Patents by Inventor Yulun WU

Yulun WU 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).

  • Patent number: 11682194
    Abstract: A training method for a robust neural network based on feature matching is provided in this disclosure, which includes following steps. Step A, a first stage model is initialized. The first stage model includes a backbone network, a feature matching module and a fullple loss function. Step B, the first stage model is trained by using original training data to obtain a second stage model. Step C, the second stage model is attacked so as to generate PGD adversarial samples of the original training data, and the second stage model is trained again with the generated adversarial samples and the original training data. Step D, training parameters are adjusted and the second stage model is trained again, and parameters for which the model has highest accuracy on an original test set are saved.
    Type: Grant
    Filed: September 22, 2022
    Date of Patent: June 20, 2023
    Assignee: National University of Defense Technology
    Inventors: Yanming Guo, Jian Li, Songyang Lao, Liang Bai, Yingmei Wei, Yulun Wu
  • Publication number: 20230089335
    Abstract: A training method for a robust neural network based on feature matching is provided in this disclosure, which includes following steps. Step A, a first stage model is initialized. The first stage model includes a backbone network, a feature matching module and a fullple loss function. Step B, the first stage model is trained by using original training data to obtain a second stage model. Step C, the second stage model is attacked so as to generate PGD adversarial samples of the original training data, and the second stage model is trained again with the generated adversarial samples and the original training data. Step D, training parameters are adjusted and the second stage model is trained again, and parameters for which the model has highest accuracy on an original test set are saved.
    Type: Application
    Filed: September 22, 2022
    Publication date: March 23, 2023
    Inventors: Yanming GUO, Jian LI, Songyang LAO, Liang BAI, Yingmei WEI, Yulun WU
  • Publication number: 20210405674
    Abstract: The present disclosure provides a low dropout regulator circuit including an error amplifying circuit, a charge pump circuit, and an output feedback circuit. An output terminal of the error amplifying circuit is connected to an input terminal of the charge pump circuit. The charge pump circuit includes a first switch module, a first energy storage module, a third switch module, a parasitic capacitor, a second switch module, a second energy storage module, and a clock control circuit which is respectively connected to the first switch module, the third switch module, and the second switch module to control turn-on and turn-off of the first switch module, the third switch module and the second switch module. An output terminal of the charge pump circuit is connected to the output feedback circuit, and the output feedback circuit is connected to the error amplifying circuit, thereby providing a higher voltage, and reducing voltage fluctuations.
    Type: Application
    Filed: September 8, 2021
    Publication date: December 30, 2021
    Inventors: Jianfeng XUE, Mengwen ZHANG, Yunning LI, Yulun WU
  • Patent number: D985176
    Type: Grant
    Filed: November 1, 2022
    Date of Patent: May 2, 2023
    Inventor: Yulun Wu