公开(公告)号：US20200183985A1

公开(公告)日：2020-06-11

申请号：US16791209

申请日：2020-02-14

Applicant: Alibaba Group Holding Limited

Inventor： Wangshu Zhang ,  Zhiwei Shi ,  Junhong Liu ,  Jie Cai ,  Xiang Hu ,  Defeng Mao

IPC: G06F16/9032 ,  G06Q30/00 ,  G06F16/9035 ,  G06F16/903

Abstract: A query is received from a user at a data service engine. The query includes a string of characters. A number of candidate topics are identified by the data service engine based on the query. A similarity score is determined between the query and each of the plurality of candidate topics based on a Vector Propagation On a Click Graph (VPCG) model trained based on user click data. A number of candidate topics are ranked based on the similarity scores. One or more topics are selected from the ranked candidate topics. The selected topics are outputted via a user interface (UI).

公开(公告)号：US10824679B2

公开(公告)日：2020-11-03

申请号：US16791209

申请日：2020-02-14

Applicant: Alibaba Group Holding Limited

Inventor： Wangshu Zhang ,  Zhiwei Shi ,  Junhong Liu ,  Jie Cai ,  Xiang Hu ,  Defeng Mao

IPC: G06F7/00 ,  G06F16/00 ,  G06F16/9032 ,  G06F16/9035 ,  G06F16/903 ,  G06Q30/00

Abstract: A query is received from a user at a data service engine. The query includes a string of characters. A number of candidate topics are identified by the data service engine based on the query. A similarity score is determined between the query and each of the plurality of candidate topics based on a Vector Propagation On a Click Graph (VPCG) model trained based on user click data. A number of candidate topics are ranked based on the similarity scores. One or more topics are selected from the ranked candidate topics. The selected topics are outputted via a user interface (UI).

公开(公告)号：US10592555B1

公开(公告)日：2020-03-17

申请号：US16390173

申请日：2019-04-22

Applicant: Alibaba Group Holding Limited

Inventor： Wangshu Zhang ,  Zhiwei Shi ,  Junhong Liu ,  Jie Cai ,  Xiang Hu ,  Defeng Mao

IPC: G06F7/00 ,  G06F17/30 ,  G06F16/9032 ,  G06F16/9035 ,  G06F16/903 ,  G06Q30/00

Abstract: A query is received from a user at a data service engine. The query includes a string of characters. A number of candidate topics are identified by the data service engine based on the query. A similarity score is determined between the query and each of the plurality of candidate topics based on a Vector Propagation On a Click Graph (VPCG) model trained based on user click data. A number of candidate topics are ranked based on the similarity scores. One or more topics are selected from the ranked candidate topics. The selected topics are outputted via a user interface (UI).

公开(公告)号：US11816440B2

公开(公告)日：2023-11-14

申请号：US17971047

申请日：2022-10-21

Applicant: Alibaba Group Holding Limited

Inventor： Ling Li ,  Zhiwei Shi ,  Yanjie Liang

IPC: G06F40/30 ,  G06F40/216 ,  G06V30/262 ,  G06F18/24 ,  G06V30/40 ,  G06V30/148 ,  G06V10/40

CPC classification number: G06F40/30 ,  G06F18/24 ,  G06F40/216 ,  G06V10/40 ,  G06V30/153 ,  G06V30/274 ,  G06V30/40

Abstract: The disclosed embodiments describe methods, systems, and apparatuses for determining user intent. A method is disclosed comprising obtaining a session text of a user; calculating, by the processor, a feature vector based on the session text; determining probabilities that the session text belongs to a plurality of intent labels, the probabilities calculated using a multi-level hierarchal intent classification model, the intent labels assigned to levels in the multi-level hierarchal intent classification model; and assigning a user intent to the session text based on the probabilities.

公开(公告)号：US11514245B2

公开(公告)日：2022-11-29

申请号：US16352159

申请日：2019-03-13

Applicant: Alibaba Group Holding Limited

Inventor： Ling Li ,  Zhiwei Shi ,  Yanjie Liang

IPC: G06F40/253 ,  G06F40/30 ,  G06K9/62 ,  G06F40/216 ,  G06V30/262

Abstract: The disclosed embodiments describe methods, systems, and apparatuses for determining user intent. A method is disclosed comprising obtaining a session text of a user; calculating, by the processor, a feature vector based on the session text; determining probabilities that the session text belongs to a plurality of intent labels, the probabilities calculated using a multi-level hierarchal intent classification model, the intent labels assigned to levels in the multi-level hierarchal intent classification model; and assigning a user intent to the session text based on the probabilities.