Spread of COVID-19 can be slowed down by Artificial Intelligence

Artificial Intelligence and relation to Covid-19

In Greece, excellent findings from a novel approach using machine learning to the COVID-19 exam have been obtained. In order to identify and restrict the importation of symptomless COVID-19 instances among arriving foreign passengers between between August & November 2020, the technology, known as Eva, dynamically used latest testing results gathered at the Greek border. This helped contain this same number of cases and fatalities in the nation.

More asymptomatic, infected tourists can be found using a machine-learning technique than with traditional random testing.

According to the data, Eva discovered 1.85 times as many asymptomatic sick travellers than would have been found through regular, random monitoring testing. The identification of rates of infection was up between three to four times greater during the busiest travel months of September and October than it was during random testing.

Food Safety and the Coronavirus Disease 2019 (COVID-19) | FDA

The project’s findings are detailed in the paper “Deploying an Artificial Intelligence Scheme for COVID-19 Test results at the Greek Border,” written by Kimon Drakopoulos & Vishal Patel out from University of Southern California, Jon Vlachogiannis from investment firm Agent Risk, Christos Hadjicristodoulou from of the University of Thrace, and Hamsa Bastani, a School of business professor of operations, information, and decisions and affiliated teaching staff at Analytics.

Is AI Hard To Learn? A Guide To Getting Started in 2023

According to the publication, “Our approach opens the door for utilising [artificial intelligence] & actual information for public health purposes, such as border checks during a pandemic.” Eva also holds the possibility of optimising the already overworked test rig in most nations due to the fast emergence of a brand-new coronavirus strain.

“The important question was whether we could run the tests in a better way using dynamic surveillance to detect more diseased travellers, given the set budget for testing,” said Bastani. The lack of the test infrastructure at its national boundaries to realistically screen every entering passenger is one of the main problems nations encounter when dealing with COVID-19. Because such extensive testing would be expensive and time-consuming, the majority of nations either screen inbound travellers from particular nations or provide random COVID-19 testing.

Additionally, Eva made it feasible for Greece to detect COVID-19 infection spikes nine days sooner on average than would have been the case with machine learning-based techniques utilising solely publically accessible data.

According to Bastani, the technology behind Eva is a “contextual bandit algorithm,” a machine-learning framework created for “sequential decision-making” that takes into consideration a variety of real-world issues including time-varying information & port-specific testing budgets. The method strikes a compromise between the necessity to keep accurate surveillance estimations of COVID-19 incidence across nations and the distribution of few test results to identify potentially infected passengers. While similar algorithms have been used in online advertising & A/B testing, Eva is the 1st instance of that tech being used to solve a public health issue, she noted.

Overcoming Data Challenges during Covid-19

Eva is an improvement over traditional border control methods since it does not rely on data that has been made publicly available, which has number of disadvantages.

The main reason why publicly available data is of “poor quality” is because various nations have varying reporting standards and testing procedures. Although it is usual practise to concentrate testing efforts on symptomatic individuals, the prevalence rate that results may not accurately represent the symptomatic group that is most prone to travel. According to Bastani, there is frequently a reporting delay because of inadequate infrastructure. “We can identify a country’s COVID instances are increasing based on the information we are actively gathering at borders 10 days before you would see that reflect in the public data.”

When the Greek deployment was beginning in July, Bastani noted in an interview with the Wharton Finance Daily radio show on SiriusXM, “Testing is normally aimed towards sick folks rather than asymptomatic individuals.” “You may reasonably assume that those who are arriving are asymptomatic.” This emphasises how important it is to use data that truly reflects the frequency of asymptomatic COVID-19 travellers across nations rather than depending on publicly available statistics.

By constantly gathering test results at the Greek border, Eva’s algorithm overcomes the low quality of publicly available data and maintains high-quality surveillance estimations of the prevalence in each nation. The article observed that the Greece deployment “prevented more sick people from entering by adaptively changing border rules nine days sooner.” According to Bastani, “it is a significant time period in which many high-risk folks would have undoubtedly came in & infected other citizens.”

Publicly available information is frequently used in border control procedures, although Bastani noted that this information is frequently incorrect and inconsistent between nations. The discrepancies result from testing data censoring by certain nations and even different definitions of the a COVID-19 fatality, she continued. Like an instance of incorrect data, she cited the recent finding that COVID-19 fatalities in care facilities in New York were not properly counted. When comparing mortality tolls across nations, the problem is made worse since some countries keep highly reliable records while others don’t.

Greece was the first nation to implement border restrictions using Eva’s dynamic random observation testing methodology. The model outlines the infrastructure needed to gather COVID-19 test results, utilising those to generate estimates and guide decisions for further tests in a continuous feedback loop.

Using the Eva model, Greece mandated that every person or family wishing to enter the country complete a digital “passenger locator form” 24 hours prior to arrival. On this form, individuals submitted basic basic information about themselves, such as other countries they had visited in the previous year. Everyone who turned in the forms was given a tracking QR code. Eva’s programme analyses the data on the forms to determine who needs to have a COVID-19 test. The border control authorities in Greece handled around 38,500 forms every day on average; about 18% of the individuals who submitted the papers did not ultimately appear.

Keeping COVID-19 at Bay

Greece kept its case count “extremely low pretty much throughout the summer,” according to Bastani, thanks to Eva’s focused testing that enabled adaptable border control procedures. Contrary to many nations that had to fully shut down, the nation was able to continue some level of economic activity, she said. After a rise in COVID-19 cases, Greece implemented a second lockdown & travel restrictions in November.

Crisis Management Minister Hardialias Checks COVID Preparadness

At a news conference in July of last year, the Greek government commended Eva for her efforts. Nikos Hardalias, Greece’s civil protection & deputy minister for crisis response, who is in charge of the nation’s COVID-19 Response Taskforce, said that the AI system created by Bastani, Drakopoulos, Gupta, & Vlachogiannis has been helpful in both preparing for the country’s opening to tourists from all over the world and in allowing flexibility in judgement regarding our COVID-19 strategy.

Covid-19 and free-to-use Technology

Since Eva is an accessible technology, Bastani & her team will give it for free to any nation that requests it. They have presented before COVID task teams in a number of European Union nations. Designing passenger locating forms that are acceptable for various immigration procedures and integrating back-end resources like testing laboratories would be required to adapt it to other nations.

Bastani presented a compelling case for governments to collect private information, such that produced by the traveler locator form used in the deployment in Greece, and tailor them to their unique circumstances.

No nation should only rely on publicly available information, she argued. Instead, countries should actively watch who enters their borders, test at least a portion of them, and use the results to determine the most effective methods of border security.

“That said, if a nation doesn’t have the means to do that, it’s probably preferable to implement a policy that replicates another nation that has done so instead of depending simply on public statistics,” the author stated in the statement.

Hamsa Bastani: Personalized medical decision-making - YouTube

Bastani and her coworkers are improving Eva to include more passenger-specific data than they did in the deployment in Greece. Due to the Data Protection Act in Europe, they were only able to utilise anonymized, aggregated data with restricted demographic information for Eva. Other nations with less strict data privacy laws may be able to collect more information, including information on occupation, according to Bastani. We are aware that some vocations have a significantly greater COVID-19 risk than the others.

She said that Eva may be taught to utilise pooling to lessen the restrictions encountered by testing laboratories. According to her, overburdened laboratories might exchange their data with other laboratories that might have extra space at any particular time. Eva said that, in a similar vein, dynamic data might be used to assist in figuring out the ideal personnel numbers at laboratories and other sites in the testing infrastructure.

edited and proofread by nikita sharma

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button

Adblock Detected

Please consider supporting us by disabling your ad blocker