The Defense Advanced Research Projects Agency has named the research teams to develop artificial intelligence tools and approaches designed to help subject matter experts automate knowledge discovery, model creation and simulation.
The selected teams will work on the four major technical areas of the Automating Scientific Knowledge Extraction and Modeling program, which is divided into two phases that will both run for 21 months, DARPA said Friday.
Phase 1 of the ASKEM initiative will focus on enhancing the accuracy, maintainability and timeliness of expert models, while phase 2 will center on the models’ generalizability and scalability.
The selected teams are:
- Harvard Medical School
- Julia Computing, Massachusetts Institute of Technology
- Lum.ai, University of Arizona, SIFT, University of Utah
- Pacific Northwest National Laboratory, Basis, University of Texas at Austin
- Uncharted Software Inc., Jataware
- University of Florida, University of Colorado Denver, Topos Institute
- University of Pittsburgh
- University of Wisconsin-Madison’s Morgridge Center for Public Service, ETH Zurich
The teams will demonstrate the effectiveness of the models and simulators for various use cases such as viral epidemics and the causes and impacts of space weather.