Patents by Inventor Jiayi Liu

Jiayi Liu 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: 11096374
    Abstract: A pet house accessory with functions of air purification, sterilization and temperature regulation, including a shell, a mounting plate, a thermal conduction device, a semiconductor refrigeration sheet and a sterilization device, wherein a heat dissipating hole is arranged at the center of the bottom of the shell, the thermal conduction device is arranged at the center of the bottom of the mounting plate, the semiconductor refrigeration sheet is fixedly installed at the bottom of the thermal conduction device, and a through hole is defined inside the thermal conduction device. The pet house accessory with functions of air purification, sterilization and temperature regulation provided by the present disclosure is capable of cooling and heating by controlling the semiconductor refrigeration sheet through the control switch to cool down and warm the pet house, and is convenient for pets to live.
    Type: Grant
    Filed: January 27, 2021
    Date of Patent: August 24, 2021
    Inventor: Jiayi Liu
  • Publication number: 20210110274
    Abstract: A computer-implemented method includes: initializing model parameters for training a neural network; performing a forward pass and backpropagation for a first minibatch of training data; determining a new weight value for each of a plurality of nodes of the neural network using a gradient descent of the first minibatch; for each determined new weight value, determining whether to update a running mean corresponding to a weight of a particular node; based on a determination to update the running mean, calculating a new mean weight value for the particular node using the determined new weight value; updating the weight parameters for all nodes based on the calculated new mean weight values corresponding to each node; assigning the running mean as the weight for the particular node when training on the first minibatch is completed; and reinitializing running means for all nodes at a start of training a second minibatch.
    Type: Application
    Filed: August 13, 2020
    Publication date: April 15, 2021
    Applicant: LG ELECTRONICS INC.
    Inventors: Samarth TRIPATHI, Jiayi LIU, Unmesh KURUP, Mohak SHAH
  • Publication number: 20200389601
    Abstract: Systems and methods for image registration and self-localization for onsite and offsite viewing are provided. In one aspect, systems and methods for spherical image based registration and self-localization for onsite and offsite viewing, such as augmented reality viewing and virtual reality viewing, are provided. In one embodiment, a system includes a portable electric device fitted with a device camera, an external omnidirectional spherical camera, and a remote server. In one embodiment, a method of use employs a set of fiducial markers inserted in images collected by a spherical camera to perform image registration of spherical camera images and device camera images.
    Type: Application
    Filed: June 1, 2020
    Publication date: December 10, 2020
    Inventors: Nathaniel Boyless, Jiayi Liu, Sriram Siva, Hao Zhang
  • Patent number: 10776668
    Abstract: A search framework for finding effective architectural building blocks for deep convolutional neural networks is disclosed. The search framework described herein utilizes a building block which incorporates branch and skip connections. At least some operations of the architecture of the building block are undefined and treated as hyperparameters which can be automatically selected and optimized for a particular task. The search framework uses random search over the reduced search space to generate a building block and repeats the building block multiple times to create a deep convolutional neural network.
    Type: Grant
    Filed: December 7, 2018
    Date of Patent: September 15, 2020
    Assignee: Robert Bosch GmbH
    Inventors: Jayanta Kumar Dutta, Jiayi Liu, Unmesh Kurup, Mohak Shah
  • Publication number: 20190188537
    Abstract: A search framework for finding effective architectural building blocks for deep convolutional neural networks is disclosed. The search framework described herein utilizes a building block which incorporates branch and skip connections. At least some operations of the architecture of the building block are undefined and treated as hyperparameters which can be automatically selected and optimized for a particular task. The search framework uses random search over the reduced search space to generate a building block and repeats the building block multiple times to create a deep convolutional neural network.
    Type: Application
    Filed: December 7, 2018
    Publication date: June 20, 2019
    Inventors: Jayanta Kumar Dutta, Jiayi Liu, Unmesh Kurup, Mohak Shah
  • Patent number: D941531
    Type: Grant
    Filed: December 22, 2020
    Date of Patent: January 18, 2022
    Inventor: Jiayi Liu