SBIR-STTR Award

In order for manufacturers to adopt data-driven decision making, they must tap into unstructured data that sits unused. The process of annotation and knowledge extraction from unstructured text data is a time consuming and challenging bottleneck to adopting machine learning at scale. This Phase I SBIR proposes implementation of an enhanced cloud-based platform enabling maintenance and other organizational users to tag and extract knowledge from large volumes of unstructured text data more efficiently. In Nestor, NIST provides an NLP toolkit that assigns tags and rules to unstructured data by a ranked tagging procedure. Combining this approach with RedShred’s state of the art enrichment platform will accelerate an organization’s ability to rapidly develop and deploy these models to support evolving organizational needs. The extended system, ENT, will include customizable ranking allowing organizations to tailor which data are prioritized based on business value as well as interactive UI dashboards providing real-time feedback. This combined system provides a valuable low-friction solution to accelerate adoption of NLP technologies in manufacturing to unlock value from previously idle log data repositories.

In order for manufacturers to unlock the promise of AI for data-driven decision making, they need visibility into large volumes of unstructured knowledge trapped in technical language like work order notes. The process of annotation and knowledge extraction from unstructured technical language data is currently a time consuming and challenging bottleneck to adopting machine learning at scale. In Nestor, NIST provides an NLP toolkit that assigns tags and rules to unstructured data by a ranked tagging procedure. Combining this approach with RedShred’s state of the art enrichment platform accelerates an organization’s ability to rapidly develop and deploy these models to support evolving organizational needs. In Phase I we validated an enhanced cloud-based implementation of this combined platform. In Phase II we will add enterprise management and analysis capabilities to facilitate deploying this solution at-scale. This combined system provides a valuable low-friction solution to accelerate adoption of NLP technologies in manufacturing to unlock value from previously idle technical language data repositories.