Unable to test every tourist and unable to turn them away, Greece used ML to pick visitors for COVID-19 checks Faced with limited resources in a pandemic, Greece turned to machine-learning software to decide which sorts of travelers to test for COVID-19 as they arrived in the country. The system in question used reinforcement learning, specifically multi-armed bandit algorithms, to identify which potentially infected, asymptomatic …
https://docs.microsoft.com/en-us/office/vba/language/reference/user-interface-help/rnd-function
Eh no. the relevant piece of the article is here:
The code identified 1.85 times as many asymptomatic, infected travelers as random testing methods, "with up to 2-4 times as many during peak travel," according to the team. "To achieve the same effectiveness as Eva, random testing would have required 85 per cent more tests at each point of entry," their paper stated, regarding that first figure.
Even the WHO has long acknowledged that the asymptomatic do not drive “the pandemic” in any form. Most of them do not spread the virus, as their virus load is much too low. In other words: stop testing healthy people. Oh, wait – we cannot do that as the whole show would be over then!
Government's hounds voting me down in 3, 2, 1 …
Dunno what you're looking at, but every viral load work I've looked at has found essentially no difference in viral load between {a,}symptomatic. The WHO has absolutely disgraced itself during this pandemic -- I wouldn't be placing any weight on anything they pronounce.
On the other hand, the only non-pathetic close study I've seen of surgical mask use vs viral broadcast showed that rhinovirus basically ignored them, flu got mostly caught but a fair whack got out, but coronaviruses A/ had 0 get thru, completely stopped, B/ only about a third of victims had any detectable live virus coming out of their mouth/nose. Pre Covid but could be the same.
"On the other hand, the only non-pathetic close study I've seen of surgical mask use vs viral broadcast showed that rhinovirus basically ignored them, flu got mostly caught but a fair whack got out, but coronaviruses A/ had 0 get thru, completely stopped, B/ only about a third of victims had any detectable live virus coming out of their mouth/nose. Pre Covid but could be the same."
Oh boy. This isn't science. When somebody tells you that a red basketball fits through the hoop, but the same ball in yellow doesn't, you know that he's taking a shit on you. ;-)
Errr... your apparent ignorance is truly astounding. Viruses are not all the same size or shape. Rhinoviruses are much smaller than coronaviruses, for example.
You might want to tone down your vehemence until such time as you learn something about what you're shouting about. Right now, you clearly don't have the first clue about any aspect of viruses even in general, let alone SARS-CoV-2.
First of all, not all viruses are the same size - not even close, in fact.
But that's not the point; the point is that masks don't need to stop individual viral particles to be effective. They need to stop the water droplets that carry them through the air. Those droplets are enormous compared to the virus itself, and many of them can easily be stopped by a mask.
stop testing healthy people
Make that "apparently healthy people". The incubation period for Covid (i.e., the time from infection to symptom onset) is roughly 1-14 days, with an average of 5-6 days. During that period, you may well be contagious. Some studies have concluded that you are in fact most contagious 1-2 days before the onset of symptoms, when the viral load is highest and your immune system has not fully kicked in.
https://medical.mit.edu/covid-19-updates/2020/07/how-long-symptom-onset-person-contagious
https://www.nature.com/articles/s41591-020-0869-5
> most contagious 1-2 days
Delta appears even faster. Australia had one well documented instance of a chap becoming infectious within 1-2 *hours* of exposure. Unfortunately, they were both on their way to the MCG for a large event...
You've got to test healthy people to keep the statistics working properly. Remember, there isn't an absolute 100% yes/no diagnostic, at least one that doesn't take an impractical length of time. So of the people you test you're going to discover four groups --positive and are sick, positive and are not sick, negative and sick, negative and not sick. This combined with general knowledge about the population in general will tell you what are the chances any individual will be infected. What those rather clever Greeks are up to is tweaking the probabilities based on experience in order to refine the test odds. Its very logical when you think about it but in our world of absolutes -- there is only blackest black and whitest white in the popular imagination -- the subtleties of what they're up to will escape most people.
(This 'refining on results' is what people used to call "Artificial Intelligence".)
This post has been deleted by its author
Even the WHO has long acknowledged that the asymptomatic do not drive “the pandemic” in any form.
The study I see says that the infectivity of some asymptomatic SARS-CoV-2 carriers might be weak. "Weak" != "null".
Most of them do not spread the virus
So some do, don't they?
The Greeks have EVA,
The Welsh and Scots have ZOE (https://covid.joinzoe.com/)
and the English have... DIDO
I don't think the tech is anywhere close to being able to detect at risk groups but the idea is sound. I would hope the tech is good enough to mostly filter out the lone travellers who are minding their own business*.
* For entirely selfish reasons - that would often be me (I usually holiday by myself).
"Eva was terminated after November. “When the tourist season was over, the number of arriving international passengers became very low, and so there was very little benefit to permitting non-essential travel to the country,” Drakopoulos told The Reg."
This is where the whole "ML" study falls apart entirely. Having data about people the model wouldn't have selected in the first place, in order, say ... to learn, like in "Machine Learning" would have been priceless.
But they shut it down !
So, they end up with a model only learning about the sample it did select. So hardly more than tossing a coin.
"The researchers declined to say how many people total were tested after being singled out by Eva, citing privacy reasons"
Wow, the bullshitometer just went through the roof, here !
"Gupta told El Reg that accuracy and Eva’s false positive and negative rates weren't relevant in the study."
Thanks $DEITY, since the whole methodology sent those metrics out of the window. Glad they didn't care :)
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2023
