Cobalt Speech and Language, Inc. SBIR Phase I Award, August 2019

Cobalt is developing a Conversational Artificial Intelligence (ConvAI) for use by maintenance technicians to support final inspection procedures for the F-15E engines. - Focus Area: 16 – Immersive Training Solutions. - User Need: 16 – Machine-Assisted Engine Inspection The ConvAI system is an adaptive tool based on deep learning techniques that will guide a technician through a complex series of inspection procedures by interacting with the technician during each step of the inspection. It will also support the use of their voice to ask various questions throughout the process, such as “where is the fuse panel?� The ConvAI system used in this project is built on Cobalt’s ConvAI platform incorporating Natural Language Understanding (NLU), Dialogue Manager (DM), and Natural Language Generation (NLG). The ConvAI is integrated with a custom knowledgebase created from various technical information and detailed inspection procedures. The ConvAI will collect various analytics that can use by machine learning techniques to retrain the ConvAI system to improve its efficiency and accuracy in communicating with the technician. Other analytics collected can be used to monitor the effectiveness of the technician and possibly identify in real-time mistakes a technician might be making. (e.g. “Did you spend enough time on step