Patents by Inventor Stephen Joseph MERITY

Stephen Joseph MERITY 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).

  • QUASI-RECURRENT NEURAL NETWORK BASED ENCODER-DECODER MODEL
    Publication number: 20210103816
    Abstract: The technology disclosed provides a quasi-recurrent neural network (QRNN) encoder-decoder model that alternates convolutional layers, which apply in parallel across timesteps, and minimalist recurrent pooling layers that apply in parallel across feature dimensions.
    Type: Application
    Filed: December 15, 2020
    Publication date: April 8, 2021
    Inventors: James BRADBURY, Stephen Joseph MERITY, Caiming XIONG, Richard SOCHER
  • QUASI-RECURRENT NEURAL NETWORK BASED ENCODER-DECODER MODEL
    Publication number: 20180129931
    Abstract: The technology disclosed provides a quasi-recurrent neural network (QRNN) encoder-decoder model that alternates convolutional layers, which apply in parallel across timesteps, and minimalist recurrent pooling layers that apply in parallel across feature dimensions.
    Type: Application
    Filed: January 31, 2017
    Publication date: May 10, 2018
    Applicant: salesforce.com, inc.
    Inventors: James BRADBURY, Stephen Joseph MERITY, Caiming XIONG, Richard SOCHER
  • QUASI-RECURRENT NEURAL NETWORK
    Publication number: 20180129937
    Abstract: The technology disclosed provides a quasi-recurrent neural network (QRNN) that alternates convolutional layers, which apply in parallel across timesteps, and minimalist recurrent pooling layers that apply in parallel across feature dimensions.
    Type: Application
    Filed: January 31, 2017
    Publication date: May 10, 2018
    Applicant: salesforce.com, inc.
    Inventors: James BRADBURY, Stephen Joseph MERITY, Caiming XIONG, Richard SOCHER
  • POINTER SENTINEL MIXTURE ARCHITECTURE
    Publication number: 20180082171
    Abstract: The technology disclosed provides a so-called “pointer sentinel mixture architecture” for neural network sequence models that has the ability to either reproduce a token from a recent context or produce a token from a predefined vocabulary. In one implementation, a pointer sentinel-LSTM architecture achieves state of the art language modeling performance of 70.9 perplexity on the Penn Treebank dataset, while using far fewer parameters than a standard softmax LSTM.
    Type: Application
    Filed: January 31, 2017
    Publication date: March 22, 2018
    Applicant: salesforce.com, inc.
    Inventors: Stephen Joseph MERITY, Caiming XIONG, James BRADBURY, Richard SOCHER