Charles River Analytics building tools for decision-making under uncertain conditions
NewsJuly 25, 2017
CAMBRIDGE, Mass. Intelligent-systems developer Charles River Analytics has received a one-year contract extension -- awarded by the Defense Advanced Research Projects Agency (DARPA) and valued at close to $500,000 -- for a system supporting Probabilistic Model-Based Programming Techniques for Prediction, Analysis, and Control, or PROMPT. The system, says the company, is intended for use by system engineers to reason under uncertain conditions.
Charles River Analytics says that PROMPT is designed to extend model-based systems engineering (MBSE) languages -- widely used in software design to add safe and effective performance features to mission-critical systems -- with probabilistic programming capabilities. According to the company, PROMPT enables engineers to link existing MBSE language elements to model-based probabilistic programming language constructs, which produces a probabilistic program that may be executed and analyzed, thereby allowing engineers to reason about the behaviors of systems operating under uncertainty, such as a minesweeping system that relies on noisy sensors for target identification.
Dr. Avi Pfeffer, chief scientist at Charles River Analytics, says of the effort: "We’re helping engineers create system models to reason about the operating behavior of systems in uncertain environments. We’re developing probabilistic extensions to the standard SysML systems modeling language, which we compile to our Figaro probabilistic programming language, allowing engineers to explore potential outcomes and tradeoffs through plugins to their systems engineering tools."
