公开(公告)号：US12147431B2

公开(公告)日：2024-11-19

申请号：US18060736

申请日：2022-12-01

Applicant: BEIJING BAIDU NETCOM SCIENCE TECHNOLOGY CO., LTD.

Inventor： Haocheng Liu ,  Jingyu Xu

IPC: G06F16/2457 ,  G06F16/28

Abstract: A method for pushing a resource includes obtaining a collection of users to which a resource is to be pushed, in which the collection includes a plurality of users; obtaining respective predicted conversion rates of the resource for the plurality of users; obtaining user groups by dividing the plurality of users into groups based on respective predicted conversion rates for the plurality of users, in which the user groups correspond to different conversion rate ranges; and for each user group, selecting a target user from the user group based on a real conversion rate related to pushed users corresponding to the conversion rate range of the user group, and pushing the resource to the target user.

公开(公告)号：US20250005018A1

公开(公告)日：2025-01-02

申请号：US18745835

申请日：2024-06-17

Applicant: BEIJING BAIDU NETCOM SCIENCE TECHNOLOGY CO., LTD.

Inventor： Haocheng Liu ,  Haiyang Xu ,  Liangjie Zhang ,  Qian Sun

IPC: G06F16/242 ,  G06F16/21 ,  G06F16/248

Abstract: A computer-implemented for information processing based on a large language model (LLM) includes: obtaining a query request; determining, by the LLM and based on the query request, a query condition, a source of a query result, and a select field for generating a structured query language (SQL) statement; generating by the LLM, the SQL statement by combining the query condition, the source of the query result, and the select field; and obtaining the query result corresponding to the query request, by inputting the SQL statement into the LLM for execution

公开(公告)号：US20230095947A1

公开(公告)日：2023-03-30

申请号：US18060736

申请日：2022-12-01

Applicant: BEIJING BAIDU NETCOM SCIENCE TECHNOLOGY CO., LTD.

Inventor： Haocheng Liu ,  Jingyu Xu

IPC: G06F16/2457 ,  G06F16/28

Abstract: A method for pushing a resource includes obtaining a collection of users to which a resource is to be pushed, in which the collection includes a plurality of users; obtaining respective predicted conversion rates of the resource for the plurality of users; obtaining user groups by dividing the plurality of users into groups based on respective predicted conversion rates for the plurality of users, in which the user groups correspond to different conversion rate ranges; and for each user group, selecting a target user from the user group based on a real conversion rate related to pushed users corresponding to the conversion rate range of the user group, and pushing the resource to the target user.