COVID-19 might be incurable as yet, but humans have some potent weapons in our own arsenal, including artificial intelligence (AI) and high-performance computing (HPC).
Here are ten examples of how AI and HPC are helping the fight against the virus.
1. Identifying the outbreak
Early awareness is critical for stalling a disease outbreak, especially if it starts where bad news might get covered up.
This was exactly why Kamran Khan founded BlueDot after the SARS epidemic of 2003. BlueDot's algorithm hunts for outbreaks by scanning ‘foreign’ language news sources for trending stories and uses global airline ticketing data to predict where the disease will spread.
According to CNBC, BlueDot detected the outbreak in Wuhan and warned its clients on 31 December.
2. Predicting the spread of disease
An epidemic is a highly complex system of systems with millions of variables. Figuring out which of the variables are the main deterministic factors and measuring their likely impact requires immense computing resources.
HPC running machine learning is particularly suited to modelling such complex scenarios, to produce algorithms capable of predicting outcomes when the variables change. Predictive analytics is being used to develop forecasts like this one from Australia to predict the future course of the Covid-19 outbreak.
3. Tracing people who could be infected
The early containment stage for COVID-19 involves tracing anyone who might have been in contact with an infected person, or in the vicinity of the pathogen, and isolating them.
China's exceptional advances in state-wide surveillance technology and their national database of faces enabled them to track one individual from Hangzhou, first through automated number plate recognition (ANPR) and later through automated facial recognition (AFR), when he breached his own self-isolation, according to Reuters.
4. Communicating with citizens
Government health helplines are quickly swamped with calls from anxious citizens, even in countries with highly developed healthcare infrastructure, as UK's National Health Service helpline, NHS 111, soon discovered.
To resolve the overload, NHS 111 introduced a digital version of its helpline, using an AI-powered chatbot and online symptom-checker to remove the burden from call centres, hospital ERs and doctors’ offices. In its first five days in service, NHS 111 Online has handled more than one million enquiries.
5. Diagnosing for infection
AI-powered image analysis is being used in China to speed up the processing of CT scans, enabling a diagnosis to be achieved in 15 seconds with 90% accuracy, compared to 15 minutes for a human radiologist.
By re-purposing a system that was designed to spot lung cancer in CT images, Ping An Smart Healthcare in China has been able to deliver the COVID-19 smart image-reading system, helping to triage 5 000 new cases in more than 1 500 locations in its first two weeks of operation, reports Bioworld.com.
6. Finding an effective treatment
Deep Learning enables pharmaceutical researchers to predict the characteristics of new chemical compounds that don’t exist yet, or predict which existing molecules might be effective against a new disease - substantially reducing the lead time to develop a new drug.
Last week, researchers used the Oak Ridge National Laboratory supercomputer to test how 8 000 chemical compounds would interact with the Covid-19 virus, identifying 77 compounds that could help with future research.
There are already successful precedents in drug discovery. Researchers from MIT recently published a paper in Cell, describing how they used AI to discover a powerful new antibiotic, named Halicin, capable of killing many of the world's nastiest pathogens.
7. Enforcing the quarantine
AI-powered unmanned aerial vehicles (UAV) or smart drones are commonly used to access areas that are hard-to-reach or dangerous, such as power lines and cell towers.
Smart survey drones are now being re-tooled in China to police quarantined areas from above, to perform decontamination protocols and to deliver lightweight essential supplies. A report in the South China Morning Post describes the initiative in detail.
8. Delivering essential supplies on the ground
The mobility sector is another casualty of the lockdown in China, but driverless vehicles have taken over where humans are prohibited.
Driverless logistics vehicles are re-supplying hospitals in infected areas, while autonomous robots are now delivering meals within hospitals for more than 40 Chinese cities, as this report from the World Economic Forum describes.
9. Eliminating fake news
Fake news, and dangerous advice, about coronavirus has been spreading faster than the virus itself, and there's no shortage of it, according to a BBC report.
AI is being used to flag fake news by learning which Web sites are more authoritative, identifying posts that use sensational language or images that don't fit with the right date and location for the 'story' they support. Scott Tong explores the detail in his article published last month on Marketplace.
10. Managing complex supply chains
Empty shelves are a familiar sight already, but panic-buying of hand sanitiser is only the tip of iceberg. If we follow supply chains that originate in China, and many do, upstream to their source, COVID-19 is proving to be a wrecking ball, preventing workers from even going to work.
Disruption so far upstream in any highly optimised supply chain will have enormous knock-on effects downstream.
Mark Balte's article illustrates the extraordinary degree to which AI is now intrinsic in the SCP domain. Rebooting global supply chains and getting back to a new 'normal' will be impossible without it.
AI is the ace in our hand
Humankind is generally at its best when confronted by the need for urgency and an existential threat. COVID-19 brings us both. In the weeks ahead, the greatest minds will come up with something new. It's what we've always done in the past, only this time we will have AI to extend the limits of our brainpower as never before.
* Andrew Quixley is an independent AI and new technologies specialist, with twenty years of experience in the global enterprise software industry. Most recently, he led IBM's go-to-market drive into the nascent market for AI software and services.
Share