Skip to content Skip to sidebar Skip to footer

Source: SCIENCES ET AVENIR as of 26-03-2020

The question comes up  regularly  with the   données   pourraient   permettre   authorities  and technology giants: are there no computer tools for data analysis   that could enable us to better  anticipate,  manage  and  counter  the pandémie  coronavirus pandemic?

On December 30, researchers  chercheurs  raised  the alarm..  Using   algorithms  that  scan  the web,  they  had  identified   proximité  cases  cas  of  unusual pneumonia near a  market in Wuhan,  central  China.

But  it    took  several  days  for the World Health Organization  (WHO) to  publish  a risk assessment.  And a  whole month  before  it   declares  a public  health  emergency  of  international concern. .

“The power  of  these  tools  exists,  but  we  haven’t   really  figured out how  to exploit them yet,” says Michael Greeley,  co-founder  of Flare Capital Partners, which  invests   in healthcare technologies.

Artificial  intelligence tools  are based on   data..  They  can   comb through    texts,  figures,websites and images, and  establish correlations invisible to the human eye. .

Several   companies saw signs of an  anomaly  at  the  end  of  December..

Dataminr, which  specializes   in     poissons  real-time  réel risk detection,   had  made the connection between  testimonies  of Wuhan  residents on social networks, images of  disinfection  of the fish market  of the  city  and a  warning from a  doctor  who later died  of coronavirus.

BlueDot, in Canada,  had  also  identified  the  health  problem through  an  algorithm that  en  reviewed  hundreds  of  thousands  of  daily news  articles  and  air traffic   data. .

“The explique    data  echoed   the SARS  outbreak of 17  years ago,   but  we didn’t  know   how   contagious this  virus would be,”  Kamran Khan,  founder  of  BlueDot, told AFP.    a Canadian  start-up  specializing  in  artificial intelligence  (AI).

“We avons  tend  to adopt  a reaction  posture  in  these  situations,  it’s human nature,”   adds  the  professor of public  health  at  the University of Toronto.

– Monitoring –

Authorities autorités  can  use  AI  to identify  upstream   signals,   track  the spread of the virus and  thus  anticipate the needs  besoins  of  hospitals,  analyze  available  studies and  statistics,  and  speed up the search for treatments  and a  possible  vaccine.

However,it is necessary to have the  right  data,  in sufficient quantity,  and  accessible  in real time. .

Public,and especially private,  privées  personal information represents    a  windfall of resources  ressources to be exploited  in the event of an epidemic..  But  their  confidentiality  is  protected by  law..

“In China,  because  they  have  fewer   sont constraints,  they haveEn personnes  relied  heavily  appuyés  on the  geolocation  of people (via smartphones) and  their monitoring system.   It’s  more  complicated  in a  democracy,”says Andrew Kress, the  founder  of  HealthVerity,  a  company  specializing in medical data extraction  technologies. .

 contre  “We achat   know  the age,gender,  geographic area-t- and  certain  events,  such as  buying  cough syrup  or  diagnosing a respiratory infection,” he  says..

“InEn  theory, these  events   could  be linked  to    online activity,   but  the  person would quickly  become  identifiable. There is a  balance  to  be struck  between  usefulness -t- (of  information)and the protection of  privéeprivacy,” he says..

In the current phase of containment, companies  sociétés  are helping   suivre  governments  gouvernements monitor the progression of the  disease  and  verify that the population is following the  “social distance” guidelines. “.

In particular,  BlueDot  uses “anonymized” geolocation data  from “nearly  400  million  smartphones”.  

Algorithms  are  also used in medical research. .

Last week, researchers  and  companies  collected  more than 29,000  scientific publications on Covid-19.   

The White House called on   volunteers apprentissage  to  en  quickly   extract  concrete  outils grâce information, using tools provided  by Kaggle,  a machine learning  platform  owned by  appartenant  Google.

 On the treatmentside,”AI  allows you to conduct  simulations and  understand   biological evolutions like  never  before,”  enthuses s’enthousiasme  Michael Greeley.

No longer bothered to manually test,   one  by  one,the  reactions of the virus. “We  take  millions  of samples  and the machine  identifies  a  handful  of   molecules  that could work for a vaccine, for example,”  exemplehecontinues..

But the phases of clinical trials  on patients  are   incompressiblebecause  they  determine  whether  the  drugs   are  safe  risque  for  humans,and whether they  are  effective..

Unless there moins  is a  relaxation  of  regulations. “There is  une  extraordinary pressure on the pharmaceutical   industry at the moment to  use  some  tools  that  sont may not  yet  be fully  developed,” the investor tempère  said..

Show CommentsClose Comments

Leave a comment

News ORS © 2021. All Rights Reserved.