US 11,854,672 B2
Method and apparatus for generating chemical structure
Youngchun Kwon, Yongin-si (KR); Seokho Kang, Suwon-si (KR); Won-joon Son, Yongin-si (KR); Dongseon Lee, Yongin-si (KR); and Younsuk Choi, Suwon-si (KR)
Assigned to SAMSUNG ELECTRONICS CO., LTD., Suwon-si (KR); and RESEARCH & BUSINESS FOUNDATION SUNGKYUNKWAN UNIVERSITY, Suwon-si (KR)
Filed by SAMSUNG ELECTRONICS CO., LTD., Suwon-si (KR); and RESEARCH & BUSINESS FOUNDATION SUNGKYUNKWAN UNIVERSITY, Suwon-si (KR)
Filed on Jun. 10, 2020, as Appl. No. 16/897,456.
Claims priority of provisional application 62/913,297, filed on Oct. 10, 2019.
Claims priority of application No. 10-2019-0149114 (KR), filed on Nov. 19, 2019.
Prior Publication US 2021/0110892 A1, Apr. 15, 2021
Int. Cl. G05B 19/4155 (2006.01); G16C 60/00 (2019.01); G16C 20/70 (2019.01); G16C 20/10 (2019.01); G16C 20/20 (2019.01)
CPC G16C 60/00 (2019.02) [G16C 20/10 (2019.02); G16C 20/20 (2019.02); G16C 20/70 (2019.02)] 14 Claims
OG exemplary drawing
 
1. A method of generating a second chemical structure using a chemical structure generation model, the method comprising:
receiving a first descriptor of a first chemical structure;
encoding the first descriptor into a latent variable using an encoder of the chemical structure generation model;
generating a second descriptor by decoding the latent variable using a decoder of the chemical structure generation model;
generating the second chemical structure corresponding to the second descriptor;
determining whether the second chemical structure corresponds to a valid chemical structure; and
based on determining that the second chemical structure corresponds to a valid chemical structure, updating the chemical structure generation model based on the second descriptor,
wherein updating the chemical structure generation model based on the second descriptor comprises:
inputting the second descriptor into a validity evaluation model, wherein the validity evaluation model is configured to:
output a first value as a reward based on the second descriptor representing a valid chemical structure, and
output a second value as the reward based on the second descriptor representing an invalid chemical structure; and
updating a weight of the chemical structure generation model based on the reward for the second descriptor having the first value.