Google Inc.

Microsoft Technology Licensing, LLC

Saurabh Gupta

Li Deng

John Carlton Platt

Jianfeng Gao

Xiaodong He

Sergey Levine

Geoffrey G Zweig

Vance Julius Vagell

Rahul Sukthankar

Jeffrey B Hamblin

Margaret Ann Mitchell

Joseph M Sherman

James Christopher Davidson

Rupesh Kumar Srivastava

Forrest Iandola

Jitendra Malik

Justin Neddo

Hao Fang

Mike Harm

Sergey Vladimir Levine

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for navigation using visual inputs. One of the systems includes a mapping subsystem configured to, at each time step of a plurality of time steps, generate a characterization of an environment from an image of the environment at the time step, wherein the characterization comprises an environment map identifying locations in the environment having a particular characteristic, and wherein generating the characterization comprises, for each time step: obtaining the image of the environment at the time step, processing the image to generate a first initial characterization for the time step, obtaining a final characterization for a previous time step, processing the characterization for the previous time step to generate a second initial characterization for the time step, and combining the first initial characterization and the second initial characterization to generate a final characterization for the time step.

Agent navigation using visual inputs

Disclosed herein are technologies directed to discovering semantic similarities between images and text, which can include performing image search using a textual query, performing text search using an image as a query, and/or generating captions for images using a caption generator. A semantic similarity framework can include a caption generator and can be based on a deep multimodal similar model. The deep multimodal similarity model can receive sentences and determine the relevancy of the sentences based on similarity of text vectors generated for one or more sentences to an image vector generated for an image. The text vectors and the image vector can be mapped in a semantic space, and their relevance can be determined based at least in part on the mapping. The sentence associated with the text vector determined to be the most relevant can be output as a caption for the image.

Discovery of semantic similarities between images and text

A method and system for providing modular application for use within a cloud-based file includes storing on a cloud computing service a link to a modular application configured to be used with a cloud-based file, receiving a request to access the link to the modular application, determining whether the request source is associated with the cloud computing service, and providing an option to access the requested modular application. The option can be for a trial of the modular application or to create a cloud-based file using the modular application.

System and method for accessing modular applications

A method for providing applications for use within collaborative documents includes storing, on a cloud computing service, a collaborative document accessible by a plurality of collaborators and a plurality of modular applications for a document editing application used by the plurality of collaborators to view the collaborative document, where the plurality of modular applications are usable within the collaborative document. The cloud computing service receives a request from a first collaborator in the plurality of collaborators to activate the use of a first modular application in the plurality of modular applications within the collaborative document, and activates the use of the first modular application within the collaborative document, where the first modular application executes on the cloud computing service and each of the plurality of collaborators may use the first modular application within the collaborative document.

Collaborative use and management of modular applications

A virtual disk is comprised of segments of unused capacity of physical computer-readable storage media co-located with computing devices that are communicationally coupled to one another through network communications. The computing devices execute one or more of a client process, a storage process and a controller process. The controller processes manage the metadata of the virtual disk, including a virtual disk topology that defines the relationships between certain ones of the physical computer-readable storage media and a particular virtual disk. The client process provide data for storage to certain ones of the computing devices executing the storage processes, as defined by a virtual disk topology, and also read data from storage from those computing devices. The client process additionally expose the virtual disk in the same manner as any other computer-readable medium.

Virtual disks constructed from unused distributed storage

Growing community of inventors

Berkeley, CA, United States of America

Saurabh Gupta