公开(公告)号：US20240055071A1

公开(公告)日：2024-02-15

申请号：US18494372

申请日：2023-10-25

申请人： TENCENT TECHNOLOGY (SHENZHEN) COMPANY LIMITED ,  ZHEJIANG UNIVERSITY

发明人： Xujun ZHANG ,  Benben LIAO ,  Shengyu ZHANG ,  Tingjun HOU

IPC分类号： G16B15/30 ,  G16B40/20 ,  G16C20/30 ,  G16C20/50 ,  G16C20/60 ,  G06N3/0455 ,  G06N3/063

CPC分类号： G16B15/30 ,  G16B40/20 ,  G16C20/30 ,  G16C20/50 ,  G16C20/60 ,  G06N3/0455 ,  G06N3/063

摘要： An artificial intelligence-based compound processing method and apparatus, an electronic device, a computer-readable storage medium, and a computer program product relates to an artificial intelligence technology. The method includes obtaining an active compound for a target protein; performing compound generation processing on an attribute property of the active compound to obtain a first candidate compound; performing molecular docking processing on the active compound and the target protein to obtain molecular docking information respectively corresponding to a plurality of molecular conformations of the active compound; screening the plurality of molecular conformations based on the molecular docking information respectively to identify a second candidate compound corresponding to the active compound; and constructing a compound library for the target protein based on the first candidate compound and the second candidate compound.

公开(公告)号：US20240346315A1

公开(公告)日：2024-10-17

申请号：US18755350

申请日：2024-06-26

申请人： TENCENT TECHNOLOGY (SHENZHEN) COMPANY LIMITED

发明人： Zhenxing WU ,  Chang-Yu HSIEH ,  Shengyu ZHANG ,  Tingjun HOU

IPC分类号： G06N3/08

CPC分类号： G06N3/08

摘要： A method includes determining a plurality of neural network models each corresponding to one of a plurality of molecular representations, and, for each molecular representation in the plurality of molecular representations, determining, using the neural network model corresponding to the molecular representation, a molecular property prediction result and prediction confidence corresponding to unlabeled data in an unlabeled data set, obtaining at least a portion of the unlabeled data as reference unlabeled data, the reference unlabeled data having corresponding prediction confidence higher than a preset threshold, and determining, based on the reference unlabeled data and a molecular property prediction result corresponding to the reference unlabeled data, pseudo-labeled data of a neural network model corresponding to another molecular representation in the plurality of molecular representations. The method further includes performing training on the plurality of neural network models respectively based on corresponding pseudo-labeled data of the plurality of neural network models.