Human-centric incorporation of domain knowledge for predictive maintenance (PMx)

AI stands to learn a lot from human knowledge. This project will focus on formal identification and definition of subject matter expert (SME) knowledge elements, including rules-of-thumb and operational norms that significantly influence model building for predictive maintenance (PMx) contexts. It will also require development and assessment of methods for connecting SME knowledge to the AI modeling process.

“What did the model actually learn?” is a longstanding, open-ended question in AI. By distilling core knowledge from humans (e.g. the law of conservation of momentum) and incorporating this knowledge in the AI design process, either during training or during model evaluation, we will hopefully lead to more trustworthy systems because we have proof that an AI is making decisions in a manner similar to the way that a trustworthy human adjudicator would make their decisions.