Method of controlling for undesired factors in machine learning models

A computer system configured to train and use a neural network to evaluate an insurance applicant as part of an underwriting process, wherein the neural network controls for consideration of one or more undesired factors which might otherwise be considered by the neural network, the computer system comprising one or more processors configured to:

• train the neural network to probabilistically correlate an aspect of appearance with a health-related characteristic by providing the neural network with a training data set of images of individuals having known health-related characteristics, including the one or more undesired factors;
• identify the one or more undesired factors;
• identify one or more relevant interaction terms between the one or more undesired factors;
• receive via a communication element an image of the insurance applicant;
• analyze with the neural network the image of the insurance applicant to probabilistically determine the health-related characteristics for the insurance applicant, wherein such analysis excludes the identified one or more undesired factors;
• and
• suggest or recommend with the neural network an appropriate insurance premium based at least in part on the probabilistically determined health-related characteristic but not on the one or more undesired factors to control for undesired prejudice or discrimination in machine learning models.

A computer-implemented method for training and using a neural network to evaluate an insurance applicant as part of an underwriting process, wherein the neural network controls for consideration of one or more undesired factors which might otherwise be considered by the neural network, the computer-implemented method comprising, via one or more processors:

• training the neural network to probabilistically correlate an aspect of appearance with a health-related characteristic by providing the neural network with a training data set of images of individuals having known health-related characteristics, including the one or more undesired factors;
• identifying the one or more undesired factors;
• identifying one or more relevant interaction terms between the one or more undesired factors;
• receiving via a communication element a selfie image of the insurance applicant taken with a smartphone and transmitted via a wireless communications network;
• analyzing with the neural network the selfie image of the insurance applicant to probabilistically determine the health-related characteristics for the insurance applicant, wherein such analysis excludes the identified one or more undesired factors;
• and
• suggesting with the neural network an appropriate insurance premium based at least in part on the probabilistically determined health-related characteristic but not on the one or more undesired factors to control for undesired prejudice or discrimination in the neural network.