| CPC G06N 3/08 (2013.01) [G06F 16/90335 (2019.01)] | 12 Claims |
|
1. A computer-implemented method to generate synthetic data instances using a generative model, the method comprising the following steps:
generating, by the generative model, a synthetic data instance for an input variable value of an input variable supplied to the generative model;
classifying, by a classification model, the synthetic data instance to generate a classification result;
determining, for the classification result, a loss function value of a loss function, the loss function evaluating the classification result;
determining a gradient of the loss function with respect to the input variable at the input variable value;
based on an absolute value of the gradient being smaller than or equal to a predefined threshold:
modifying the input variable value to generate a plurality of modified input variable values,
determining, for each modified input variable value of the plurality of modified input variable values, a gradient of the loss function with respect to the input variable at the respective modified input variable value,
combining the gradients of the loss function with respect to the input variable at the respective modified input variable values to generate an estimated gradient at the input variable value, and
modifying the input variable value in a direction determined by the estimated gradient to generate a further input variable value; and
generating, by the generative model, a further synthetic data instance for the further input variable value.
|