Patents by Inventor Linnan WANG

Linnan WANG 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).

  • GENERATING NEURAL NETWORKS
    Publication number: 20240119267
    Abstract: Apparatuses, systems, and techniques to selectively use one or more neural network layers. In at least one embodiment, one or more neural network layers are selectively used based on, for example, one or more iteratively increasing neural network performance metrics.
    Type: Application
    Filed: September 21, 2022
    Publication date: April 11, 2024
    Inventors: Slawomir Kierat, Piotr Karpinski, Mateusz Sieniawski, Pawel Morkisz, Szymon Migacz, Linnan Wang, Chen-Han Yu, Satish Salian, Ashwath Aithal, Alexandru Fit-Florea
  • Accelerating deep neural network training with inconsistent stochastic gradient descent
    Patent number: 10572800
    Abstract: Aspects of the present disclosure describe techniques for training a convolutional neural network using an inconsistent stochastic gradient descent (ISGD) algorithm. Training effort for training batches used by the ISGD algorithm are dynamically adjusted according to a determined loss for a given training batch which are classified into two sub states—well-trained or under-trained. The ISGD algorithm provides more iterations for under-trained batches while reducing iterations for well-trained ones.
    Type: Grant
    Filed: February 2, 2017
    Date of Patent: February 25, 2020
    Assignee: NEC Corporation
    Inventors: Linnan Wang, Yi Yang, Renqiang Min, Srimat Chakradhar
  • ACCELERATING DEEP NEURAL NETWORK TRAINING WITH INCONSISTENT STOCHASTIC GRADIENT DESCENT
    Publication number: 20170228645
    Abstract: Aspects of the present disclosure describe techniques for training a convolutional neural network using an inconsistent stochastic gradient descent (ISGD) algorithm. Training effort for training batches used by the ISGD algorithm are dynamically adjusted according to a determined loss for a given training batch which are classified into two sub states—well-trained or under-trained. The ISGD algorithm provides more iterations for under-trained batches while reducing iterations for well-trained ones.
    Type: Application
    Filed: February 2, 2017
    Publication date: August 10, 2017
    Inventors: Linnan WANG, Yi YANG, Renqiang MIN, Srimat CHAKRADHAR