Patents by Inventor Zhao ZHONG

Zhao ZHONG has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).

  • Publication number: 20250013877
    Abstract: This application discloses a data processing method and apparatus in the artificial intelligence field, to improve prediction accuracy of a neural predictor. The neural predictor uses a small quantity of training samples. In the data processing method, a hyperparameter combination sampled from a hyperparameter search space corresponding to a user task, a plurality of samples included in a training set, and evaluation metrics of the plurality of samples are used as inputs to the neural predictor, and a prediction metric corresponding to the hyperparameter combination is determined by using the neural predictor. A hyperparameter sample and an evaluation metric of the hyperparameter sample are used to assist in predicting the hyperparameter combination sampled from the hyperparameter search space. Because the hyperparameter combination is predicted based on the evaluation metric and the hyperparameter sample that already has the evaluation metric, accuracy can be improved.
    Type: Application
    Filed: September 24, 2024
    Publication date: January 9, 2025
    Inventors: Yucong Zhou, Zhao Zhong
  • Publication number: 20240249115
    Abstract: An input of an optimized query Query feature transformation module is obtained based on an output feature of at least one previous network layer of the optimized attention layer. An input of an optimized key Key feature transformation module is obtained based on an output feature of at least one previous network layer of the optimized attention layer. An input of an optimized value Value feature transformation module is obtained based on an output feature of at least one previous network layer of the optimized attention layer. An input of at least one feature transformation module in the optimized query Query feature transformation module, the optimized key Key feature transformation module, and the optimized value Value feature transformation module is obtained based on an output feature of at least one non-adjacent previous network layer of the optimized attention layer.
    Type: Application
    Filed: March 15, 2024
    Publication date: July 25, 2024
    Inventors: Yunxiao SUN, Yucong ZHOU, Zhao ZHONG
  • Publication number: 20240233358
    Abstract: Embodiments of this application disclose an image classification method, a device and a storage medium, and belong to the field of image processing. In this method, a target image is processed by using a current neural network model, to obtain a current classification result. The current neural network model is a neural network model i corresponding to a largest probability in a selection result output by a neural network model a, and the selection result includes probabilities corresponding to p neural network models in m neural network models. A current integration result is determined based on the current classification result, and a category of the target image is determined based on the current integration result.
    Type: Application
    Filed: December 29, 2023
    Publication date: July 11, 2024
    Inventors: Yikang Zhang, Zhao Zhong
  • Publication number: 20240135698
    Abstract: Embodiments of this application disclose an image classification method, a device and a storage medium, and belong to the field of image processing. In this method, a target image is processed by using a current neural network model, to obtain a current classification result. The current neural network model is a neural network model i corresponding to a largest probability in a selection result output by a neural network model a, and the selection result includes probabilities corresponding to p neural network models in m neural network models. A current integration result is determined based on the current classification result, and a category of the target image is determined based on the current integration result.
    Type: Application
    Filed: December 29, 2023
    Publication date: April 25, 2024
    Inventors: Yikang Zhang, Zhao Zhong
  • Publication number: 20240135174
    Abstract: This application discloses a data processing method, and a neural network model training method and apparatus in the field of artificial intelligence. The data processing method includes: processing to-be-processed data by using a target neural network quantization model, where the target neural network quantization model includes a plurality of groups of fusion parameters, the target neural network quantization model is obtained by quantizing a target neural network model, an activation function of the target neural network model includes a piecewise linear function (PWL), the PWL includes a plurality of intervals, and there is a correspondence between the plurality of groups of fusion parameters and the plurality of intervals. According to the method in this application, a model that uses the PWL as an activation function can be quantized, thereby improving an inference speed of the model.
    Type: Application
    Filed: December 29, 2023
    Publication date: April 25, 2024
    Applicant: HUAWEI TECHNOLOGIES CO., LTD.
    Inventors: Yucong Zhou, Zhao Zhong, Yannan Xiao, Genshu Liu
  • Publication number: 20240078428
    Abstract: A neural network model training method, a data processing method, and an apparatus are disclosed. The neural network model training method includes: training a neural network model based on training data, where an activation function of the neural network model includes at least one piecewise function, and the piecewise function includes a plurality of trainable parameters; and updating the plurality of trainable parameters of the at least one piecewise function in a training process. According to the method, the activation function suitable for the neural network model can be obtained. This can improve performance of the neural network model.
    Type: Application
    Filed: July 19, 2023
    Publication date: March 7, 2024
    Inventors: Yucong ZHOU, Zezhou ZHU, Zhao ZHONG
  • Publication number: 20230385642
    Abstract: This application discloses a model training method, which may be applied to the field of artificial intelligence. The method includes: obtaining a first neural network model; replacing a first convolutional layer in the first neural network model with a linear operation to obtain a plurality of second neural network models; and performing model training on a plurality of second neural network models, to obtain a neural network model with a highest model precision in a plurality of trained second neural network models. In this application, a convolutional layer in a to-be-trained neural network is replaced with a linear operation equivalent to a convolutional layer. A manner with highest precision is selected from a plurality of replacement manners, to improve precision of a trained model.
    Type: Application
    Filed: August 8, 2023
    Publication date: November 30, 2023
    Inventors: Yucong ZHOU, Zhao ZHONG
  • Publication number: 20230289572
    Abstract: A neural network structure determining method is disclosed. The method includes: obtaining a to-be-trained initial neural network, where the initial neural network includes M first blocks block and a second block, the second block is connected to each first block, and each first block corresponds to one trainable target weight; performing model training on the initial neural network, to obtain M updated target weights; and updating a connection relationship between the second block and the M first blocks in the initial neural network based on the M updated target weights, to obtain a first neural network.
    Type: Application
    Filed: May 12, 2023
    Publication date: September 14, 2023
    Inventors: Yifan XIAO, Jian ZHANG, Zhao ZHONG
  • Publication number: 20230222639
    Abstract: This application provides a data processing method, system, and apparatus, and relates to the field of artificial intelligence (AI). The data processing method may be performed by a server, or may be performed by a device having a data processing function. During execution, reference data is first obtained. The reference data includes RGB image data and a device parameter of an image device. Then, a plurality of conversion parameters required for converting the RGB image data into RAW data are determined. Finally, the RGB image data is processed into the RAW data based on the plurality of conversion parameters. The RAW data matches the device parameter of the image device. In this application, the RGB image data is converted into the RAW data based on the plurality of conversion parameters rather than manual experience. Therefore, the described data processing method, system, and apparatus improve data processing efficiency.
    Type: Application
    Filed: March 13, 2023
    Publication date: July 13, 2023
    Inventors: Xinyu ZHANG, Bing YU, Zhao ZHONG
  • Publication number: 20230186103
    Abstract: This application relates to the field of artificial intelligence technologies, and describes a classification model training method, a hyperparameter search method, and an apparatus. The training method includes obtaining a target hyperparameter of a to-be-trained classification model. The target hyperparameter is used to control a gradient update operation of the to-be-trained classification model. The to-be-trained classification model includes a scaling invariance linear layer. The scaling invariance linear layer enables a predicted classification result output when a weight parameter of the to-be-trained classification model is multiplied by any scaling coefficient to remain unchanged. The method further includes updating the weight parameter of the to-be-trained classification model based on the target hyperparameter and a target training manner, to obtain a trained classification model.
    Type: Application
    Filed: February 6, 2023
    Publication date: June 15, 2023
    Inventors: Yucong ZHOU, Zhao ZHONG
  • Publication number: 20220319154
    Abstract: This application discloses a neural network model update method, an image processing method, and an apparatus in the field of artificial intelligence. The neural network model update method includes: obtaining a structure of a neural network model and a related parameter of the neural network model; training the neural network model based on the related parameter of the neural network model to obtain a trained neural network model; and if an evaluation result of the trained neural network model does not meet a preset condition, updating at least two items of the related parameter of the neural network model and the structure of the neural network model until an evaluation result of an updated neural network model meets a preset condition and/or a quantity of updates reaches a preset quantity of times. According to the method in this application, efficiency of updating a neural network model can be improved.
    Type: Application
    Filed: June 17, 2022
    Publication date: October 6, 2022
    Applicant: HUAWEI TECHNOLOGIES CO., LTD.
    Inventors: Xinyu ZHANG, Peng YUAN, Muyuan FANG, Zhao ZHONG
  • Publication number: 20220222934
    Abstract: This application discloses a neural network construction method and apparatus, and an image processing method and apparatus in the field of artificial intelligence. The neural network construction method includes: constructing search space based on an application requirement of a target neural network, where the search space includes M elements, the M elements are used to indicate M network structures, each of the M elements includes a quantity of blocks in a stage in a corresponding network structure and a channel quantity of each block, and M is a positive integer (S710); and selecting a target network structure from the M network structures based on a distribution relationship among unevaluated elements in the search space (S720). According to the method, a neural network satisfying a performance requirement can be efficiently constructed.
    Type: Application
    Filed: March 21, 2022
    Publication date: July 14, 2022
    Applicant: HUAWEI TECHNOLOGIES CO., LTD.
    Inventors: Yikang ZHANG, Zhao ZHONG
  • Publication number: 20220215259
    Abstract: Technical solutions in this application are applied to the field of artificial intelligence. This application provides a neural network training method, a method for performing data processing by using a neural network trained by using the method, and a related apparatus. According to the training method in this application, a target neural network is trained in an adversarial manner, so that a policy search module can continuously discover a weakness of the target neural network, generate a policy of higher quality according to the weakness, and perform data augmentation according to the policy to obtain data of higher quality. A target neural network of higher quality can be trained according to the data. In the data processing method in this application, data processing is performed by using the foregoing target neural network, so that a more accurate processing result can be obtained.
    Type: Application
    Filed: March 22, 2022
    Publication date: July 7, 2022
    Inventors: Xinyu ZHANG, Peng YUAN, Zhao ZHONG
  • Patent number: 11270190
    Abstract: Embodiments of the present application disclose a method and apparatus for generating a neural network structure, an electronic device, and a storage medium. The method comprises: sampling a neural network structure to generate a network block, the network block comprising at least one network layer; constructing a sampling neural network based on the network block; training the sampling neural network based on sample data, and obtaining an accuracy corresponding to the sampling neural network; and in response to that the accuracy does not meet a preset condition, regenerating a new network block according to the accuracy until a sampling neural network constructed by the new network block meets the preset condition, and using the sampling neural network meeting the preset condition as a target neural network.
    Type: Grant
    Filed: November 26, 2018
    Date of Patent: March 8, 2022
    Assignee: Beijing SenseTime Technology Development Co., Ltd.
    Inventors: Zhao Zhong, Junjie Yan, Chenglin Liu
  • Publication number: 20190095780
    Abstract: Embodiments of the present application disclose a method and apparatus for generating a neural network structure, an electronic device, and a storage medium. The method comprises: sampling a neural network structure to generate a network block, the network block comprising at least one network layer; constructing a sampling neural network based on the network block; training the sampling neural network based on sample data, and obtaining an accuracy corresponding to the sampling neural network; and in response to that the accuracy does not meet a preset condition, regenerating a new network block according to the accuracy until a sampling neural network constructed by the new network block meets the preset condition, and using the sampling neural network meeting the preset condition as a target neural network.
    Type: Application
    Filed: November 26, 2018
    Publication date: March 28, 2019
    Applicant: Beijing SenseTime Technology Development Co., Ltd
    Inventors: Zhao ZHONG, Junjie YAN, Chenglin LIU
  • Publication number: 20170136375
    Abstract: A reversible toy comprising a first figure and a second figure, each figure comprising a hollow body portion with a upper portion and a base portion and each figure having a conically shaped hollow and a shared base portion. The toy is reversible between a first position and a second position by turning it inside out via the shared base portion such that one of the figures is substantially concealed when the other is visible.
    Type: Application
    Filed: June 12, 2015
    Publication date: May 18, 2017
    Inventors: Sini Aikkattukuzhi Narayanan, Patrick S. Baran, Joseph Michael Onderko, Zhao Zhong Sun
  • Publication number: 20090203885
    Abstract: The present invention relates to Astrocyte Modulated Genes (AMGs). AMGs are genes whose expression are modulated in human astrocytes grown in primary cell culture following the exposure of these cells to either the human immunodeficiency virus HIV-1 or to the HIV-1 protein gp120. AMGs comprise both Astrocyte Enhanced Genes (AEGs) and Astrocyte Suppressed Genes (ASGs). Thus, the present invention further relates to Astrocyte Enhanced Genes (AEGs), the expression of which are up-regulated in human astrocytes grown in primary cell culture that are exposed to either the human immunodeficiency virus HIV-1 or to the HIV-1 protein gp120, and to Astrocyte Suppressed Genes (ASGs), the expression of which are downregulated in human astrocytes grown in primary cell culture that are exposed to either the human immunodeficiency virus HIV-1 or to the HIV-1 protein gp120. Because they may play a role in HIV-associated dementia (“HAD”), AMGs may be used as markers in methods for screening for drugs that treat or prevent HAD.
    Type: Application
    Filed: April 13, 2009
    Publication date: August 13, 2009
    Inventors: Paul B. Fisher, Zhao-zhong Su, Dong-chul Kang
  • Patent number: 7517973
    Abstract: The present invention relates to Astrocyte Enhanced Genes (AEGs), the expression of which is upregulated in human astrocytes grown in primary cell cultures that are exposed to either the human immunodeficiency virus (HIV-1) or to the HIV-1 glycoprotein gp120.
    Type: Grant
    Filed: October 27, 2004
    Date of Patent: April 14, 2009
    Inventors: Paul B Fisher, Zhao-zhong Su, Dong-chul Kang
  • Publication number: 20080108067
    Abstract: The present invention relates to the human Retinoic Acid Inducible Gene-1 (hereafter, “RIG-1”) promoter. The present invention provides for the promoter itself which is inducible by interferon, virus infection, retinoic acid and double-stranded RNA. A further embodiment includes expression constructs comprising the RIG-1 promoter operatively linked to a gene of interest, which may be a reporter gene or a therapeutic gene. It also provides for cells and non-human transgenic animals comprising such expression constructs. In addition, the present invention provides for methods of screening agents for anti-viral activity using RIG-1 promoter activation or inhibition as the basis of the screening system.
    Type: Application
    Filed: April 19, 2007
    Publication date: May 8, 2008
    Inventors: Paul B. Fisher, Zhao-zhong Su
  • Publication number: 20060127980
    Abstract: The nucleic acid sequence of the human Excitatory Amino Acid Transporter-2 Gene (hEAAT2) promoter, a nucleic acid sequence that hybridizes to the hEAAT2 promoter nucleic acid sequence under stringent hybridization conditions, and a nucleic acid sequence that is functionally equivalent to the hEAAT2 promoter sequence are provided, as are vectors containing these nucleic acid sequences. In addition, methods for the use of these nucleic acids to achieve tissue- or cell-specific gene expression are provided, as are methods for the use of these hEAAT2 promoter nucleic acids to identify agents that can modulate glutamate transport or the activity of the glutamate promoter. Such agents may be useful in the prevention, palliation or treatment of neurodegenerative and/or cerebrovascular diseases.
    Type: Application
    Filed: August 4, 2005
    Publication date: June 15, 2006
    Inventors: Paul Fisher, Zhao-zhong Su