Title:
METHODS AND DEVICES FOR ENSURING THE REPRODUCIBILITY OF SOFTWARE SYSTEMS
Document Type and Number:
WIPO Patent Application WO/2023/028996
Kind Code:
A1
Abstract:
Methods and devices are described for validating reproducibility of a machine learning model. Upon the initiation of a first machine learning model training instance, a training profile containing random values used in the first machine learning training instance is obtained. Upon initiation of a second machine learning model training instance, an interception module is used to automatically intercept non-determinism introducing functions called during the training process, retrieve corresponding random values from the training profile and use the retrieved random values in the second machine learning model training instance. In response to validating that the results returned from the first and second machine learning model training instances are identical, the machine learning model is validated as a reproducible machine learning model.
Inventors:
CHEN BOYUAN (CA)
WEN MINGZHI (CN)
SHI YONG (CN)
JIANG ZHENMING (CA)
WEN MINGZHI (CN)
SHI YONG (CN)
JIANG ZHENMING (CA)
Application Number:
PCT/CN2021/116475
Publication Date:
March 09, 2023
Filing Date:
September 03, 2021
Export Citation:
Assignee:
HUAWEI TECH CO LTD (CN)
International Classes:
G06N3/08; G06N20/00
Foreign References:
| CN111427541A | 2020-07-17 | |||
| CN111612168A | 2020-09-01 | |||
| CN112560939A | 2021-03-26 | |||
| US20190156216A1 | 2019-05-23 | |||
| US20200401644A1 | 2020-12-24 | |||
| US20210150330A1 | 2021-05-20 |
Download PDF: