Self-Healing Data

Maintenance and supply data associated with Navy platforms provides essential information used by numerous communities in planning, logistics, and all maintenance disciplines. However, poor data quality due to errors in the data capture process is hindering the Navy's ability to gain vulnerable and accurate insight from their data. Illumination Works (ILW) offers the Navy a novel solution, employing cutting-edge natural language processing and machine learning techniques, for automatically correcting errors in their data systems. In this Phase I SBIR effort, ILW will build upon our existing TRL 4 solution developed for the Air Force, to augment the Automated Data Cleansing and Analysis Tool (ADCAT) capable of self-healing errors in part numbers and other coded fields within the Navy's data systems and provide root cause analysis for part failures across two Navy Platforms. ADCAT offers a robust solution that leverages natural language processing and machine learning techniques applied to text narratives that capture information relevant for correcting error-prone data. ADCAT will significantly improve the accuracy of Navy data and improve trending of part defects as well as the underlying cause of repeated part/equipment failure.