Soliloquy Ph II

Automatic Speech Recognition (ASR) allows trainees to practice verbal communication skills in representative training environments without the need for human role players. However, these technologies have yet to become a feature of most training systems. Building on Phase I work, SoarTech along with our academic partners at UC Davis, proposes developing Soliloquy, a novel interface that allows non-expert users to customize ASR systems for Department of Defense training applications. Soliloquy offers an interactive “human-in-the-loop” paradigm for developing, editing, and customizing grammatical constructs and investigates big data approaches for developing ASR systems from raw source data.