Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

AI tools have been built to catch Covid, but none of them helped

technologyreview.com

Post reply on HN
Print view View original

61–70 of 136 posts

Re: AI tools have been built to catch Covid, but none of them helped

#61

Post by forcry »

new299 wrote: post #5

> Many unwittingly used a data set that contained chest scans of children who did not have covid as their examples of what non-covid cases looked like. But as a result, the AIs learned to identify kids, not covid. > Driggs’s group trained its own model using a data set that contained a mix of scans taken when patients were lying down and standing up. Because patients scanned while lying down were more likely to be se…

>I wouldn't trust any method of fixing errors in data collection. You need to redo data collection and perform a proper case-control study.

Climate research would like to have a word with you.

Re: AI tools have been built to catch Covid, but none of them helped

#62

Post by new299 »

LeetHacks wrote: post #56
new299 wrote: post #5

> Many unwittingly used a data set that contained chest scans of children who did not have covid as their examples of what non-covid cases looked like. But as a result, the AIs learned to identify kids, not covid. > Driggs’s group trained its own model using a data set that contained a mix of scans taken when patients were lying down and standing up. Because patients scanned while lying down were more likely to be se…

If properly done, that approach would most likely generate a model that works really well and detects COVID for the right reasons. Put it out in the wild and it breaks down again because all radiologist would have to use the exact same method of scanning patients. Practice shows that scan parameters, position, annotations vastly differ across hospital, scanner and radiologist. A client of mine wanted me to look into…

It's a first step I guess. At least if you can get the classification to work with controlled data collection you have reasonable evidence that the approach can work.

If all other parameters area randomized between case and control, this is also fine. I'd guess you can also add illumination and other artifacts to the datasets to make the training robust to this.

But ultimately in the extreme case (like wanting to submit images from a point and shoot camera) I suspect you'll have a hard time building a system that's robust to that... or robust enough to be used in a diagnostic context...

I personally don't think I'd be comfortable working without a standard configuration used by the radiologists and a controlled protocol.

Re: AI tools have been built to catch Covid, but none of them helped

#63

Post by tablespoon »

briefcomment wrote: post #22
tablespoon wrote: post #17

Earlier quoted context omitted.

> You think a 100% failure rate on the hottest topic in the world right now is likely? Yeah. Attempts at powered flight had a 100% failure rate in the 1800s. It's kind of absurd to think that "AI" (especially in its current incarnation) must be able to solve any "hot" problem that it's thrown at.

You think AI is so primitive that we can't make any headway into the biggest problem we currently have? That seems like a fairly fringe view. Especially on a problem that has well defined data like medical imaging.

> You think AI is so primitive that we can't make any headway into the biggest problem we currently have?

More or less. It's almost certain if "AI" makes any headway at all, it will fall far short of the hype.

> That seems like a fairly fringe view. Especially on a problem that has well defined data like medical imaging.

I doubt it.

Re: AI tools have been built to catch Covid, but none of them helped

#64

Post by forcry »

JohnJamesRambo wrote: post #31

Should have built AI tools to educate the ignorant populace and encourage them to get a vaccine. I know there have to be tons of adboys and girls out there whose only job is to persuade people to do something that could have knocked this out of the park. I see we are still at about 50% vaccination and Walmart is begging people to come get a free vaccine. I’m sorry my tone is negative but this is really getting me fru…

> The stupid is winning.

Calling someone stupid just because they do not share your blind trust in a bunch of people is kind of malicious.

Re: AI tools have been built to catch Covid, but none of them helped

#65

Post by tablespoon »

afurculita wrote: post #3

Yep, that's because they didn't use a blockchain. That would have solved everything.

> Yep, that's because they didn't use a blockchain. That would have solved everything.

I just bought the COVID-19 NFT and burned it. Pandemic over.

Re: AI tools have been built to catch Covid, but none of them helped

#66

Post by new299 »

forcry wrote: post #61
new299 wrote: post #5

> Many unwittingly used a data set that contained chest scans of children who did not have covid as their examples of what non-covid cases looked like. But as a result, the AIs learned to identify kids, not covid. > Driggs’s group trained its own model using a data set that contained a mix of scans taken when patients were lying down and standing up. Because patients scanned while lying down were more likely to be se…

>I wouldn't trust any method of fixing errors in data collection. You need to redo data collection and perform a proper case-control study. Climate research would like to have a word with you.

It seems like a very different scenario as we don't have control Earths that we do case-control studies on...

Re: AI tools have been built to catch Covid, but none of them helped

#67

Post by Romanulus »

JohnJamesRambo wrote: post #31

Should have built AI tools to educate the ignorant populace and encourage them to get a vaccine. I know there have to be tons of adboys and girls out there whose only job is to persuade people to do something that could have knocked this out of the park. I see we are still at about 50% vaccination and Walmart is begging people to come get a free vaccine. I’m sorry my tone is negative but this is really getting me fru…

... and yet, CDC says 74% of people infected in Massachusetts Covid outbreak were fully vaccinated.

Re: AI tools have been built to catch Covid, but none of them helped

#68

Post by trhway »

JohnJamesRambo wrote: post #31

Should have built AI tools to educate the ignorant populace and encourage them to get a vaccine. I know there have to be tons of adboys and girls out there whose only job is to persuade people to do something that could have knocked this out of the park. I see we are still at about 50% vaccination and Walmart is begging people to come get a free vaccine. I’m sorry my tone is negative but this is really getting me fru…

>to educate the ignorant populace

ok, lets educate. And before educating others lets educate ourselves.

https://www.cnbc.com/2021/07/30/cdc-study-shows-74percent-of...

"About three-fourths of people infected in a Massachusetts Covid-19 outbreak were fully vaccinated, according to new data published Friday by the CDC.

The new data, published in the U.S. agency’s Morbidity and Mortality Weekly Report, also found that fully vaccinated people who get infected carry as much of the virus in their nose as unvaccinated people. "

And another point related to education. Couple days ago a White House official stated :"We're dealing with an evolving virus. There is no precedent, and there is no playbook". Note, it isn't the White House of the previous buffoon, it is supposedly educated current White House. No wonder that being led by the people like this the current pandemic response isn't that much better than the ones of the previous centuries.

Re: AI tools have been built to catch Covid, but none of them helped

#69

Post by iamstupidsimple »

gumby wrote: post #24
briefcomment wrote: post #22

Earlier quoted context omitted.

You think AI is so primitive that we can't make any headway into the biggest problem we currently have? That seems like a fairly fringe view. Especially on a problem that has well defined data like medical imaging.

Medical diagnosis has been one of the core ai domains domains since the early 1960s so people have been breaking the picks at that coalface for almost 60 years. And those older systems has more intelligence in them. Todays’s “AI” is a small number of tricks aimed at large amounts of data, with an unprecedented and enormous amount of marketing added.

> And those older systems has more intelligence in them. Todays’s “AI” is a small number of tricks aimed at large amounts of data, with an unprecedented and enormous amount of marketing added.

If you're talking about deep neural networks, I can understand this viewpoint. But generally the last decade has proven widely successful for high quality vision recognition models that just weren't available before then. And with something like transfer learning, you can take a powerful off-the-shelf model and specialize it without a huge dataset.

The real barriers IME are around legal and privacy implications. There's also a strong argument about if these models creating enough value in the first place, but they can work on a technical level.

Re: AI tools have been built to catch Covid, but none of them helped

#70

Post by forcry »

new299 wrote: post #66
forcry wrote: post #61

Earlier quoted context omitted.

>I wouldn't trust any method of fixing errors in data collection. You need to redo data collection and perform a proper case-control study. Climate research would like to have a word with you.

It seems like a very different scenario as we don't have control Earths that we do case-control studies on...

Yes, of course, but not being able to do the proper minimum viable thing 'x', does imply the possible alternative is good enough..

Also I was mostly talking about correcting errors in existing data..

Post reply on HN