Patents by Inventor Harrison Lynch

Harrison Lynch 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).

  • Training a deep neural network model to generate rich object-centric embeddings of robotic vision data
    Patent number: 11887363
    Abstract: Training a machine learning model (e.g., a neural network model such as a convolutional neural network (CNN) model) so that, when trained, the model can be utilized in processing vision data (e.g., from a vision component of a robot), that captures an object, to generate a rich object-centric embedding for the vision data. The generated embedding can enable differentiation of even subtle variations of attributes of the object captured by the vision data.
    Type: Grant
    Filed: September 27, 2019
    Date of Patent: January 30, 2024
    Assignee: GOOGLE LLC
    Inventors: Soeren Pirk, Yunfei Bai, Pierre Sermanet, Seyed Mohammad Khansari Zadeh, Harrison Lynch
  • OPTIMIZING POLICY CONTROLLERS FOR ROBOTIC AGENTS USING IMAGE EMBEDDINGS
    Publication number: 20230150127
    Abstract: There are provided systems, methods, and apparatus, for optimizing a policy controller to control a robotic agent that interacts with an environment to perform a robotic task. One of the methods includes optimizing the policy controller using a neural network that generates numeric embeddings of images of the environment and a demonstration sequence of demonstration images of another agent performing a version of the robotic task.
    Type: Application
    Filed: January 23, 2023
    Publication date: May 18, 2023
    Inventors: YEVGEN CHEBOTAR, Pierre Sermanet, Harrison Lynch
  • Optimizing policy controllers for robotic agents using image embeddings
    Patent number: 11559887
    Abstract: There are provided systems, methods, and apparatus, for optimizing a policy controller to control a robotic agent that interacts with an environment to perform a robotic task. One of the methods includes optimizing the policy controller using a neural network that generates numeric embeddings of images of the environment and a demonstration sequence of demonstration images of another agent performing a version of the robotic task.
    Type: Grant
    Filed: September 20, 2018
    Date of Patent: January 24, 2023
    Assignee: Google LLC
    Inventors: Yevgen Chebotar, Pierre Sermanet, Harrison Lynch
  • TRAINING A DEEP NEURAL NETWORK MODEL TO GENERATE RICH OBJECT-CENTRIC EMBEDDINGS OF ROBOTIC VISION DATA
    Publication number: 20210334599
    Abstract: Training a machine learning model (e.g., a neural network model such as a convolutional neural network (CNN) model) so that, when trained, the model can be utilized in processing vision data (e.g., from a vision component of a robot), that captures an object, to generate a rich object-centric embedding for the vision data. The generated embedding can enable differentiation of even subtle variations of attributes of the object captured by the vision data.
    Type: Application
    Filed: September 27, 2019
    Publication date: October 28, 2021
    Inventors: Soeren Pirk, Yunfei Bai, Pierre Sermanet, Seyed Mohammad Khansari Zadeh, Harrison Lynch
  • OPTIMIZING POLICY CONTROLLERS FOR ROBOTIC AGENTS USING IMAGE EMBEDDINGS
    Publication number: 20200276703
    Abstract: There are provided systems, methods, and apparatus, for optimizing a policy controller to control a robotic agent that interacts with an environment to perform a robotic task. One of the methods includes optimizing the policy controller using a neural network that generates numeric embeddings of images of the environment and a demonstration sequence of demonstration images of another agent performing a version of the robotic task.
    Type: Application
    Filed: September 20, 2018
    Publication date: September 3, 2020
    Inventors: Yevgen Chebotar, Pierre Sermanet, Harrison Lynch