AI’s Impact on Coronavirus
AI is on the front line of the fight against coronavirus. From initial detection and forecasting its spread to efforts to flatten the curve and develop a vaccine quickly, AI is accelerating the fight in dozens of ways.
Our lives have been fundamentally altered by the rise of COVID-19. The current outbreak has led to unprecedented shutdowns around the world, and billions of people are isolating to help reduce the spread. It’s been more than a century since the world has seen an outbreak on this scale. Although that has created countless unforeseen challenges, it means we have a much more powerful set of tools with which to fight the virus — including artificial intelligence.
The tools with which we make predictions, research treatments, and distribute much-needed supplies are powered by the world’s most advanced algorithms. Here are what some of the responses look like, how AI is impacting the front lines, and the role AI is playing in the home.
Predicting the Spread of COVID-19
Epidemiology relies almost entirely on data to predict the potential spread of disease. Although there are established patterns that can be followed and historical data to highlight what a pandemic can do to the population, statistical analysis of infection rates and the impact of social distancing policies can be used to get a much more precise measure of the virus’s spread.
The first indications of the outbreak were delivered by BlueDot’s infectious disease tracking algorithms when they identified a string of pneumonia cases in Wuhan, China. The company uses both machine learning and natural language processing to identify patterns from hundreds of data sources — everything from flight patterns to weather reports. Similar technology is being used now to process huge volumes of data that could help identify hot spots and treatment options.
Roni Rosenfeld, a computer science professor at Carnegie Mellon University who runs the Delphi research group, was recently tasked by the Center for Disease Control (CDC) to help predict the spread of COVID-19 using the team’s machine learning algorithms. Their technology is designed to improve disease forecasting and put it more on par with weather forecasts that we routinely rely on. They’ve used «wisdom of crowds,» which aggregates the feedback of dozens of individuals to make specific forecasts.
In China, machine vision algorithms and hardware are being used to screen hundreds of people at a time for raised temperatures in Beijing’s railway stations. Another system has been developed that can identify COVID-19 with 95 percent accuracy in chest CT scans.
Flattening the COVID-19 Curve
AI is being used to help flatten the curve and limit the number of concurrent cases of the disease so we don’t overwhelm hospitals. One example is the work of CloudMedx, drawing data from payers, providers, and patients to make targeted predictions about the demand for medical equipment, staff, and the flow of patients both geographically and by duration.
Companies such as CloudMedx are leading the way in self-assessment tools already being leveraged by Medicare.gov and Covid2019.health. With so few tests completed early in the spread of the disease in the United States and so many cases considered «mild,» this allows modeling of potential cases even when those individuals do not visit a doctor or receive a test.
There are other companies working on similar tools. Alphabet’s Verily has launched a testing pilot in four California counties. Oxford researchers in the U.K. have been developing a home test that would allow for self-evaluation. There are dozens of potential solutions being developed, and AI is enabling their rapid creation and deployment.
Other tools are being leveraged to identify high-risk candidates and protect them proactively from exposure. Medical Home Network, for example, is using an AI system to identify their highest risk patients and target their outreach efforts to those individuals. Instead of contacting 122,000 individuals in their database, they can narrow their focus to the 4.4 percent that would be most at risk. The system can account for medical risk factors as well as social factors — those who work high-risk jobs as well as those who may be socially isolated during the outbreak with little or no family around.
Vaccines may still be a year away, but AI is enabling rapid development. Chinese researchers were able to decode the genetic sequence of the disease and upload it to a public database by January 10, and researchers around the globe are now using computer algorithms to generate potential vaccine designs. San Diego-based biotech firm Inovio Pharmaceuticals was able to use its machine learning system to develop a potential vaccine candidate in just three hours. Moderna Therapeutics did the same in just two days.
AI Is Keeping the Internet Running
Our lives are highly connected by email, video chat, and applications such as Facebook. We tend not to think of them as human operated, but many are. Social networks, in particular, rely heavily on human moderators to screen for and remove content that isn’t appropriate for the public, which puts those companies in an interesting position.
Most content moderators are contractors and are not permitted to work from home by their agencies. Facebook, however, has sent them home anyway, meaning the company is now, more than ever, relying on its machine learning algorithms to police and remove content that may violate its policies. The most sensitive work is being dealt with internally, but for less-sensitive content, AI is being deployed at scale as never before. It’s not as precise, and these companies will freely admit it. YouTube even posted a notice on its site warning that some content may be removed that hasn’t violated any policies.
The situation shows that simultaneously AI will be capable of filling this gap and reducing the stress and anxiety such jobs can cause human moderators. It also shows that we’re not quite there yet and more work is needed to fully implement such a system.
AI in the Home for Remote Workers
There are many parts of our fight against COVID-19. Identifying and eventually stopping the spread of the disease is the most important, but millions are staying home to help in those efforts. By limiting contact with others and staying home, we can help flatten the curve.
AI is helping to make this a reality. Remote monitoring and response to potential cybersecurity threats have been implemented in many of the country’s largest companies. With entire workforces sent home to work remotely, these systems are being tested at scale for the first time with IT workers unable to address issues on site. Chatbots are allowing websites to continue providing consistent service to their visitors, even when the workforce is more distributed and resources are not readily available. Call centers are leveraging AI to operate remotely without impacting quality of service. Emotion AI is still able to capture data from individual calls and inform the response in a time of heightened stress and anxiety.
AI Is on the Front Lines of the COVID-19 Response
With so many people at home — schools out, businesses closed, and resources stretched thin — AI is filling a vital gap like never before. Some tools are being accelerated to market while others were built for this situation.
From faster response times to spreading infection rates to smarter quarantine recommendations and symptom identification, technology is playing a vital role in the fight against this overwhelming public health crisis.
Fuente: By Rana Gujral
2021 Predictions: As Cloud Adoption Accelerates, Governance and Security Become Critical
The predictions below encompass what we can expect to see for multi-cloud adoption, cloud security and governance challenges, the role of
How to Avoid Inefficiencies and Engender Trust in a Data-Driven Enterprise
New realities in light of COVID-19 have caused many of us to reimagine our businesses, industries, and lives. As we try to find our way t
The Growing Role of Data Lineage in Modern Data Management
End consumers of BI reports, analyses, data sets, and other data-driven products have questions that they regularly ask: Where did the da