Patents by Inventor Joshua Samuel ROMOFF

Joshua Samuel ROMOFF 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).

  • Hybrid reward architecture for reinforcement learning
    Patent number: 10977551
    Abstract: Aspects provided herein are relevant to machine learning techniques, including decomposing single-agent reinforcement learning problems into simpler problems addressed by multiple agents. Actions proposed by the multiple agents are then aggregated using an aggregator, which selects an action to take with respect to an environment. Aspects provided herein are also relevant to a hybrid reward model.
    Type: Grant
    Filed: June 27, 2017
    Date of Patent: April 13, 2021
    Assignee: Microsoft Technology Licensing, LLC
    Inventors: Harm Hendrik Van Seijen, Seyed Mehdi Fatemi Booshehri, Romain Michel Henri Laroche, Joshua Samuel Romoff
  • HYBRID REWARD ARCHITECTURE FOR REINFORCEMENT LEARNING
    Publication number: 20180165603
    Abstract: Aspects provided herein are relevant to machine learning techniques, including decomposing single-agent reinforcement learning problems into simpler problems addressed by multiple agents. Actions proposed by the multiple agents are then aggregated using an aggregator, which selects an action to take with respect to an environment. Aspects provided herein are also relevant to a hybrid reward model.
    Type: Application
    Filed: June 27, 2017
    Publication date: June 14, 2018
    Applicant: Microsoft Technology Licensing, LLC
    Inventors: Harm Hendrik VAN SEIJEN, Seyed Mehdi FATEMI BOOSHEHRI, Romain Michel Henri LAROCHE, Joshua Samuel ROMOFF
  • SCALABILITY OF REINFORCEMENT LEARNING BY SEPARATION OF CONCERNS
    Publication number: 20180165602
    Abstract: Aspects provided herein are relevant to machine learning techniques, including decomposing single-agent reinforcement learning problems into simpler problems addressed by multiple agents. Actions proposed by the multiple agents are then aggregated using an aggregator, which selects an action to take with respect to an environment. Aspects provided herein are also relevant to a hybrid reward model.
    Type: Application
    Filed: June 27, 2017
    Publication date: June 14, 2018
    Applicant: Microsoft Technology Licensing, LLC
    Inventors: Harm Hendrik VAN SEIJEN, Seyed Mehdi FATEMI BOOSHEHRI, Romain Michel Henri LAROCHE, Joshua Samuel ROMOFF