| CPC G06F 18/2163 (2023.01) [G06F 18/2113 (2023.01); G06F 18/2431 (2023.01); G06N 3/08 (2013.01); G06N 5/01 (2023.01); G06N 20/20 (2019.01)] | 17 Claims |
|
1. A computer-implemented method (CIM) comprising:
receiving a full training data set including a plurality of individual training data set, with each individual training data set including a plurality subset of individual training data sets, with each individual training data set including data relating to a historical instantiation of a situation that is relevant to training;
dividing the plurality of individual training sets into N classes, where N is an integer greater than three;
dividing the N classes into M full data classes and N-M partial data classes;
training a fixed size machine learning (ML) classification model to obtain a trained fixed size machine learning (ML) classification model and a trained in-class confidence model, wherein training classes are split to each computing node, including the following types of training:
training a fixed sized machine learning (ML) classification model using the individual training data sets in the full data classes,
feeding a full data of assigned classes into a fixed-sized ML classification model;
feeding the full data of assigned classes and partial data of other classes into the in-class confidence model;
training an in-class confidence model using the individual training data sets in the full data classes, and
training the in-class confidence model using the individual training data sets in the partial data classes, wherein the in-class confidence model calculates a set of prediction values and associate a set of confidence values;
outputting a first set of prediction value(s) based on the performance of training, and a respectively corresponding first set of confidence value(s);
distributing each class of the N classes of individual training data sets to a different node of a distributed machine learning system; and
applying the distributed classes to generate and output, from the nodes of the distributed machine learning system, a second set of prediction value(s) for each class of the N classes based on the performance of training, and a respectively corresponding second set of confidence value(s), with the second set of prediction value(s) and the second set of confidence value(s) being determined by the trained in-class confidence model and the trained ML classification model.
|