Methods and Systems for the Measurement of Relative Trustworthiness for Technology Enhanced With AI Learning Algorithms
AssigneeNuEnergy.ai
InventorsNiraj BHARGAVA, Fred SPECKEEN, Evan W. STEEG, Jorge DELIGIANNIS, Gaston GONNET
Life of this patent
Estimating the term…
Abstract
The present disclosure provides a system and method for use in evaluating the trustworthiness of an artificial intelligence (AI) process. The trustworthiness evaluation may include both automated and manual evaluation of the AI process. Further, the system is provided with functionality for automatically evaluating and modifying the criteria used in evaluating the AI process.
Claims21
A method for updating evaluation criteria used in determining a machine trust index of artificial intelligence processes, the method comprising:
- determining a plurality of evaluation criteria from a pool of evaluation criteria for use in generating a machine trust index (MTI) score for an artificial intelligence (AI) process, each of the evaluation criteria associated with an evaluation process to be used in evaluating the respective evaluation criteria;
- evaluating each of the plurality of determined evaluation criteria according to the evaluation process associated with the respective evaluation criteria;
- generating the MTI based on the evaluated evaluation criteria and storing in an MTI database the MTI in association with a time the MTI was generated;
- using a plurality of generated MTIs stored in the MTI database, determining which evaluation criteria in the pool of evaluation criteria do not contribute to the generated MTI;
- and
- removing the evaluation criteria determined not to contribute to the MTI scores from the evaluation criteria pool.
A system for updating evaluation criteria used in determining a machine trust index of artificial intelligence processes, the system comprising:
- at least one processing unit configured to execute instructions;
- and
- at least one computer readable memory unit communicatively coupled to the at least one processing unit, the computer readable memory unit storing instructions, which when executed by one or more of the at least one processing unit configure the system to:
- determine a plurality of evaluation criteria from a pool of evaluation criteria for use in generating a machine trust index (MTI) score for an artificial intelligence (AI) process, each of the evaluation criteria associated with an evaluation process to be used in evaluating the respective evaluation criteria;
- evaluate each of the plurality of determined evaluation criteria according to the evaluation process associated with the respective evaluation criteria;
- generate the MTI based on the evaluated evaluation criteria and storing in an MTI database the MTI in association with a time the MTI was generated;
- using a plurality of generated MTIs stored in the MTI database, determine which evaluation criteria in the pool of evaluation criteria do not contribute to the generated MTI;
- and
- remove the evaluation criteria determined not to contribute to the MTI scores from the evaluation criteria pool.
A non-transitory computer readable memory storing instructions, which when executed by a processing unit of a system configure the system to:
- determine a plurality of evaluation criteria from a pool of evaluation criteria for use in generating a machine trust index (MTI) score for an artificial intelligence (AI) process, each of the evaluation criteria associated with an evaluation process to be used in evaluating the respective evaluation criteria;
- evaluate each of the plurality of determined evaluation criteria according to the evaluation process associated with the respective evaluation criteria;
- generate the MTI based on the evaluated evaluation criteria and storing in an MTI database the MTI in association with a time the MTI was generated;
- using a plurality of generated MTIs stored in the MTI database, determine which evaluation criteria in the pool of evaluation criteria do not contribute to the generated MTI;
- and
- remove the evaluation criteria determined not to contribute to the MTI scores from the evaluation criteria pool.
Dependencies are inferred from the claim text.