News Alibaba Claims AI Can Identify Coronavirus Patients With 96% Accuracy

[Admin]

Alibaba reportedly said its artificial intelligence can detect new Coronavirus infections with 96% accuracy.

Alibaba Claims AI Can Identify Coronavirus Patients With 96% Accuracy : Read more

 

[derekullo]

But can Alibaba see why kids love the taste of Cinnamon Toast Crunch?

https://knowyourmeme.com/memes/why-do-kids-love-the-taste-of-cinnamon-toast-crunch

 

[Rdslw]

This metric is MATRIX BASED. true positive , false positive , false negative true negative.
Are the 96% true positive then where goes the last 4 % ? If it lets go of infected people, its not good.
Also another thing I've seen such companies to ignore false positives in such metrics.
And it uses CT so throughput is to low for china...

 

[escksu]

96% is just not good enough. At 10,000 thats 400 pple..... if 100 of then is false negative, its gg.....

 

[Samat]

This metric is MATRIX BASED. true positive , false positive , false negative true negative.
Are the 96% true positive then where goes the last 4 % ? If it lets go of infected people, its not good.
Also another thing I've seen such companies to ignore false positives in such metrics.
And it uses CT so throughput is to low for china...

Current testing methology is only 40% accurate with 30% chance for false positive or negative. So 96% accuracy would be a huge inprovement.
But definitely some accurate method, without having to take a CT scan, is needed.

 

[Chung Leong]

96% is just not good enough. At 10,000 thats 400 pple..... if 100 of then is false negative, its gg.....

It's good enough as long as one undetected patient does not transmit the decease to more than 25 others. Reducing the transmission rate to below 1 is the key to halting the epidemic. Of course, from the point of view of individual patients, 96% accuracy is decidedly unsatisfactory.

 

[spiketheaardvark]

I'm curious how many of these patients are routinely getting CT scans. CT scan time isn't cheap and in the middle of an out break the machines are going to swamped. I also doubt that it will be useful for asymptomatic people. I can see the usefulness of running the analysis on people that are getting CT scans as standard of care but not much else. And then with more standard testing for follow up.

Current testing methology is only 40% accurate with 30% chance for false positive or negative. So 96% accuracy would be a huge inprovement.
But definitely some accurate method, without having to take a CT scan, is needed.

I really hope the RT-pcr test is better than that. I don't have to live up medical diagnose standards in a research lab, but if I had a pcr with that detection rate I'd consider it very broken. I would be very surprised if the CDC pushed out a test that poor. Where you take the sample from and how you handle it though are going to be very important, and it's possible the application of a good test is wanting. Our best bet probably a combination of PCR and serological test. Both are fast, relatively cheap, routinely done in clinical labs, and easy to do in very large numbers.

 

[bit_user]

Are the 96% true positive then where goes the last 4 % ? If it lets go of infected people, its not good.

They can usually bias the results one way or another. As you point out, you probably want to minimize false negatives, depending on the availability of quarantine facilities.

Edit: I didn't find the research cited by the article, but here's a similar attempt by another group:

[2002.09334] Deep Learning System to Screen Coronavirus Disease 2019 Pneumonia

Also another thing I've seen such companies to ignore false positives in such metrics.

Those can be addressed by follow-up testing with another methodology.

And it uses CT so throughput is to low for china...

Maybe, but maybe not. China is throwing a lot of resources at this, and it sounds like the test is only for those with significant symptoms.

Edit: It seems the patients should've already been getting CT scans. From the article Tom's cited:

The Chinese National Health Commission on Feb. 5 widened the criteria for the diagnosis of new infections, adding CT scan results on top of the previous nucleic acid test method, to ensure patients with clinical symptoms will receive standard treatment as soon as possible.

 

[bit_user]

Current testing methology is only 40% accurate with 30% chance for false positive or negative.

On the news, I heard someone (presumably in quarantine) tested negative 6 times and they finally got a true positive, on the 7th.

There are also reports of a couple people testing negative, being released, and then testing positive again. It's unlikely they got re-infected, so that points to more accuracy problems with the test (or how it was administered).

 

[bit_user]

I also doubt that it will be useful for asymptomatic people.

Maybe not for completely asymptomatic subjects, but here's an attempt at a cheap, unobtrusive method:

[2002.05534] Abnormal respiratory patterns classifier may contribute to large-scale screening of people infected with COVID-19 in an accurate and unobtrusive manner

 

[Chung Leong]

To add to my previous comment, 96% accuracy is not as good as it sounds when what you're trying to detect is relatively rare. Imagine we have 10000 patients: 100 are infected with the coronavirus, 9900 have ordinary pneumonia. The test would yield 96 true positives and 380 false positives. So only one in five diagnoses is correct. From the patients' point of view that's pretty bad.

 

[bit_user]

To add to my previous comment, 96% accuracy is not as good as it sounds when what you're trying to detect is relatively rare. Imagine we have 10000 patients: 100 are infected with the coronavirus, 9900 have ordinary pneumonia. The test would yield 96 true positives and 380 false positives. So only one in five diagnoses is correct. From the patients' point of view that's pretty bad.

You have to consider that this test is only being administered to those exhibiting respiratory distress. So, the ground truth wouldn't be 100/10,000 (1%), it would more likely be in the 20% to 80% range, depending on how common other types of pneumonia are, in the region in which the test is being deployed.

 

[Chung Leong]

You have to consider that this test is only being administered to those exhibiting respiratory distress. So, the ground truth wouldn't be 100/10,000 (1%), it would more likely be in the 20% to 80% range, depending on how common other types of pneumonia are, in the region in which the test is being deployed.

Pneumonia is among the leading causes of death in China. False positives are going to outnumber true positives, specially outside hot zones. If health officials choose to quarantine all those tested positive, a lot of uninfected people will end up catching the virus as a result.

 

[bit_user]

Pneumonia is among the leading causes of death in China. False positives are going to outnumber true positives, specially outside hot zones.

You mean other causes of pnemuonia will outnumber cases involving Covid-19, I assume. Talk of true/false positives/negatives makes it sound like you're still talking about the test.

If health officials choose to quarantine all those tested positive, a lot of uninfected people will end up catching the virus as a result.

Huh? It's not as if quarantined persons are all stuffed in a room, together. Especially when you know that the test isn't 100% accurate.

I don't imagine anyone is proposing this one test as the only criteria for assessing a positive diagnosis. Honestly, you're starting to sound a bit less rational.

 

[Chung Leong]

Huh? It's not as if quarantined persons are all stuffed in a room, together. Especially when you know that the test isn't 100% accurate.

I don't know if you have seen photos of the new hospital they built in Wuhan. It's just a giant hall with rows and rows of beds. The chance that people with weaken immune systems will catch the virus there is probably near 100%. Ideally, officials would wait for results from a more accurate test before transferring patients . I don't see that happening in Communist China. After all, these false positives will inevitably "correct" themselves.

 

[david germain]

Alibaba reportedly said its artificial intelligence can detect new Coronavirus infections with 96% accuracy.

Alibaba Claims AI Can Identify Coronavirus Patients With 96% Accuracy : Read more

Only been 3000 confirmed cases - i guess it dont the same one more than once??

 

[Olle P]

"Researchers trained the AI on more than 5,000 confirmed cases..."
How many cases with similar symptoms but confirmed not Covid-19 were used?

Are the 96% true positive then where goes the last 4 % ? Also another thing I've seen such companies to ignore false positives in such metrics.

My guess is they opt to prefer false positive to false negative. All positives will get treatment and possibly a second evaluation/verification.

And it uses CT so throughput is to low for china...

I'm curious how many of these patients are routinely getting CT scans. CT scan time isn't cheap and in the middle of an out break the machines are going to swamped. ...

I've read that CT is being increasingly used for patients with symptoms, because it's much faster than the regular lab test.
It's cheaper to backtrack earlier contacts of a contaminated patient the faster a contamination can be established.

Maybe not for completely asymptomatic subjects, but here's an attempt at a cheap, unobtrusive method:

[2002.05534] Abnormal respiratory patterns classifier may contribute to large-scale screening of people infected with COVID-19 in an accurate and unobtrusive manner

That method seems fairly superfluous. Any patient short of breath should get treatment for their respiratory problems and be properly tested for Covid-19. No screening necessary.

 

[ron baker]

The unpalatable FACT is thet everyone will get Covid19.. the only thing the authoritys can do is hope to slow the spread. Then after the first 18 months it will be better understood and incorpprated into your annual flu jab. Its the influx of millions of heavily infected that the health systems cant cope with. I am self-isolating for two months , with another 6-9 months yet. Economy is gonna take a beating , real beating.

 

[ilacıtanı]

How to keep yourself away from Coronavirus

View: https://youtu.be/rw1EEyxskss

 

[bit_user]

The unpalatable FACT is thet everyone will get Covid19..

China seems to have stopped it. Let's see if anyone else can replicate that feat.

the only thing the authoritys can do is hope to slow the spread.

There are other therapeutics that could be effective.

I think you're underestimating how much is still being learned about it.

 

[ron baker]

We are learning .. but I think facts are being withheld. . Why is Africa so quiet? is there a quiet decimation happening.. ALso south America .. we live in a highly connected world and something like this will be everywhere. Narco states, Failed states, War zones Klepto-demaGogerys Turkmenistan ?

 

[bit_user]

We are learning .. but I think facts are being withheld. .

Aw, c'mon. Use your head. Just because you don't hear about something doesn't mean it's being witheld.

Why is Africa so quiet?

Because they don't have test kits or much resources to devote to it. They also have other epidemics they're dealing with, and this one hasn't quite risen to the top of the heap.

Here's the March 8th edition of a free, weekly science podcast that contains a segment on it.

https://www.bbc.co.uk/programmes/w3csym36​

There's, of course, more recent news. Did you try a Google News Search?

https://news.google.com/search?q=covid-19 africa&hl=en-US&gl=US&ceid=US:en​

ALso south America .. we live in a highly connected world and something like this will be everywhere.

Yes, there too.

https://news.google.com/search?q=covid-19 brazil&hl=en-US&gl=US&ceid=US:en​

Narco states, Failed states, War zones

Yes, people are worried about those, too. However, it's hard to get information out of those places, especially now that reporters for big news organizations can no longer travel freely. Life in war zones or countries like Venezuela is already so bad that the disease has to hit pretty hard to overshadow all of the other misery in those places. However, even Venezuela recently closed its borders out of concerns about COVID-19.

https://news.google.com/search?q=covid-19 venezuela&hl=en-US&gl=US&ceid=US:en​

 

[ron baker]

So how many cases in India, Brazil ,Latam ? unknown .. how many excess deaths , in those Countrys.. unknown. We'll probably never know, estimates only , in ten years time.

 

[kkevich]

even with artificial intelligence detecting Coronavirus with 96% accuracy we'll never know the true statistics for example of Africa.
Even with the development of AI in medical imaging or testing coronavirus there's a huge lack of resources to take tests and say about true statistics