Four ways AI could help COVID-19 efforts

As coronavirus (COVID-19) continues its spread, we learn more about it each day. The more we learn, the better we can fight.

And a major asset in that fight could be artificial intelligence (AI). AI has been around since the 1950s as a computer science concept in which machines become “smart” and perform tasks they progressively learn from the data we give them.

“AI requires a massive amount of data,” said Tanya Berger-Wolf, director of Ohio State’s Translational Data Analytics Institute (TDAI) and a computer science and engineering professor. “You want to see what kinds of patterns exist in data; you give it data and these algorithms start finding repeated patterns.”

Those repeated patterns can prove critical in how we halt the outbreak. Here are at least four ways Berger-Wolf believes AI could prove helpful to our efforts going forward.

  1. One way AI can prove beneficial is in learning how individuals respond to COVID-19, which might help us make accurate predictions for populations.

    “There’s starting to be a lot of DNA sequence data, so that’s where patterns can start being learned. Can we take the information about an individual person with symptoms, the genomics of the host and the virus, the medical history and the clinical data, and predict what would be the course of the disease for a person?

    “Can we then start aggregating those into predictions for a population? What happens if we lift restrictions? What happens if we change the hours of operations of the stores? All of these ‘what if’ scenarios.

    “We can start switching from models that say, ‘Here’s how we think this is going to operate’ to ‘let’s learn from what the data is telling us.’ We can create predictive analytics all the way from an individual to global scales and that will help us optimize decisions.”

  1. Text mining and natural language processing could be invaluable.

    “There’s a push to get through the massive amount of scientific literature about coronavirus and related viruses, SARS and MERS specifically, and find the patterns and connections and extract semantic information that will help scientific understanding.

    “There are emerging discoveries that were buried across many articles and often across disciplines. This text summarization also can be applied to mining medical records and understanding the connections between different conditions and treatments.

    “Language processing could also help with actual communication to help people who are lonely have a conversation partner. There’s going to be a huge mental health crisis with this pandemic, and people have started building AI conversation companions, but it’s never been this massive.”

  2. AI’s ability to process visual images and audio has multiple applications we could use.

    “There is still this issue of a false negative for the coronavirus test from the swab, so quite often it’s verified through a CT scan. AI applications can process images. Can we speed up processing CT scans and X-rays and find clues for the disease? This can also support aspects of telemedicine as well, processing clinical data quickly for patient evaluations.

    “Also, our ability to lift restrictions on human mobility will be tightly reliant on testing but also contact tracing (knowing when a person comes in contact with someone infected). AI can use imaging and audio for contact tracing to predict the outcome of interactions. In places such as Singapore and China, they’ve been able to do that by taking all your cell phone data and extracting information. That’s really hard in a society like the United States, which values privacy. So we can ask people to self-report through apps or websites, but it’s going to be limited because of so many privacy issues and reporting biases.”

  3. You want one more way AI can help? How about by making all those Zoom meetings better?

    “Our whole social life is changing. With going virtual in all of our conversations happening on Zoom or Skype or Hangouts, there’s a lot AI can do in making these kinds of conversations more natural. So out of the 20 people on a call, who are you talking to? How do we make that directional conversation more meaningful? AI could support everything from mundane things such as noise cancellation and making us look better on screen to actually facilitating a multi-part conversation in a natural way.”

First Posted to Ohio State Insights Page. To read the original article, click here.

Share this page
Suggested Articles
Grad students Vedula and Maneriker win Web Conference 2020 Best Paper award

OSU graduate students Nikhita Vedula and Pranav Maneriker received the Best Paper award at the Web Conference 2020 (formerly WWW) with their paper “Open Intent Extraction from Natural Language Interactions,”...

Growing Research Opportunities- The GRO Academy

The Office of Research invites self-nominations from all colleges to a new leadership development program: the Growing Research Opportunities (GRO) Academy. This program will equip faculty with skills needed to...

Visual analytics for management: Translational science and applications in practice

An annotated excerpt from the forthcoming text by Elliot Bendoly and Sasha Clark; Taylor & Francis, 2017  Amidst a data ecosystem’s collectors, analysts and users, important meaning can be distorted...

Health and Life Sciences Community of Practice aims to connect computational scientists to subject area researchers in science-related fields

The Computational Health and Life Sciences Community of Practice hosted by TDAI is focused on developing big-data analytics that have translational values and can be deployed in healthcare and life...

New Data Science Career Offerings (November 2020)

As of this November, the Translational Data Analytics Institute has posted three new career offerings onto the jobsatosu site. The new positions, all full-time and revolving around research and data...