System for Embedded ASW Training Using Artificial Intelligence (SEATRAIN)

To provide an Artificial Intelligence (AI) Assistant for P-8A acoustic operators and trainees, Boston Fusion Corp. proposes to demonstrate a System for Embedded ASW Training Using Artificial Intelligence (SEATRAIN). SEATRAIN integrates four core capabilities. First, SEATRAIN’s AI Assistant uses machine learning to coach trainees by showing how their assessments differ from its own assessments and explaining its rationale. The Phase II version will highlight how trainees’ actions differ from its actions and show how proven procedures/techniques guided its actions. Second, SEATRAIN incorporates online learning to help it continually improve its classification performance. Third, SEATRAIN uses machine learning to filter out clutter and perform cluster classification, enabling it to generate timely alerts for valid contacts and anomalous detections—including detections for which there are no reference signatures or training data. SEATRAIN can classify contacts accurately because its computational resources are focused only on relevant detections. Fourth, SEATRAIN is embedded in an extensible fusion architecture that is readily modified to accept new input data (e.g., onboard non-acoustic sensors and external data links) and perform higher levels of fusion/assessment. Adapting SEATRAIN for the operational P-8A ASW mission system will help operators provide more timely, relevant, and accurate tactical information during actual ASW operations.