Title:
トレーニングされた機械学習モデルのパフォーマンスを改善するための方法
Document Type and Number:
Japanese Patent JP6862426
Kind Code:
B2
Abstract:
A method for improving performance of a trained machine learning model includes adding a second classifier with a second objective function to a first classifier with a first objective function. Rather than minimizing a function of errors for the first classifier, the second objective function is used to directly reduce the number errors of the first classifier.
Inventors:
Tarati, Sachin Subash
Baltak, Aniket
Baltak, Aniket
Application Number:
JP2018510504A
Publication Date:
April 21, 2021
Filing Date:
August 11, 2016
Export Citation:
Assignee:
QUALCOMM INCORPORATED
International Classes:
G06N20/00; G06N3/08; G06V10/764
Domestic Patent References:
| JP5290013A | ||||
| JP9185394A | ||||
| JP2011048262A |
Other References:
Geoffrey Hinton et al.,"Distilling the Knowledge in a Neural Network",arXiv.org [online],Cornell University,2015年 3月,arXiv:1503.02531,pp.1-9,[令和2年8月28日検索],インターネット
Attorney, Agent or Firm:
Kurata Masatoshi
Yoshihiro Fukuhara
Morisezo Iseki
Takashi Okada
Nakamaru Yoshihiro
Yoshihiro Fukuhara
Morisezo Iseki
Takashi Okada
Nakamaru Yoshihiro