Title:
PROPERTY PREDICTION SYSTEM, PROPERTY PREDICTION METHOD, AND PROPERTY PREDICTION PROGRAM
Document Type and Number:
WIPO Patent Application WO/2022/225009
Kind Code:
A1
Abstract:
In the present invention, an input data generation system 10 comprises at least one processor and is for generating input data for machine learning to predict properties of materials based on raw materials of a known structure. The at least one processor: acquires partial structure data, representative of a partial structure, from a database; receives at least input of raw material structure data specifying the structures of raw materials and blending ratio data representative of the ratios of blending of the raw materials; generates partial structure input data D representative of a partial structure present in the structure of the raw materials, on the basis of the partial structure data and the raw material structure data; and causes the blending ratio data to be reflected in the partial structure input data D of the raw materials, thereby generating input data, and causes the input data to be inputted to a machine learning model.
More Like This:
Inventors:
HANAOKA KYOHEI (JP)
Application Number:
PCT/JP2022/018416
Publication Date:
October 27, 2022
Filing Date:
April 21, 2022
Export Citation:
Assignee:
SHOWA DENKO MATERIALS CO LTD (JP)
International Classes:
G16C20/30; G06N20/00
Foreign References:
| JP2019028879A | 2019-02-21 | |||
| JP2019028879A | 2019-02-21 |
Other References:
OKUNO YOSHISHIGE: "Reducing the number of attempts to synthesize thermosetting polymers by AI analysis", THE 2019 PROJECT RESULT REPORT MATERIAL OF ULTRA-ADVANCED MATERIALS ULTRA-HIGH-SPEED DEVELOPMENT BASIC TECHNOLOGY, 13 April 2020 (2020-04-13), XP055978466, Retrieved from the Internet [retrieved on 20221107]
MINAMI TAKUYA: "Prediction of thermoset resin properties by using classification of raw materials and machine learning", THE SYMPOSIUM ON CHEMOINFORMATICS, 27 October 2018 (2018-10-27), XP055978469, Retrieved from the Internet [retrieved on 20221107]
IKEBATA HISAKI: "Polymer informatics utilizing PoLyInfo", CICSJ BULLETIN, 30 December 2019 (2019-12-30), pages 94 - 98, XP055978471, Retrieved from the Internet [retrieved on 20221107], DOI: 10.11546/cicsj.37.94
MINAMI TAKUYA: "Prediction of thermoset resin properties by using classification of raw materials and machine learning", THE SYMPOSIUM ON CHEMOINFORMATICS, 27 October 2018 (2018-10-27), XP055978469, Retrieved from the Internet
IKEBATA HISAKI: "Polymer informatics utilizing PoLyInfo", CICSJ BULLETIN, 30 December 2019 (2019-12-30), pages 94 - 98, XP055978471, Retrieved from the Internet
Attorney, Agent or Firm:
HASEGAWA Yoshiki et al. (JP)
Download PDF: