Method for reducing bias in deep learning classifiers using ensembles

A non-transitory computer readable medium, storing instructions for generating a model ensemble to reduce bias and control operation of a physical system, the instructions comprising:

• receiving sensor data from the physical system;
• training a plurality of machine learning models from data comprising the sensor data from the physical system, each of the plurality of machine learning models trained from a first subset of the data and validated from a second subset of the data, each of the first subset and the second subset being different for each of the plurality of machine learning models;
• determining accuracy of each of the plurality of machine learning models based on validation against the second subset of the data;
• pruning the plurality of machine learning models based on the accuracy to generate a subset of the plurality of machine learning models;
• and
• forming the model ensemble from the subset of the plurality of machine learning models;
• deploying the model ensemble to predict maintenance actions for the physical system based on the sensor data;
• and
• automatically controlling the physical system based on predictions from the model ensemble.

An apparatus for controlling operation of a physical system, comprising:

• a processor, configured to:
• receive sensor data from the physical system;
• train a plurality of machine learning models from data comprising the sensor data from the physical system, each of the plurality of machine learning models trained from a first subset of the data and validated from a second subset of the data, each of the first subset and the second subset being different for each of the plurality of machine learning models;
• determine accuracy of each of the plurality of machine learning models based on validation against the second subset of the data;
• prune the plurality of machine learning models based on the accuracy to generate a subset of the plurality of machine learning models;
• and
• form a model ensemble from the subset of the plurality of machine learning models;
• deploy the model ensemble to predict maintenance actions for the physical system based on the sensor data;
• and
• automatically control the physical system based on predictions from the model ensemble.