US 11,966,819 B2
Training classifiers in machine learning
Qingzi Liao, White Plains, NY (US); Yunfeng Zhang, Chappaqua, NY (US); Michael Desmond, White Plains, NY (US); and Rachel Katherine Emma Bellamy, Bedford, NY (US)
Assigned to International Business Machines Corporation, Armonk, NY (US)
Filed by INTERNATIONAL BUSINESS MACHINES CORPORATION, Armonk, NY (US)
Filed on Dec. 4, 2019, as Appl. No. 16/703,371.
Prior Publication US 2021/0174239 A1, Jun. 10, 2021
Int. Cl. G06N 20/00 (2019.01)
CPC G06N 20/00 (2019.01) 17 Claims
OG exemplary drawing
 
1. A method for training classifiers used in machine learning, the method comprising:
receiving, by one or more processors of a computer system, a corpus of training data;
generating, by the one or more processors of the computer system, one or more clusters of the training data according to features of the training data;
comparing, by a cluster refining rule elicitation module of the computer system, neighboring clusters of the one or more clusters to automatically extract boundary rules for suggestion of a user for selection, confirmation and/or editing;
interacting with the user, by the cluster refining rule elicitation module of the elicit user-specified rules by suggesting the extracted boundary rules;
automatically refining, by the one or more processors of the computer system, the one or more clusters using the user-specified rules, wherein the refining further includes: assigning subsets of the corpus of training data in or out of clusters of the one or more clusters using the user-specified rules, and using a generative model to resolve conflicts in the assigning; and
training, by one or more processors of a computer system, multiple classifiers for use in machine learning based upon the refined one or more clusters.