Patents by Inventor Christian Szegedy

Christian Szegedy 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: 12125257
    Abstract: A neural network system that includes: multiple subnetworks that includes: a first subnetwork including multiple first modules, each first module including: a pass-through convolutional layer configured to process the subnetwork input for the first subnetwork to generate a pass-through output; an average pooling stack of neural network layers that collectively processes the subnetwork input for the first subnetwork to generate an average pooling output; a first stack of convolutional neural network layers configured to collectively process the subnetwork input for the first subnetwork to generate a first stack output; a second stack of convolutional neural network layers that are configured to collectively process the subnetwork input for the first subnetwork to generate a second stack output; and a concatenation layer configured to concatenate the pass-through output, the average pooling output, the first stack output, and the second stack output to generate a first module output for the first module.
    Type: Grant
    Filed: July 9, 2021
    Date of Patent: October 22, 2024
    Assignee: Google LLC
    Inventors: Vincent O. Vanhoucke, Christian Szegedy, Sergey Ioffe
  • Patent number: 11809955
    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for image processing using deep neural networks. One of the methods includes receiving data characterizing an input image; processing the data characterizing the input image using a deep neural network to generate an alternative representation of the input image, wherein the deep neural network comprises a plurality of subnetworks, wherein the subnetworks are arranged in a sequence from lowest to highest, and wherein processing the data characterizing the input image using the deep neural network comprises processing the data through each of the subnetworks in the sequence; and processing the alternative representation of the input image through an output layer to generate an output from the input image.
    Type: Grant
    Filed: September 28, 2022
    Date of Patent: November 7, 2023
    Assignee: Google LLC
    Inventors: Christian Szegedy, Vincent O. Vanhoucke
  • Patent number: 11651218
    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for adversarial training of a neural network. One of the methods includes obtaining a plurality of training inputs; and training the neural network on each of the training inputs, comprising, for each of the training inputs: processing the training input using the neural network to determine a neural network output for the training input; applying a perturbation to the training input to generate an adversarial perturbation of the training input; processing the adversarial perturbation of the training input using the neural network to determine a neural network output for the adversarial perturbation; and adjusting the current values of the parameters of the neural network by performing an iteration of a neural network training procedure to optimize an adversarial objective function.
    Type: Grant
    Filed: August 15, 2022
    Date of Patent: May 16, 2023
    Assignee: Google LLC
    Inventors: Christian Szegedy, Ian Goodfellow
  • Publication number: 20230014634
    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for image processing using deep neural networks. One of the methods includes receiving data characterizing an input image; processing the data characterizing the input image using a deep neural network to generate an alternative representation of the input image, wherein the deep neural network comprises a plurality of subnetworks, wherein the subnetworks are arranged in a sequence from lowest to highest, and wherein processing the data characterizing the input image using the deep neural network comprises processing the data through each of the subnetworks in the sequence; and processing the alternative representation of the input image through an output layer to generate an output from the input image.
    Type: Application
    Filed: September 28, 2022
    Publication date: January 19, 2023
    Inventors: Christian Szegedy, Vincent O. Vanhoucke
  • Patent number: 11462035
    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for image processing using deep neural networks. One of the methods includes receiving data characterizing an input image; processing the data characterizing the input image using a deep neural network to generate an alternative representation of the input image, wherein the deep neural network comprises a plurality of subnetworks, wherein the subnetworks are arranged in a sequence from lowest to highest, and wherein processing the data characterizing the input image using the deep neural network comprises processing the data through each of the subnetworks in the sequence; and processing the alternative representation of the input image through an output layer to generate an output from the input image.
    Type: Grant
    Filed: March 12, 2021
    Date of Patent: October 4, 2022
    Assignee: Google LLC
    Inventors: Christian Szegedy, Vincent O. Vanhoucke
  • Patent number: 11416745
    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for adversarial training of a neural network. One of the methods includes obtaining a plurality of training inputs; and training the neural network on each of the training inputs, comprising, for each of the training inputs: processing the training input using the neural network to determine a neural network output for the training input; applying a perturbation to the training input to generate an adversarial perturbation of the training input; processing the adversarial perturbation of the training input using the neural network to determine a neural network output for the adversarial perturbation; and adjusting the current values of the parameters of the neural network by performing an iteration of a neural network training procedure to optimize an adversarial objective function.
    Type: Grant
    Filed: November 22, 2019
    Date of Patent: August 16, 2022
    Assignee: Google LLC
    Inventors: Christian Szegedy, Ian Goodfellow
  • Publication number: 20210334605
    Abstract: A neural network system that includes: multiple subnetworks that includes: a first subnetwork including multiple first modules, each first module including: a pass-through convolutional layer configured to process the subnetwork input for the first subnetwork to generate a pass-through output; an average pooling stack of neural network layers that collectively processes the subnetwork input for the first subnetwork to generate an average pooling output; a first stack of convolutional neural network layers configured to collectively process the subnetwork input for the first subnetwork to generate a first stack output; a second stack of convolutional neural network layers that are configured to collectively process the subnetwork input for the first subnetwork to generate a second stack output; and a concatenation layer configured to concatenate the pass-through output, the average pooling output, the first stack output, and the second stack output to generate a first module output for the first module.
    Type: Application
    Filed: July 9, 2021
    Publication date: October 28, 2021
    Inventors: Vincent O. Vanhoucke, Christian Szegedy, Sergey Ioffe
  • Patent number: 11062181
    Abstract: A neural network system that includes: multiple subnetworks that includes: a first subnetwork including multiple first modules, each first module including: a pass-through convolutional layer configured to process the subnetwork input for the first subnetwork to generate a pass-through output; an average pooling stack of neural network layers that collectively processes the subnetwork input for the first subnetwork to generate an average pooling output; a first stack of convolutional neural network layers configured to collectively process the subnetwork input for the first subnetwork to generate a first stack output; a second stack of convolutional neural network layers that are configured to collectively process the subnetwork input for the first subnetwork to generate a second stack output; and a concatenation layer configured to concatenate the pass-through output, the average pooling output, the first stack output, and the second stack output to generate a first module output for the first module.
    Type: Grant
    Filed: August 26, 2019
    Date of Patent: July 13, 2021
    Assignee: Google LLC
    Inventors: Vincent O. Vanhoucke, Christian Szegedy, Sergey Ioffe
  • Publication number: 20210201092
    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for image processing using deep neural networks. One of the methods includes receiving data characterizing an input image; processing the data characterizing the input image using a deep neural network to generate an alternative representation of the input image, wherein the deep neural network comprises a plurality of subnetworks, wherein the subnetworks are arranged in a sequence from lowest to highest, and wherein processing the data characterizing the input image using the deep neural network comprises processing the data through each of the subnetworks in the sequence; and processing the alternative representation of the input image through an output layer to generate an output from the input image.
    Type: Application
    Filed: March 12, 2021
    Publication date: July 1, 2021
    Inventors: Christian Szegedy, Vincent O. Vanhoucke
  • Patent number: 10977529
    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for image processing using deep neural networks. One of the methods includes receiving data characterizing an input image; processing the data characterizing the input image using a deep neural network to generate an alternative representation of the input image, wherein the deep neural network comprises a plurality of subnetworks, wherein the subnetworks are arranged in a sequence from lowest to highest, and wherein processing the data characterizing the input image using the deep neural network comprises processing the data through each of the subnetworks in the sequence; and processing the alternative representation of the input image through an output layer to generate an output from the input image.
    Type: Grant
    Filed: April 13, 2020
    Date of Patent: April 13, 2021
    Assignee: Google LLC
    Inventors: Christian Szegedy, Vincent O. Vanhoucke
  • Publication number: 20200311491
    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for image processing using deep neural networks. One of the methods includes receiving data characterizing an input image; processing the data characterizing the input image using a deep neural network to generate an alternative representation of the input image, wherein the deep neural network comprises a plurality of subnetworks, wherein the subnetworks are arranged in a sequence from lowest to highest, and wherein processing the data characterizing the input image using the deep neural network comprises processing the data through each of the subnetworks in the sequence; and processing the alternative representation of the input image through an output layer to generate an output from the input image.
    Type: Application
    Filed: April 13, 2020
    Publication date: October 1, 2020
    Inventors: Christian Szegedy, Vincent O. Vanhoucke
  • Patent number: 10650289
    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for image processing using deep neural networks. One of the methods includes receiving data characterizing an input image; processing the data characterizing the input image using a deep neural network to generate an alternative representation of the input image, wherein the deep neural network comprises a plurality of subnetworks, wherein the subnetworks are arranged in a sequence from lowest to highest, and wherein processing the data characterizing the input image using the deep neural network comprises processing the data through each of the subnetworks in the sequence; and processing the alternative representation of the input image through an output layer to generate an output from the input image.
    Type: Grant
    Filed: January 11, 2018
    Date of Patent: May 12, 2020
    Assignee: Google LLC
    Inventors: Christian Szegedy, Vincent O. Vanhoucke
  • Patent number: 10521718
    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for adversarial training of a neural network. One of the methods includes obtaining a plurality of training inputs; and training the neural network on each of the training inputs, comprising, for each of the training inputs: processing the training input using the neural network to determine a neural network output for the training input; applying a perturbation to the training input to generate an adversarial perturbation of the training input; processing the adversarial perturbation of the training input using the neural network to determine a neural network output for the adversarial perturbation; and adjusting the current values of the parameters of the neural network by performing an iteration of a neural network training procedure to optimize an adversarial objective function.
    Type: Grant
    Filed: September 28, 2016
    Date of Patent: December 31, 2019
    Assignee: Google LLC
    Inventors: Christian Szegedy, Ian Goodfellow
  • Publication number: 20190377985
    Abstract: A neural network system that includes: multiple subnetworks that includes: a first subnetwork including multiple first modules, each first module including: a pass-through convolutional layer configured to process the subnetwork input for the first subnetwork to generate a pass-through output; an average pooling stack of neural network layers that collectively processes the subnetwork input for the first subnetwork to generate an average pooling output; a first stack of convolutional neural network layers configured to collectively process the subnetwork input for the first subnetwork to generate a first stack output; a second stack of convolutional neural network layers that are configured to collectively process the subnetwork input for the first subnetwork to generate a second stack output; and a concatenation layer configured to concatenate the pass-through output, the average pooling output, the first stack output, and the second stack output to generate a first module output for the first module.
    Type: Application
    Filed: August 26, 2019
    Publication date: December 12, 2019
    Inventors: Vincent O. Vanhoucke, Christian Szegedy, Sergey Ioffe
  • Patent number: 10467493
    Abstract: Systems, methods, and apparatus, including computer programs encoded on a computer storage medium. In one aspect, a system includes initial neural network layers configured to: receive an input image, and process the input image to generate a plurality of first feature maps that characterize the input image; a location generating convolutional neural network layer configured to perform a convolution on the representation of the first plurality of feature maps to generate data defining a respective location of each of a predetermined number of bounding boxes in the input image, wherein each bounding box identifies a respective first region of the input image; and a confidence score generating convolutional neural network layer configured to perform a convolution on the representation of the first plurality of feature maps to generate a confidence score for each of the predetermined number of bounding boxes in the input image.
    Type: Grant
    Filed: July 14, 2017
    Date of Patent: November 5, 2019
    Assignee: Google LLC
    Inventors: Dominik Roblek, Christian Szegedy, Jacek Slawosz Jurewicz
  • Patent number: 10460211
    Abstract: A neural network system that includes: multiple subnetworks that includes: a first subnetwork including multiple first modules, each first module including: a pass-through convolutional layer configured to process the subnetwork input for the first subnetwork to generate a pass-through output; an average pooling stack of neural network layers that collectively processes the subnetwork input for the first subnetwork to generate an average pooling output; a first stack of convolutional neural network layers configured to collectively process the subnetwork input for the first subnetwork to generate a first stack output; a second stack of convolutional neural network layers that are configured to collectively process the subnetwork input for the first subnetwork to generate a second stack output; and a concatenation layer configured to concatenate the pass-through output, the average pooling output, the first stack output, and the second stack output to generate a first module output for the first module.
    Type: Grant
    Filed: December 30, 2016
    Date of Patent: October 29, 2019
    Assignee: Google LLC
    Inventors: Vincent O. Vanhoucke, Christian Szegedy, Sergey Ioffe
  • Publication number: 20190019050
    Abstract: Systems, methods, and apparatus, including computer programs encoded on a computer storage medium. In one aspect, a system includes initial neural network layers configured to: receive an input image, and process the input image to generate a plurality of first feature maps that characterize the input image; a location generating convolutional neural network layer configured to perform a convolution on the representation of the first plurality of feature maps to generate data defining a respective location of each of a predetermined number of bounding boxes in the input image, wherein each bounding box identifies a respective first region of the input image; and a confidence score generating convolutional neural network layer configured to perform a convolution on the representation of the first plurality of feature maps to generate a confidence score for each of the predetermined number of bounding boxes in the input image.
    Type: Application
    Filed: July 14, 2017
    Publication date: January 17, 2019
    Inventors: Dominik Roblek, Christian Szegedy, Jacek Slawosz Jurewicz
  • Patent number: 10007867
    Abstract: Systems and methods of identifying entities are disclosed. In particular, one or more images that depict an entity can be identified from a plurality of images. One or more candidate entity profiles can be determined from an entity directory based at least in part on the one or more images that depict the entity. The one or more images that depict the entity and the one or more candidate entity profiles can be provided as input to a machine learning model. One or more outputs of the machine learning model can be generated. Each output can include a match score associated with an image that depicts the entity and at least one candidate entity profile. The entity directory can be updated based at least in part on the one or more generated outputs of the machine learning model.
    Type: Grant
    Filed: April 4, 2016
    Date of Patent: June 26, 2018
    Assignee: Google LLC
    Inventors: Qian Yu, Liron Yatziv, Yeqing Li, Christian Szegedy, Sacha Christopher Arnoud, Martin C. Stumpe
  • Publication number: 20180137396
    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for image processing using deep neural networks. One of the methods includes receiving data characterizing an input image; processing the data characterizing the input image using a deep neural network to generate an alternative representation of the input image, wherein the deep neural network comprises a plurality of subnetworks, wherein the subnetworks are arranged in a sequence from lowest to highest, and wherein processing the data characterizing the input image using the deep neural network comprises processing the data through each of the subnetworks in the sequence; and processing the alternative representation of the input image through an output layer to generate an output from the input image.
    Type: Application
    Filed: January 11, 2018
    Publication date: May 17, 2018
    Inventors: Christian Szegedy, Vincent O. Vanhoucke
  • Publication number: 20180068207
    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for image processing using deep neural networks. One of the methods includes receiving data characterizing an input image; processing the data characterizing the input image using a deep neural network to generate an alternative representation of the input image, wherein the deep neural network comprises a plurality of subnetworks, wherein the subnetworks are arranged in a sequence from lowest to highest, and wherein processing the data characterizing the input image using the deep neural network comprises processing the data through each of the subnetworks in the sequence; and processing the alternative representation of the input image through an output layer to generate an output from the input image.
    Type: Application
    Filed: November 10, 2017
    Publication date: March 8, 2018
    Inventors: Christian Szegedy, Vincent O. Vanhoucke