SAS launches free AI-driven environment to mine COVID-19 research

Globally, thousands of scientists are working tirelessly on COVID-19 mitigation efforts, including the search for new treatments and a vaccine. Analytics leader SAS brings a powerful resource to the fight with COVID-19 Scientific Literature Search and Text Analysis, a free visual text analysis environment that uses artificial intelligence (AI) and machine learning to quickly search tens of thousands of research articles on COVID-19 and deliver potentially lifesaving answers to these scientists.

Leading research groups have gathered and released to the public more than 50,000 full-text scientific research articles on COVID-19 and other coronaviruses through the COVID-19 Open Research Dataset (CORD-19). The articles include studies on treatment effectiveness, vaccine development, mitigation efforts, genetic analysis, economic impact and more. With so much scientific literature available, it’s impractical – if not impossible – to analyze it all manually.

“Effectively mining unstructured text from scientific literature can take teams of people, many needing deep subject matter expertise, and substantial amounts of time to effectively categorize and determine relevancy,” said 2000

Drawing on AI, natural language processing, linguistic rules and sophisticated modeling techniques, SAS’ COVID-19 Scientific Literature Search and Text Analysis environment enables quick and intelligent extraction of relevant text and numerical data from CORD-19. The free and publicly available environment aims to quickly and efficiently connect the global research community with the most relevant scientific literature through its self-guided, cloud-based system.

“SAS has put forth a solution that expedites this process and enables researchers across the globe to support their COVID-19 related efforts such as understanding the effectiveness of treatments or better understanding genetic variables in COVID mutations,” said Cullen.

With SAS’ new visual text analysis environment, users can interactively explore relevant research on coronavirus topics such as incubation period, genetic variations, risk assessment and more. They can also visualize extracted keywords and summarized quantitative data, quickly identify co-citations and the authority of papers using network analysis visualization, and search for key terms in free text.

“The COVID-19 pandemic has led to an explosion of medical and scientific research, far more than any one person can meaningfully consume,” said Jeffrey Engel, MD, of the Council of State and Territorial Epidemiologists. “The SAS environment, using its advanced analytic and AI capabilities, provides global researchers, public policy analysts, epidemiologists, healthcare providers and academia the ability to focus on topics that are relevant, sources that are authoritative, and findings that are impactful in their efforts to stop COVID-19.”

SAS^® Viya^® powers the interactive environment through advanced analytics and AI methods that help researchers explore the CORD-19 dataset. Using SAS Visual Text Analytics and SAS Visual Data Mining and Machine Learning, SAS augments the dataset with models developed by linguists and health and life sciences experts. Researchers can explore topics such as PPE effectiveness, social distancing efficacy and the transmission environment. The models also extract and visualize quantitative data, such as the incubation period and reproduction number of COVID-19.

“With COVID-19 research moving at high volume and velocity it can be a monumental challenge to connect researchers with the relevant literature,” said Sajung Yun, PhD, Adjunct Faculty at Johns Hopkins University and CEO of ZtoMed. “SAS has leveraged some of the most innovative AI techniques to intelligently sift through the body of data and connect users with relevant research in an intuitive and expeditious fashion.”

“SAS is committed to using advanced technology in innovative ways to help customers across every industry combat the pandemic,” said Bryan Harris, Senior Vice President of Engineering at SAS. “Our new AI-powered environment goes beyond current offerings in the market and provides an intuitive, visual way to find complex connections in research. Because we’re not only relying on machine learning, but also human-in-the-loop approaches, users can view different data slices to find exactly what they need to support their own research needs.”

As part of its COVID-19 response, SAS has also released the COVID-19 Epidemiological Scenario Analysis, an interactive environment that builds on medical resource optimization models created jointly with Cleveland Clinic. The models run different virus-projection scenarios to predict the impact of a disease outbreak and quantify the effectiveness of public health mitigation strategies.

Additionally, SAS’ dedicated COVID-19 Data Analytics Resources Hub includes free analytical models, a public dashboard to monitor the spread of the pandemic, a data-discovery environment built on SAS Viya, and access to free training.