TEACHR: Technical Explanation of Algorithmic Choice Hierarchy and Reasoning

Today, predictive analytics solutions provide workflow automation and data connectivity to enable maintenance recommendations. There is a USAF-identified defense need for explainability of the resultant recommendations for decision support, enabling human users to understand, appropriately trust, and effectively manage the emerging generation of machine learning outputs. Questions to explain predictive machine analytics to humans include: - Why this recommendation and not something else? - What are the criteria for success or failure? - Where are humans in the loop required? - How do we course correct for shifting mission priorities? - How can we trust this data? - How can we trust this analysis? Hence, the need to enhance Colvin Run’s predictive analytics solutions with Technical Explanation of Algorithmic Choice Hierarchy and Reasoning, or TEACHR. We will leverage the latest research in Explainable Artificial Intelligence (XAI) in this Phase I feasibility study, with our current and prospective USAF customer and end-user involvement, enhancing our unique portfolio of predictive analytics solutions powered by market-leading commercial business intelligence software. The STTR effort will include partnership and research from the University of Colorado Boulder Department of Information Science.