Title:
MACHINE LEARNING SYSTEM AND MODEL TRAINING METHOD
Document Type and Number:
WIPO Patent Application WO/2023/011470
Kind Code:
A1
Abstract:
A machine learning system and a model training method. The system comprises a master control module (100), an original data acquisition module (110), a data screening module (120), a sample labeling module (130) and a learning module (140), wherein the master control module (100) calls, according to a target calling sequence, target modules that are selected by a user, and trains a model; the data screening module (120) screens out valuable target original data from the current original data that is obtained by the original data acquisition module (110); the sample labeling module (130) outputs data to be labeled to the user for labeling; and when called, the learning module (140) trains, according to a corresponding learning mode thereof, a machine learning model to be trained, so as to obtain a corresponding model. The machine learning system can acquire data of the current scenario, and train a model according to a learning mode that is selected by a user; and there is no need for a developer to re-design the model when the model is applied to a different application scenario, thereby realizing adaptive learning for an actual application scenario.
More Like This:
Inventors:
CHENG ZHANZHAN (CN)
Application Number:
PCT/CN2022/109701
Publication Date:
February 09, 2023
Filing Date:
August 02, 2022
Export Citation:
Assignee:
SHANGHAI GOLDWAY INTELLIGENT TRANSP SYSTEM CO LTD (CN)
International Classes:
G06N20/00; G06N3/04; G06N3/08
Foreign References:
| CN113516251A | 2021-10-19 | |||
| CN112559784A | 2021-03-26 | |||
| CN110610197A | 2019-12-24 | |||
| CN112347261A | 2021-02-09 | |||
| CN112784095A | 2021-05-11 | |||
| CN104318242A | 2015-01-28 | |||
| CN108833409A | 2018-11-16 | |||
| US11043208B1 | 2021-06-22 | |||
| US20200125897A1 | 2020-04-23 |
Attorney, Agent or Firm:
PATENTSINO IP FIRM (CN)
Download PDF: