Apply

Computer Science Innovation and COVID-19: Emerging Solutions

At every level, COVID-19 has had a profound impact on government, education, our ability to earn a living, and, of course, upon human health and wellness. Further, between bombast and misinformation, it’s easy to overlook the significant innovation of data and computer scientists in combatting the novel coronavirus.

 

The truth is that innovation abounds, particularly in the areas of COVID-19 testing and treatment. Here, I will share some of those innovations, and, perhaps, a bit of hope for a safer, healthier future.

 

 

Oxford Scientists Tap Machine Learning to Detect SARS-CoV2 

 One example of innovation in testing technology comes to us from Oxford’s Department of Physics, where they have developed an accurate COVID-19 test capable of detecting SARS-CoV2 directly in patient samples using an approach based on machine learning. Far from being “just another” testing method, it offers numerous advantages, including:

  • Detecting actual virus particles, rather than relying on the presence of antibodies or other indirect signs 
  • Delivering results in fewer than five minutes 
  • Delivering rapid results without the need for sample preparation 
  • Playing a critical role in the potential development of mass testing technology 
  • Unique flexibility in testing for other pathogens 

 According to Tech Crunch contributor Darrell Etherington, “The technology that makes this possible works by labeling any virus particles found in a sample collected by a patient using short, fluorescent DNA strands that act as markers. A microscope images the sample and the labeled viruses present, and then machine learning software takes over using algorithmic analysis developed by the [Oxford] team to automatically identify the virus, using differences that each one produces in terms of its fluorescent light emitted owing to their different physical surface makeup, size, and individual chemical composition.”  

 

In an article published on medRxiv, the development team explained, “[Our approach] uses a convolutional neural network to distinguish between microscopy images of single intact particles of different viruses. Our assay achieves labeling, imaging, and virus identification in less than five minutes and does not require any lysis, purification, or amplification steps. The trained neural network was able to differentiate SARS-CoV-2 from negative clinical samples, as well as from other common respiratory pathogens such as influenza and seasonal human coronaviruses, with high accuracy. Single-particle imaging combined with deep learning offers a promising alternative to traditional viral diagnostic methods, and has the potential for significant impact. 

 

Currently, the team is attempting to form a commercial enterprise responsible for commercializing the technology, with product development expected in early 2021. However, the Oxford team is just one of many working to provide humanity with powerful new testing and identification tools. 

 

 

DePaul University Helps Narrow Race and Ethnicity Data Gap in COVID-19 Testing 

 COVID-19 testing is challenging at the best of times. However, it has proven increasingly difficult for testers to capture racial, ethnic, and other demographic-related information. The unfortunate side effect of this is a significant percentage of tests from patients of an “unknown race.” 

 

Identifying accurate race and ethnicity information is critical for developing a robust, accurate image of SARS-CoV2 and how its effects may vary from individual to individual across different races and ethnicities. Without this critical missing piece, scientists and researchers are forced to make assumptions. 

 

It is not only about correcting potentially erroneous assumptions, though. Missing race and ethnicity data deepen already-significant inequity, skewing the provision of and access to critical resources, such as PPE and testing sites, education and outreach efforts. 

 

However, data science researchers at DePaul University may have developed a solution to help fill in the “gaps.” In testing within Chicago, Illinois, the team was able to reduce the category of “unknown” race from 47% of tests to just 11%.  

 

According to  a press release issued by DePaul University, the team closed the data gap using a novel approach to imputing data. Data science professor Daniela Stan Raicu and her research team at the Center for Data Science used an algorithm to analyze U.S. Census data and available demographic information. They can predict an individual in Chicago’s race and ethnicity with 81% accuracy, according to Raicu. The team also developed a mobile application that allows city officials to easily and securely input the data with missing values. 

 

While this method has only been used in Chicago, the team continues to research methodologies that would allow it to be used in other cities with similar accuracy 

 

 

MIT Hosts AI-Cures Conference Exploring Machine Learning Approaches to COVID-19 Treatment 

 A final example comes to us from researchers from the Jameel Clinic at the MIT Stephen A. Schwarzman College of Computing, who say the ongoing COVID-19 pandemic is a prime opportunity to leverage AI technologies. On Sept. 29, representatives from 50 countries and 70 organizations joined the virtual AI Cures Conference: Data-driven Clinical Solutions for Covid-19. The goal of the conference, according to organizers, was to gather “AI researchers who develop novel methods targeting COVID-19 and clinicians who utilize them in patient care” to explore whether and how AI technology is improving COVID-19 treatment. 

 

Daniel Huttenlocher, dean of the MIT Schwarzman College of Computing, said in his welcoming remarks that “AI in health care is moving beyond the use of computing as just simple tools, to capabilities that really aid in the processes of discovery, diagnosis, and care. Attendees then heard speakers discuss technologies developed in response to the COVID-19 pandemic — from a wireless device allowing clinicians to track COVID-19 patients from a distance, to a machine learning model that identifies individuals at higher risk for needing a ventilator before they require intubation. 

 

AI Cures at Jameel Clinic is leading a broad range of efforts to bring AI solutions to the COVID-19 pandemic, from recommendations on machine learning tools for researchers to try, to large, open-source datasets with which computational experts can work.  

 

 

Hope and Progress 

 While the global pandemic continues to take an unprecedented toll, both in human lives and economic activity, there is hope on the horizon. Innovative technologies are currently in development not just across the United States, but around the world. It is exactly that united front and commitment to cutting-edge exploration that will help us find solutions that protect lives and enable a return to some semblance of normalcy. 

 

Sources: