US 12,204,859 B2
Text processing method, model training method, and apparatus
Yasheng Wang, Shenzhen (CN); Xin Jiang, Hong Kong (CN); Xiao Chen, Hong Kong (CN); Qun Liu, Hong Kong (CN); Zhengyan Zhang, Beijing (CN); Fanchao Qi, Beijing (CN); and Zhiyuan Liu, Beijing (CN)
Assigned to Huawei Technologies Co., Ltd., Shenzhen (CN); and TSINGHUA UNIVERSITY, Beijing (CN)
Filed by Huawei Technologies Co., Ltd., Shenzhen (CN); and TSINGHUA UNIVERSITY, Beijing (CN)
Filed on Nov. 15, 2021, as Appl. No. 17/526,832.
Application 17/526,832 is a continuation of application No. PCT/CN2020/072588, filed on Jan. 17, 2020.
Claims priority of application No. 201910410679.4 (CN), filed on May 16, 2019.
Prior Publication US 2022/0147715 A1, May 12, 2022
Int. Cl. G06F 40/295 (2020.01); G06F 40/211 (2020.01); G06F 40/30 (2020.01); G06F 40/237 (2020.01); G06F 40/279 (2020.01); G06F 40/284 (2020.01)
CPC G06F 40/295 (2020.01) [G06F 40/211 (2020.01); G06F 40/30 (2020.01); G06F 40/237 (2020.01); G06F 40/279 (2020.01); G06F 40/284 (2020.01)] 13 Claims
OG exemplary drawing
 
1. A text processing method, comprising:
obtaining target knowledge data, wherein the target knowledge data comprises a first named entity, a second named entity, and an association between the first named entity and the second named entity;
processing, through vectorization, the target knowledge data to obtain a target knowledge vector, wherein the target knowledge vector comprises a vector corresponding to the first named entity, a vector corresponding to the second named entity, and a vector corresponding to the association between the first named entity and the second named entity;
processing, through vectorization, to-be-processed text to obtain a target text vector, wherein the to-be-processed text comprises one or more named entities, and the one or more named entities comprise the first named entity;
fusing the target text vector and the target knowledge vector based on a target fusion model to obtain a fused target text vector and a fused target knowledge vector, wherein the fused target text vector comprises information about the to-be-processed text and at least a part of information in the target knowledge data, and wherein the fused target knowledge vector comprises the target knowledge data and semantic background information of the to-be-processed text; and
processing the fused target text vector and/or the fused target knowledge vector based on a target processing model to obtain a processing result corresponding to a target task.