公开(公告)号：US20220406034A1

公开(公告)日：2022-12-22

申请号：US17822898

申请日：2022-08-29

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

Inventor： Jingru GAN ,  Haiwei WANG ,  Jinchang LUO ,  Kunbin CHEN ,  Wei HE ,  Shuhui WANG

IPC: G06V10/74 ,  G06F40/295 ,  G06V10/80

Abstract: A method for extracting information, includes: obtaining an information stream comprising text and an image; generating, according to the text, embedded representations of textual entity mentions and a textual similarity matrix of the textual entity mentions and candidate textual entities; generating, according to the image, embedded representations of image entity mentions and an image similarity matrix of the image entity mentions and candidate image entities; and determining, based on an optimal transport, target textual entities of the textual entity mentions and target image entities of the image entity mentions according to the embedded representations of the textual entity mentions, the embedded representations of the image entity mentions, the textual similarity matrix and the image similarity matrix.

公开(公告)号：US20250013876A1

公开(公告)日：2025-01-09

申请号：US18889928

申请日：2024-09-19

Applicant: Beijing Baidu Netcom Science Technology Co., Ltd.

Inventor： Xianwei XUE ,  Qiutong PAN ,  Jinchang LUO ,  Bolei HE ,  Wei HE

IPC: G06N3/0985 ,  G06F40/30 ,  G06F40/40 ,  G06N3/0475

Abstract: An apparatus for training a large language model includes: at least one sample text instruction is input into a target large language model to obtain at least one standard response text, and the at least one sample text instruction is input into a large language model to be trained to obtain at least one predicted response text. A first sample response text is determined from the at least one standard response text according to the score difference between a first quality score of a standard response text and a second quality score of a predicted response text. A first target training sample is generated according to the first sample response text and a sample text instruction corresponding to the first sample response text, and a training dataset is constructed according to the first target training sample.

公开(公告)号：US20250013676A1

公开(公告)日：2025-01-09

申请号：US18889497

申请日：2024-09-19

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

Inventor： Jinchang LUO ,  Bolei HE ,  Kunbin CHEN ,  Wei HE

IPC: G06F16/332 ,  G06F16/33

Abstract: A computer-implemented method for information processing based on a large language model is provided. The method includes obtaining query information provided by a user. The method further includes determining memory information related to the query information. The method further includes determining, based on the query information and the memory information, a tool for processing the query information. The method further includes invoking the tool to obtain auxiliary information. The method further includes generating, based on the query information and the auxiliary information, a result of processing the query information.

公开(公告)号：US20230085599A1

公开(公告)日：2023-03-16

申请号：US18057560

申请日：2022-11-21

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

Inventor： Jinchang LUO ,  Haiwei WANG ,  Junzhao BU ,  Kunbin CHEN ,  Wei HE

IPC: G06N3/04

Abstract: The disclosure provides a method for training a tag recommendation model. The method includes: collecting training materials that comprise interest tags in response to receiving an instruction for collecting training materials; obtaining training semantic vectors that comprise the interest tags by representing features of the training materials using a semantic enhanced representation frame; obtaining training encoding vectors by aggregating social networks into the training semantic vectors; and obtaining a tag recommendation model by training a double-layer neural network structure using the training encoding vectors as inputs and the interest tags as outputs. Therefore, the interest tags obtained in the disclosure are more accurate.