Charles River Analytics Inc., developer of intelligent systems solutions, has announced a four-year contract valued over $4.4 million, awarded by the Defense Advanced Research Project Agency (DARPA) as part of its World Modelers program. Charles River will develop technology that helps analysts model national and global security threats—called Sensitivity Analysis of Uncertainty in Causal Environments (SAUCE). The SAUCE technology will facilitate analysis and understanding, leading to improved decision-making that may prevent crises.
The program will initially address food insecurity as a use case. Food insecurity is a growing problem in many parts of the world, and poses both humanitarian and security challenges. For example, South Sudan topped the list of fragile states in this year’s index released by The Fund for Peace. In this conflict-torn country, more than a million people have fled their homes since fighting erupted between different factions of South Sudan’s ruling party. The battles have prevented farmers from planting or harvesting crops, which has caused food shortages nationwide.
“Our SAUCE effort reflects evolving conflict and security patterns, which play out in cities and subdistricts more than ever,” explained Dr. Brian Ruttenberg, Senior Scientist at Charles River. “Our current national and global security tasks, such as—analyze food insecurity in each district of South Sudan two years into the future, recommend specific actions for its mitigation, then analyze the confidence and sensitivity of these recommendations—are too complex for current models and analytic technologies.”
Our effort will be supported by a world-class team of experts that includes Dr. Emanuele Borgonovo from Italy’s Bocconi University, Dr. Alexander Ihler from the University of California at Irvine, and Dr. Alexander Levis from George Mason University.
Under SAUCE, Charles River will develop new technologies for user-guided automation of causal analyses that recommend specific actions to avert crises and report uncertainty and sensitivity in the actions’ effects.
SAUCE will use an expressive representation of complex causal models to provide computationally tractable prediction and planning of optimal actions using novel approximation algorithms. The technology will also guide users through an uncertainty reporting process that unifies local and global sensitivity analyses. SAUCE’s human-machine interface will incorporate intuitive display structures and input tools within a single, coordinated workspace.
Charles River will develop SAUCE using its Figaro™ open-source probabilistic programming language. Our novel technologies will also be useful in commercial applications, such as business modeling, analytics, automated recommendations, and digital transformation.
This material is based upon work supported by the Defense Advanced Research Projects Agency (DARPA) and the Army Research Office (ARO) under Contract No. W911NF-18-C-0015. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the Defense Advanced Research Projects Agency (DARPA) or the Army Research Office (ARO). The appearance of U.S. Department of Defense (DoD) visual information does not imply or constitute DoD endorsement.