Patents by Inventor Simon Osindero
Simon Osindero 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: 11967150Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for parallel processing of video frames using neural networks. One of the methods includes receiving a video sequence comprising a respective video frame at each of a plurality of time steps; and processing the video sequence using a video processing neural network to generate a video processing output for the video sequence, wherein the video processing neural network includes a sequence of network components, wherein the network components comprise a plurality of layer blocks each comprising one or more neural network layers, wherein each component is active for a respective subset of the plurality of time steps, and wherein each layer block is configured to, at each time step at which the layer block is active, receive an input generated at a previous time step and to process the input to generate a block output.Type: GrantFiled: February 13, 2023Date of Patent: April 23, 2024Assignee: DeepMind Technologies LimitedInventors: Simon Osindero, Joao Carreira, Viorica Patraucean, Andrew Zisserman
-
Patent number: 11715009Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training a neural network including a first subnetwork followed by a second subnetwork on training inputs by optimizing an objective function. In one aspect, a method includes processing a training input using the neural network to generate a training model output, including processing a subnetwork input for the training input using the first subnetwork to generate a subnetwork activation for the training input in accordance with current values of parameters of the first subnetwork, and providing the subnetwork activation as input to the second subnetwork; determining a synthetic gradient of the objective function for the first subnetwork by processing the subnetwork activation using a synthetic gradient model in accordance with current values of parameters of the synthetic gradient model; and updating the current values of the parameters of the first subnetwork using the synthetic gradient.Type: GrantFiled: May 19, 2017Date of Patent: August 1, 2023Assignee: DeepMind Technologies LimitedInventors: Oriol Vinyals, Alexander Benjamin Graves, Wojciech Czarnecki, Koray Kavukcuoglu, Simon Osindero, Maxwell Elliot Jaderberg
-
Publication number: 20230186625Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for parallel processing of video frames using neural networks. One of the methods includes receiving a video sequence comprising a respective video frame at each of a plurality of time steps; and processing the video sequence using a video processing neural network to generate a video processing output for the video sequence, wherein the video processing neural network includes a sequence of network components, wherein the network components comprise a plurality of layer blocks each comprising one or more neural network layers, wherein each component is active for a respective subset of the plurality of time steps, and wherein each layer block is configured to, at each time step at which the layer block is active, receive an input generated at a previous time step and to process the input to generate a block output.Type: ApplicationFiled: February 13, 2023Publication date: June 15, 2023Inventors: Simon Osindero, Joao Carreira, Viorica Patraucean, Andrew Zisserman
-
Publication number: 20230124177Abstract: A computer-implemented method of training a neural network. The method comprises repeatedly determining a forward-pass set of network parameters by selecting a first sub-set of parameters of the neural network and setting all other parameters of the forward-pass set of network parameters to zero. The method then processes a training data item using the neural network in accordance with the forward-pass set of network parameters to generate a neural network output, determines a value of an objective function from the neural network output and the training data item, selects a second sub-set of network parameters, determines a backward-pass set of network parameters comprising the first and second sub-sets of parameters, and updates parameters corresponding to the backward-pass set of network parameters using a gradient estimate determined from the value of the objective function.Type: ApplicationFiled: June 4, 2021Publication date: April 20, 2023Inventors: Siddhant Madhu Jayakumar, Razvan Pascanu, Jack William Rae, Simon Osindero, Erich Konrad Elsen
-
Publication number: 20230090824Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for a system configured to select actions to be performed by an agent that interacts with an environment. The system comprises a manager neural network subsystem and a worker neural network subsystem. The manager subsystem is configured to, at each of the multiple time steps, generate a final goal vector for the time step. The worker subsystem is configured to, at each of multiple time steps, use the final goal vector generated by the manager subsystem to generate a respective action score for each action in a predetermined set of actions.Type: ApplicationFiled: November 30, 2022Publication date: March 23, 2023Inventors: Simon Osindero, Koray Kavukcuoglu, Alexander Vezhnevets
-
Patent number: 11580736Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for parallel processing of video frames using neural networks. One of the methods includes receiving a video sequence comprising a respective video frame at each of a plurality of time steps; and processing the video sequence using a video processing neural network to generate a video processing output for the video sequence, wherein the video processing neural network includes a sequence of network components, wherein the network components comprise a plurality of layer blocks each comprising one or more neural network layers, wherein each component is active for a respective subset of the plurality of time steps, and wherein each layer block is configured to, at each time step at which the layer block is active, receive an input generated at a previous time step and to process the input to generate a block output.Type: GrantFiled: January 7, 2019Date of Patent: February 14, 2023Assignee: DeepMind Technologies LimitedInventors: Simon Osindero, Joao Carreira, Viorica Patraucean, Andrew Zisserman
-
Patent number: 11537887Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for a system configured to select actions to be performed by an agent that interacts with an environment. The system comprises a manager neural network subsystem and a worker neural network subsystem. The manager subsystem is configured to, at each of the multiple time steps, generate a final goal vector for the time step. The worker subsystem is configured to, at each of multiple time steps, use the final goal vector generated by the manager subsystem to generate a respective action score for each action in a predetermined set of actions.Type: GrantFiled: May 5, 2020Date of Patent: December 27, 2022Assignee: DeepMind Technologies LimitedInventors: Simon Osindero, Koray Kavukcuoglu, Alexander Vezhnevets
-
Publication number: 20220392206Abstract: A system that is configured to receive a sequence of task inputs and to perform a machine learning task is described. The system includes a reinforcement learning (RL) neural network and a task neural network. The RL neural network is configured to: generate, for each task input of the sequence of task inputs, a respective decision that determines whether to encode the task input or to skip the task input, and provide the respective decision of each task input to the task neural network.Type: ApplicationFiled: November 13, 2020Publication date: December 8, 2022Inventors: Viorica PATRAUCEAN, Bilal PIOT, Joao CARREIRA, Volodymyr MNIH, Simon OSINDERO
-
Publication number: 20220180147Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for implementing associative memory. In one aspect a system comprises an associative memory neural network to process an input to generate an output that defines an energy corresponding to the input. A reading subsystem retrieves stored information from the associative memory neural network. The reading subsystem performs operations including receiving a given, i.e. query, input and retrieving a data element from the associative memory neural network that is associated with the given input. The retrieving is performed by iteratively adjusting the given input using the associative memory neural network.Type: ApplicationFiled: May 19, 2020Publication date: June 9, 2022Inventors: Sergey Bartunov, Jack William Rae, Timothy Paul Lillicrap, Simon Osindero
-
Publication number: 20210089908Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for controlling an agent. One of the methods includes sampling a behavior modulation in accordance with a current probability distribution; for each of one or more time steps: processing an input comprising an observation characterizing a current state of the environment at the time step using an action selection neural network to generate a respective action score for each action in a set of possible actions that can be performed by the agent; modifying the action scores using the sampled behavior modulation; and selecting the action to be performed by the agent at the time step based on the modified action scores; determining a fitness measure corresponding to the sampled behavior modulation; and updating the current probability distribution over the set of possible behavior modulations using the fitness measure corresponding to the behavior modulation.Type: ApplicationFiled: September 25, 2020Publication date: March 25, 2021Inventors: Tom Schaul, Diana Luiza Borsa, Fengning Ding, David Szepesvari, Georg Ostrovski, Simon Osindero, William Clinton Dabney
-
Publication number: 20210027064Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for parallel processing of video frames using neural networks. One of the methods includes receiving a video sequence comprising a respective video frame at each of a plurality of time steps; and processing the video sequence using a video processing neural network to generate a video processing output for the video sequence, wherein the video processing neural network includes a sequence of network components, wherein the network components comprise a plurality of layer blocks each comprising one or more neural network layers, wherein each component is active for a respective subset of the plurality of time steps, and wherein each layer block is configured to, at each time step at which the layer block is active, receive an input generated at a previous time step and to process the input to generate a block output.Type: ApplicationFiled: January 7, 2019Publication date: January 28, 2021Inventors: Simon Osindero, Joao Carreira, Viorica Patraucean, Andrew Zisserman
-
Publication number: 20200320396Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training a neural network including a first subnetwork followed by a second subnetwork on training inputs by optimizing an objective function. In one aspect, a method includes processing a training input using the neural network to generate a training model output, including processing a subnetwork input for the training input using the first subnetwork to generate a subnetwork activation for the training input in accordance with current values of parameters of the first subnetwork, and providing the subnetwork activation as input to the second subnetwork; determining a synthetic gradient of the objective function for the first subnetwork by processing the subnetwork activation using a synthetic gradient model in accordance with current values of parameters of the synthetic gradient model; and updating the current values of the parameters of the first subnetwork using the synthetic gradient.Type: ApplicationFiled: May 19, 2017Publication date: October 8, 2020Applicant: Deepmind Technologies LimitedInventors: Oriol VINYALS, Alexander Benjamin GRAVES, Wojciech CZARNECKI, Koray KAVUKCUOGLU, Simon OSINDERO, Maxwell Elliot JADERBERG
-
Publication number: 20200265313Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for a system configured to select actions to be performed by an agent that interacts with an environment. The system comprises a manager neural network subsystem and a worker neural network subsystem. The manager subsystem is configured to, at each of the multiple time steps, generate a final goal vector for the time step. The worker subsystem is configured to, at each of multiple time steps, use the final goal vector generated by the manager subsystem to generate a respective action score for each action in a predetermined set of actions.Type: ApplicationFiled: May 5, 2020Publication date: August 20, 2020Inventors: Simon Osindero, Koray Kavukcuoglu, Alexander Vezhnevets
-
Patent number: 10706889Abstract: One or more computing devices, systems, and/or methods for selective content insertion into areas of media objects are provided. For example, a media object (e.g., an image or video), is selected for composition with content, such as where a message, interactive content, a hyperlink, or other types of content is overlaid or embedded into the media object to create a composite media object. The content is added into an area of the media object that is selectively identified to reduce occlusion and/or improve visual cohesiveness between the content and the media object (e.g., added to an area with a similar or complimentary color, having an adequate size with spare amounts of visual features such as a soccer player, a ball, or other entity, etc.). In this way, the content may be add into the area of the media object to create a composite media object to provide to users.Type: GrantFiled: July 7, 2016Date of Patent: July 7, 2020Assignee: Oath Inc.Inventors: Bart Thomée, Ioannis Kalantidis, Clayton Ellis Mellina, Simon Osindero
-
Patent number: 10679126Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for a system configured to select actions to be performed by an agent that interacts with an environment. The system comprises a manager neural network subsystem and a worker neural network subsystem. The manager subsystem is configured to, at each of the multiple time steps, generate a final goal vector for the time step. The worker subsystem is configured to, at each of multiple time steps, use the final goal vector generated by the manager subsystem to generate a respective action score for each action in a predetermined set of actions.Type: GrantFiled: July 15, 2019Date of Patent: June 9, 2020Assignee: DeepMind Technologies LimitedInventors: Simon Osindero, Koray Kavukcuoglu, Alexander Vezhnevets
-
Patent number: 10558720Abstract: Disclosure includes system, method and architecture for selecting supplemental digital content using visual appearance. Digital content that is visually similar, or dissimilar, to digital content requested by a user can he identified and provided for presentation with the requested content. The additional digital content is visually congruent, or visually incongruent, with content requested by a user, such that the additional content is similar, or dissimilar, to the requested content from a visual standpoint. In a presentation of the requested content, the presentation position of each additional content item relative to the presentation position of each requested content can be identified using visual congruence, or visual incongruence.Type: GrantFiled: April 13, 2016Date of Patent: February 11, 2020Assignee: OATH INC.Inventors: Simon Osindero, Ioannis Kalantidis, David Shamma, Lyndon Kennedy, Ayman Farahat, Gerry Pesavento, Ricardo Baeza-Yates
-
Patent number: 10534981Abstract: Disclosed herein is an intelligent agent to analyze a media object. The agent comprises a trained model comprising a number of state layers for storing a history of actions taken by the agent in each of a number of previous iterations performed by the agent in analyzing a media object. The stored state may be used by the agent in a current iteration to determine whether or not to make, or abstain from making, a prediction from output generated by the model, identify another portion of the media object to analyze, end analysis. Output from the agent's model may comprise a semantic vector that can be mapped to a semantic vector space to identify a number of labels for a media object.Type: GrantFiled: April 6, 2018Date of Patent: January 14, 2020Assignee: OATH INC.Inventor: Simon Osindero
-
Publication number: 20190340509Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for a system configured to select actions to be performed by an agent that interacts with an environment. The system comprises a manager neural network subsystem and a worker neural network subsystem. The manager subsystem is configured to, at each of the multiple time steps, generate a final goal vector for the time step. The worker subsystem is configured to, at each of multiple time steps, use the final goal vector generated by the manager subsystem to generate a respective action score for each action in a predetermined set of actions.Type: ApplicationFiled: July 15, 2019Publication date: November 7, 2019Inventors: Simon Osindero, Koray Kavukcuoglu, Alexander Vezhnevets
-
Publication number: 20180225549Abstract: Disclosed herein is an intelligent agent to analyze a media object. The agent comprises a trained model comprising a number of state layers for storing a history of actions taken by the agent in each of a number of previous iterations performed by the agent in analyzing a media object. The stored state may be used by the agent in a current iteration to determine whether or not to make, or abstain from making, a prediction from output generated by the model, identify another portion of the media object to analyze, end analysis. Output from the agent's model may comprise a semantic vector that can be mapped to a semantic vector space to identify a number of labels for a media object.Type: ApplicationFiled: April 6, 2018Publication date: August 9, 2018Inventor: Simon OSINDERO
-
Patent number: 10043231Abstract: Images that comprise text are identified and output from the images is generated wherein the output comprises text from the image in textual data format. The portions of an image comprising the text data are initially identified and the text imaged by the pixels of that image portion is extracted in textual data format. The extracted text is stored so that a search for images comprising particular text is enabled.Type: GrantFiled: June 30, 2015Date of Patent: August 7, 2018Assignee: OATH INC.Inventor: Simon Osindero