WellAI Develops COVID-19 Research Tool in Response to White House’s Call to Action
SHERIDAN, Wyo., April 8, 2020— WellAI data scientists released a software application for COVID-19 researchers based on machine learning algorithms able to read and summarize vast amounts of medical literature. This artificial intelligence-powered application, available at https://wellai.health, can help scientists quickly find promising research pathways related to COVID-19 based on any set of medical concepts. It outlines possible research opportunities and shows the research studies that were used by the model to generate the suggestion.
Daniel Satchkov, the lead data scientist on the WellAI project and winner of prestigious Peter L. Bernstein award for scientific research in mathematical finance explained: “We have dedicated our most powerful technology and data science resources towards fighting COVID-19, because we recognize the urgency of the pandemic. Building such an algorithm normally takes up to a year, but we were able to do so using our broader WellAI machine learning analytics that are due to be released in May of 2020. This tool compiles and analyzes over 25 million studies to improve health outcomes. So, we were able to repurpose that work to create a very specific COVID-19 app for scientists in 4 weeks of intense effort.”
On March 16, 2020, the White House Office of Science and Technology Policy, alongside leaders in research and technology, announced “Call to Action to the Tech Community on New Machine Readable COVID-19 Dataset” (CORD-19). CORD-19 is a resource of over 45,000 scholarly articles, including over 33,000 with full text, about COVID-19, SARS-CoV-2, and related coronaviruses. This call to action was addressed to the world’s artificial intelligence experts to develop text and data mining tools to help medical community with urgent scientific questions.
“By offering this unique COVID-19 research tool, the WellAI team would like to give back to the global community. Every day counts. If we could save researchers time and perhaps point them in the right direction in the fight against COVID-19 and future pandemics, the work we have put into creating this tool will have been worth it,” Sergei Polevikov, CEO of WellAI, stated.
About WellAI’s COVID-19 Software
WellAI researchers have prioritized the creation of a model that is run specifically on studies related to the novel coronavirus. The current models consist of four neural networks that work by reading all of the 45,000+ medical studies that are available as part of the White House’s CORD-19 initiative. WellAI’s COVID-19 model and software work as follows. Researchers enter any set of concepts that they are interested in. WellAI’s COVID-19 software will show the most likely-related medical concepts that have been used in conjunction with the entered set of concepts. COVID-19 is pre-entered as a concept. The result will be a list of related concept categories (below is a sample):
These are the concepts that are most related to COVID-19 by category, ranked according to their significance in relation to COVID-19. There is long list of categories such as Genes, Amino Acids, Chemicals, Treatments, Nucleic Acids, Cells, Viruses, Diseases, Pharmacologic Substances, Molecular Functions, Immunologic Factors, Receptors, Bacteria and so on. These concepts are partially based on NIH’s SemRep concepts and are partially derived through machine learning modeling. The screenshot shows only two category results, while the full result list is more extensive.
The researcher can then click on any concept to add it to the Analyze Concepts list. For example, let’s say a researcher is interested in a Hmgn4 Protein, Human, in relation to COVID-19. Clicking the plus sign will produce a new recalculated set of results:
Note that the Gene and Amino Acid group results are now different. Likewise, all results have changed, because we are now looking for concepts related to COVID-19 and Hmgn4 Protein, Human, not just COVID-19 by itself. This process can be repeated many times. WellAI’s high performance computers will recalculate the model on the fly.
Lastly, a click on any concept will produce a list of relevant articles that the model used to make the inference. For example, if we now click on ATM Gene, we will get a list of medical studies that used all three concepts: COVID-19, Hmgn4 Protein, Human, and ATM Gene, with selected concepts highlighted.
WellAI, LLC ( https://wellai.health ) is a health and technology company dedicated to empowering people in pursuit of wellness through the use of AI models and blockchain databases. WellAI apps are able to accept medical findings, ask follow up questions, and suggest evidence-based paths for improving health outcomes. WellAI personal health applications are not meant to replace doctors but can serve as an additional resource for those who want to understand and improve their health using evidence-based science. Contact phone: 307-278-1819.