公开(公告)号：US10387749B2

公开(公告)日：2019-08-20

申请号：US15710377

申请日：2017-09-20

Applicant: Google Inc.

Inventor： Yair Movshovitz-Attias ,  King Hong Leung ,  Saurabh Singh ,  Alexander Toshev ,  Sergey Ioffe

IPC: G06K9/62 ,  G06N20/00 ,  G06K9/46 ,  G06K9/66

Abstract: The present disclosure provides systems and methods that enable distance metric learning using proxies. A machine-learned distance model can be trained in a proxy space in which a loss function compares an embedding provided for an anchor data point of a training dataset to a positive proxy and one or more negative proxies, where each of the positive proxy and the one or more negative proxies serve as a proxy for two or more data points included in the training dataset. Thus, each proxy can approximate a number of data points, enabling faster convergence. According to another aspect, the proxies of the proxy space can themselves be learned parameters, such that the proxies and the model are trained jointly. Thus, the present disclosure enables faster convergence (e.g., reduced training time). The present disclosure provides example experiments which demonstrate a new state of the art on several popular training datasets.

公开(公告)号：US10387383B2

公开(公告)日：2019-08-20

申请号：US15433815

申请日：2017-02-15

Applicant: Google Inc.

Inventor： Joshua Danziger ,  Alissa Feldman ,  Lai Kwan Wong

IPC: G06F8/61 ,  G06F8/65 ,  G06F16/10 ,  H04L29/08 ,  G06F16/172 ,  G06F16/182 ,  G06F17/30

Abstract: Systems and methods disclosed herein provide access to a data file stored at a remote data storage system. A user device can send a file access request, via a user interface provided by the remote storage server, to the application server to access a data file stored at the remote storage server. The application server can then optionally identify a relevant application for accessing the data file, and send, to the remote server, a metadata-and-content request relating to the data file. The metadata-and-content request includes version information of a previously cached version of the data file cached at the application server. When the cached version is not the latest version, the application server receives the latest version of the data file from the remote storage server. The latest version of the data file is then presented, via the relevant application on the application server, to the user device.

公开(公告)号：US10380640B2

公开(公告)日：2019-08-13

申请号：US14160664

申请日：2014-01-22

Applicant: Google Inc.

Inventor： Shumeet Baluja ,  Vibhu Mittal ,  Mehran Sahami

IPC: G06Q30/00 ,  G06Q30/02 ,  G06F17/21

Abstract: Ad rendering parameters for a set of two or more ads may be determined by (a) accepting, for a set of two or more ads, ad information including at least one ad feature having a value that depends on ad rendering parameters, and (b) determining ad rendering parameters for at least one ad from the set of two or more ads using the accepted ad information. Ad rendering parameters may be determined using accepted ad rendering constraints. The ad rendering constraints may include space available and/or a maximum number of ads permitted to be rendered. The act of determining ad rendering parameters may include maximizing a value associated with serving at least one ad from the set of two or more ads with ad rendering parameters subject to the ad rendering constraints. The ad rendering parameters may include sizes of the served ads, and/or a layout of the served ads.

公开(公告)号：US10372715B1

公开(公告)日：2019-08-06

申请号：US13856200

申请日：2013-04-03

Applicant: Google Inc.

Inventor： Gavin James ,  Justin Lewis

IPC: G06F16/00 ,  G06F16/2457

Abstract: A method for multiple entity merge in a content sharing platform is disclosed. The method includes identifying two or more entities associated with an object of a social network. The method further includes, for each of the identified two or more entities, determining an affinity score between the entity and a user to which the object is presented on the social network. The method also includes applying a function to the determined affinity scores to generate a combined affinity score for the object.

公开(公告)号：US10372317B1

公开(公告)日：2019-08-06

申请号：US14737931

申请日：2015-06-12

Applicant: Google Inc.

Inventor： Baron Winfield Arnold

IPC: G06F3/0484 ,  G06F3/0485 ,  G06F3/0488

Abstract: A method for presenting a media item of a set of media items in a user interface (UI) of a client device is disclosed. The UI includes a first scrub area associated with a first scrub rate and a second scrub area associated with a second scrub rate. The client device receives a first user input via the first scrub area of the UI to navigate through the set of media items at the first scrub rate. The client device receives a second user input that is separate from the first user input via the second scrub area of the UI to navigate through the set of media items at a second scrub rate.

公开(公告)号：US20190238944A1

公开(公告)日：2019-08-01

申请号：US13972425

申请日：2013-08-21

Applicant: Google Inc.

Inventor： Prabhu Balasubramanian ,  John Thomas ,  Akash Hasmukh Parikh ,  Natalia Ann Cohen Fisher

IPC: H04N21/4627 ,  H04N21/61

CPC classification number: H04N21/4627 ,  H04N21/6175

Abstract: An internet-based content platform that includes subscribable channels is provided. The internet-based content platform provides channels to end users. The internet-based content platform allows a channel curator to identify the paid channel comprising a plurality of media items. The internet-based content platform allows the channel curator indicate a list of types of access to the plurality of media items, a list of subscription levels, where each subscription level is associated with a different subscription fee and where each subscription level is associated with a type of access from the list of types of access. The internet-based content platform associates the list of types of access and the list of subscription levels with the paid channel.

公开(公告)号：US10366354B2

公开(公告)日：2019-07-30

申请号：US14968878

申请日：2015-12-14

Applicant: GOOGLE INC.

Inventor： Margaret Aycinena Lippow ,  Amir Jonatan Padovitz ,  Akshay Narendra Java ,  Bala Venkata Sai Ravi Krishna Kolluri

IPC: G06Q10/02 ,  G06Q30/06 ,  G06Q50/14

Abstract: A data processing system receives location data points from computing devices. The system annotates the location data points with entities and determines a duration each of the computing devices was at corresponding entities. The system aggregates the location data points into a set of sequences based on the duration and the entities and stores the set of sequences in a data record. The system accesses the database record including a set of sequences generated from location data points received from computing devices. The system receives, from a computing device, a request for a location sequence that includes a query. The system identifies an attribute of the computing device. The system identifies a sequence based on the set of sequences using the query and the attribute. The system transmits the sequence for display on a display device.

公开(公告)号：US10366286B2

公开(公告)日：2019-07-30

申请号：US15377008

申请日：2016-12-13

Applicant: Google Inc.

Inventor： Seth Glickman ,  Emil John Feig

IPC: H04N7/18 ,  H04N5/232 ,  G06K9/00 ,  G06T7/90 ,  G08G1/0962

Abstract: Systems and methods of detecting traffic light signal changes are disclosed. For instance, it can be determined that a user is stopped at an intersection having a traffic light. A plurality of images can be captured in response to detecting that the user is stopped at the intersection having a traffic light. The plurality of images do not depict the traffic light. A tonal shift in one or more color values associated with at least one image of the plurality of images can be detected. The tonal shift is indicative of a change in signal provided by the traffic light. A notification indicative of the change in signal provided by the traffic light can be provided to the user.

公开(公告)号：US10361868B1

公开(公告)日：2019-07-23

申请号：US15161775

申请日：2016-05-23

Applicant: Google Inc.

Inventor： Brandon S. Baker ,  Uday Savagaonkar

IPC: H04L9/32 ,  H04L9/06 ,  H04L9/30 ,  G06F11/36 ,  G06F21/62

Abstract: A method includes receiving a break-glass ticket scope identifying one or more secure containers of a secure container system. The secure containers are instantiated in a non-debuggable state and execute corresponding secure execution environments for contents of the corresponding secure containers. The method also includes generating a pending break-glass ticket having the break-glass ticket scope and transmitting the pending break-glass ticket to a break-glass approver for approver. In response to receiving an approved break-glass ticket from the break-glass approver, the method includes altering an access setting of the one or more secure containers defined in the break-glass ticket scope. The altered access setting allows debugging of the respective contents of the one or more secure containers executing the corresponding secure execution environments.

公开(公告)号：US10360537B1

公开(公告)日：2019-07-23

申请号：US15484933

申请日：2017-04-11

Applicant: Google Inc.

Inventor： Mike Bendersky ,  Maureen Heymans ,  Jinan Lou ,  Jie Yang ,  MyLinh Yang ,  Amitabh Saikia ,  Marc-Allen Cartright ,  Vanja Josifovski ,  Hui Tan ,  Luis Garcia Pueyo

IPC: G06F17/30 ,  G06Q10/10 ,  G06F16/248 ,  G06F16/9535 ,  H04W4/029

Abstract: Techniques are described herein for generating and applying event data extraction templates. In various implementations, a data extraction template may be applied to structured communications to extract, from each structured communication, event data associated with a transient markup language path indicated in the data extraction template. The data extraction template may include an event-related semantic data type assigned to the transient markup language path and a strength of association between the transient structural path and the event-related semantic data type. Feedback may be obtained concerning event data extracted from one or more of the structured communications. Based on the feedback, the strength of association between the transient markup language path and the event-related semantic data type may be altered. The data extraction template may then be applied to a subsequent structured communication to extract new event data from the structured communication based on the altered strength of association.