SparkCognition, Inc. — Department of Defense SBIR Phase I: N193-A01

Description

SparkCognition, an artificial intelligence and machine learning company, can provide a system capable of analyzing data related to operations, maintenance and sustainment for twelve (12) LPD-17 SAN ANTONIO class main propulsion diesel engines. SparkCognition’s SparkPredict is a turnkey solution that analyzes sensor data and uses machine learning to return actionable insights, flagging suboptimal operations and identifying impending failures before they occur. SparkCognition’s product, SparkPredict® is able to consume unlabeled data (or data without known failures and states, also known as unsupervised learning) and perform an ensembled, automated clustering technique. After specific clusters are identified, users are able to classify them via the user interface. SparkPredict is also capable of handling labeled data (Data with known failures, also known as supervised learning). It does this by leveraging automated classification and regression algorithms to optimize for the fitness of any new data in a streaming format. In the same way as the clustering algorithms, these classifications can be relabeled and modified within the user interface to retrain the system.