Systems and methods for self-learning artificial intelligence of things (AIoT) devices and services

A self-learning system for predicting a service event for an industrial asset from sensor data, comprising:

• one or more edge devices mounted on or proximate to the industrial asset, the edge devices comprising one or more sensors, a local machine learning inference engine, a wireless communication system, a memory, and a power source;
• one or more signal processors including a narrow-band classifier and a machine-learning classifier;
• the one or more edge devices are configured to determine using the machine learning inference engine that an anomaly is detected in data received from the one or more sensors, and communicate with the one or more cloud-based processors to provide at least some of the data;
• the one or more signal processors configured to receive at least some of the data and attempt to classify the anomaly using the narrow-band classifier and the machine learning-classifier;
• upon classification of the anomaly by the signal processors, determine alerts, corrective actions, and/or maintenance requests associated with the anomaly, and retrain the machine learning inference engine based at least in part on the classification by the signal processors;
• if the signal processors fail to classify the anomaly, communicate with a technician to investigate the anomaly, and upon classification of the anomaly by the technician or determination that the anomaly is a false positive, retraining the machine learning inference engine based at least in part on the classification or false positive determination by the technician.