公开(公告)号：US20230196117A1

公开(公告)日：2023-06-22

申请号：US18173310

申请日：2023-02-23

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Zewei DU ,  Hengtong HU ,  Lingxi XIE ,  Qi TIAN

IPC: G06N3/0895 ,  G06V10/82 ,  G06V10/774 ,  G06V10/771

CPC classification number: G06N3/0895 ,  G06V10/82 ,  G06V10/7753 ,  G06V10/771

Abstract: Embodiments of this application disclose a training method for a semi-supervised learning model which can be applied to computer vision in the field of artificial intelligence. The method includes: first predicting classification categories of some unlabeled samples by using a trained first semi-supervised learning model, to obtain a prediction label; and determining whether each prediction label is correct in a one-bit labeling manner, and if prediction is correct, obtaining a correct label (a positive label) of the sample, or if prediction is incorrect, excluding an incorrect label (a negative label) of the sample. Then, in a next training phase, a training set (a first training set) is reconstructed based on the information, and an initial semi-supervised learning model is retrained based on the first training set, to improve prediction accuracy of the model. In one-bit labeling, an annotator only needs to answer “yes” or “no” for the prediction label.

公开(公告)号：US20230075836A1

公开(公告)日：2023-03-09

申请号：US17986081

申请日：2022-11-14

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Fuhui TANG ,  Xiaopeng ZHANG ,  Minzhe NIU ,  Zichen WANG ,  Jianhua HAN ,  Qi TIAN

IPC: G06V10/774 ,  G06V10/40 ,  G06V10/82 ,  G06V10/764

Abstract: A model training method and a related apparatus are provided, which may be used in computer vision to perform image detection. The method includes: extracting feature information in a target image; further separately extracting features of a target object from the feature information by using a Gaussian mask to obtain a first local feature and a second local feature; determining a feature loss by using the first local feature and the second local feature; performing prediction by using the first network and the second network based on a same region proposal set to obtain a first classification predicted value and a second classification predicted value, and obtaining a classification loss based on the first classification predicted value and the second classification predicted value; and training the second network based on the classification loss and the feature loss to obtain a target network.

公开(公告)号：US20220327835A1

公开(公告)日：2022-10-13

申请号：US17852684

申请日：2022-06-29

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Ke NING ,  Longhui WEI ,  Lingxi XIE ,  Jianzhuang LIU ,  Qi TIAN

IPC: G06V20/40 ,  G06V10/80

Abstract: A video clip location technology in the field of computer vision pertaining to artificial intelligence that provides a video processing method and apparatus. The method includes: obtaining a semantic feature of an input sentence; performing semantic enhancement on a video frame based on the semantic feature to obtain a video feature of the video frame, where the video feature includes the semantic feature; and determining, based on the semantic feature and the video feature, whether a video clip to which the video frame belongs is a target video clip corresponding to the input sentence. The method helps improve accuracy of recognizing a target video clip corresponding to an input sentence.

公开(公告)号：US20210312261A1

公开(公告)日：2021-10-07

申请号：US17220158

申请日：2021-04-01

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Yixing XU ,  Kai HAN ,  Yunhe WANG ,  Chunjing XU ,  Qi TIAN

IPC: G06N3/04 ,  G06N3/08

Abstract: The present application discloses a neural network search method in the field of artificial intelligence, and the neural network search method includes: obtaining a feature tensor of each of a plurality of neural networks, where the feature tensor of each neural network is used to represent a computing capability of the neural network; inputting the feature tensor of each of the plurality of neural networks into an accuracy prediction model for calculation, to obtain accuracy of each neural network, where the accuracy prediction model is obtained through training based on a ranking-based loss function; and determining a neural network corresponding to the maximum accuracy as a target neural network. Embodiments of the present invention help improve accuracy of a network structure found through search.