Training machine learning models

The method facilitates the integration of new estimators into ML systems through network connectivity and unified handling, enhancing model accuracy and adaptability by optimizing ML models for specific tasks.

JP7870795B2Active Publication Date: 2026-06-05INTERNATIONAL BUSINESS MACHINE CORPORATION
View PDF 3 Cites 0 Cited by

Patent Information

Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
INTERNATIONAL BUSINESS MACHINE CORPORATION
Filing Date
2022-06-07
Publication Date
2026-06-05

Smart Images

  • Figure 0007870795000001
    Figure 0007870795000001
  • Figure 0007870795000002
    Figure 0007870795000002
  • Figure 0007870795000003
    Figure 0007870795000003
Patent Text Reader

Abstract

A method, computer system, and computer program product for training a machine learning model. The method includes connecting a machine learning system to a network and receiving, by the machine learning system, a new estimator and each document not included in the estimator list over the network. The method also includes adding the new estimator to the list stored in the memory. The method further includes reading the documents and providing each extracted data to a machine learning process tool, and fitting, by the machine learning process tool, at least one training dataset of the group of training datasets to the new estimator based on the extracted data. Finally, the method includes training at least a subset of the machine learning models using the new estimator with the at least one training dataset as input, the training resulting in an output.
Need to check novelty before this filing date? Find Prior Art