US Patent No. 10,921,755

METHOD AND SYSTEM FOR COMPETENCE MONITORING AND CONTIGUOUS LEARNING FOR CONTROL


Patent No. 10,921,755
Issue Date February 16, 2021
Title Method And System For Competence Monitoring And Contiguous Learning For Control
Inventorship Nurali Virani, Niskayuna, NY (US)
Abhishek Srivastav, San Ramon, CA (US)
Assignee GENERAL ELECTRIC COMPANY, Schenectady, NY (US)

Claim of US Patent No. 10,921,755

1. A competence modelling and control system comprising:a competence module to receive data from at least one data source;
a memory for storing program instructions;
a competence processor, coupled to the memory, and in communication with the competence module, and operative to execute program instructions to:
receive an objective;
select a machine learning model associated with the objective;
receive data from the at least one data source;
generate at least one next input based on the received data, wherein the at least one next input is distinct from the received data;
determine whether the at least one next input is in a competent region of training data of the machine learning model or is in an incompetent region of the machine learning model, wherein the competent region is a region in which the machine learning model is competent to make predictions;
when the at least one next input is inside the competent region, generate an output of the machine learning model;
determine an estimate of uncertainty for the generated output of the machine learning model;
when the uncertainty is below an uncertainty threshold, determine the machine learning model is competent and when the uncertainty is above the uncertainty threshold, determine the machine learning model is incompetent; and
operate the physical asset based on the state of the machine learning model, where the state is one of competent and incompetent.