The pace of scientific discovery and publication is faster today than it ever has been. For instance, the number of new, peer-reviewed research publications in PubMed, a repository of abstracts in biomedical literature, has more than doubled in the past two decades. It’s a monumental task simply to monitor the complex landscape of scientific communications, let alone make informed decisions based on new research.
That’s why Battelle scientists invented a technology that uses artificial intelligence and natural language processing to identify and search concepts through texts. The first program based on this technology is a new search engine called Continuous Automated Scanning for Technology Transformation (CASTT). It was developed for the Intelligence Advanced Research Projects Activity (IARPA), a federal organization that invests in research programs that confront the intelligence community’s most difficult challenges.
The CASTT technology was developed by a multi-disciplinary team of Battelle’s Health Analytics Division. It puts the analyst at the center of this new forecasting system. “It’s a search engine that uses machine learning instead of keyword searches and can make predictions about policy changes or scientific disruptions,” said Data Scientist Cedric Neumann who was involved in the development of the technology.
Battelle’s CASTT monitors scientific publications and adds thousands of new articles and papers per week as they are published, automatically updating search results and predictions, and notifying users via a dashboard. “We believe it’s more accurate and more relevant than keyword searches,” said Neumann. “Papers are read by Battelle language processing technology and the concepts are stored in a database that is searchable.”
For example, during a pandemic, analysts may want to predict whether there will be a policy decision or scientific breakthrough coming soon. A user could set the program to specifically monitor domains related to biomedical technology to predict the availability of vaccines or treatments, or related to spread and symptoms to predict policy-decisions.
Battelle has several spinoffs in development of the original technology for government, defense, and commercial clients so that when a radical breakthrough is on the near horizon, they can see it coming and make important proactive decisions instead of being on their heels.