Details

Autor(en) / Beteiligte

• Srinivasan, Deepak S.
• Pruthi, Rohit

Titel

Scarecrow - Intelligent Annotation platform for Engine Health Management

Ist Teil von

• Proceedings of the First International Conference on AI-ML Systems, 2021, p.1-4

Ort / Verlag

New York, NY, USA: ACM

Erscheinungsjahr

2021

Quelle

ACM Digital Library

Beschreibungen/Notizen

• Engine Health Management (EHM) in the context of applications such as gas turbines, power packs is dependent on massive amount of data captured by onboard sensors. The data streams are then processed to extract key points and trends capturing events related to failures and deterioration, which subsequently need to be enhanced by insights and judgements from Subject Matter Experts (SME). Data volumes and extremely demanding requirements, commercial and regulatory, cause human efforts to quickly emerge as bottleneck in EHM service delivery. We have developed an intelligent data annotation platform called Scarecrow which records SME inputs, generates machine learning models in near real-time which are then deployed into analytic pipelines for EHM diagnostics. Scarecrow enables seamless ML ops strategy through human assisted feature learning, model building and deployment