulation, MPG requirements, and renewal subsidies while chanting alternating chorus' of "dig baby dig" and "lock her up!"
  • At this point, AI 'predictions' are as vague as Nostradamus and his apocrypha. Why do people put so much dependence on this unproven technology in life-death situations?
  • No it was never AI-driven, what it is - is a massive pattern recognition engine - a true AI would have created a vacine by now.

Slashdot Top Deals

FORTRAN is not a flower but a weed -- it is hardy, occasionally blooms, and grows in every computer. -- A.J. Perlis

Close