Palo Alto Research Center Inc.

Google Inc.

Xerox Corporation

Conduent Business Services, LLC

Intel Corporation

Kurt Edward Partridge

James M A Begole

Robert Roy Price

Yu Ouyang

Maurice Kyojin Chu

Shumin Zhai

Daniel H Greene

Nicolas B Ducheneaut

Xiaojun Bi

Mark Jeffrey Stefik

Francisco Eduardo Torres

Alan G Bell

Craig Eldershaw

Serdar Uckun

Lance Everett Good

David Preston Cummins

Rui Zhang

Mark Webster Newman

Dragan Zivkovic

David Goldberg

Francoise Beaufays

Diana Kathryn Smetters

Françoise Beaufays

Rajiv Mathews

Andrew Hard

Robert Thomas Krivacic

Ed H Chi

Ersin Uzun

Richard Chow

James E Reich

Koji Yatani

Michael Roberts

Leigh L Klotz, Jr

Ji Fang

Victoria M E Bellotti

Mary Catherine McCorkindale

Tracy Holloway King

Manas Ashok Pathak

Cong Wang

Nicholas B Ducheneaut

Kai Kunze

Mark W Newman

Processor(s) of a client device can: receive sensor data that captures environmental attributes of an environment of the client device; process the sensor data using a machine learning model to generate a predicted output that dictates whether one or more currently dormant automated assistant functions are activated; making a decision as to whether to trigger the one or more currently dormant automated assistant functions; subsequent to making the decision, determining that the decision was incorrect; and in response to determining that the determination was incorrect, generating a gradient based on comparing the predicted output to ground truth output. In some implementations, the generated gradient is used, by processor(s) of the client device, to update weights of the on-device speech recognition model. In some implementations, the generated gradient is additionally or alternatively transmitted to a remote system for use in remote updating of global weights of a global speech recognition model.

Using corrections, of automated assistant functions, for training of on-device machine learning models

A system and method for providing distributed on-street valet parking with the aid of a digital computer is provided. Parking spaces for valet services are reserved. One of the reserved parking spaces is designated as a valet drop-off parking space. A motor vehicle is detected in the valet drop-off parking space and a notification that the motor vehicle requires parking is provided to a valet of the detected motor vehicle. A search is conducted for the reserved parking spaces available for valet services for the motor vehicle. At least one of the reserved parking spaces available for the motor vehicle is identified and the motor vehicle is parked in the identified reserved parking space.

System and method for providing distributed on-street valet parking with the aid of a digital computer

One embodiment provides a system for generating an inference model that determines an activity type for a user from contextual information. During operation, the system receives a set of contextual information associated with the user, wherein the contextual information includes at least a set of location coordinates. The system then determines an association between the contextual information and an activity type. Next, the system generates an activity inference model based in part on the association, wherein the activity inference model takes an instance of contextual information as an input parameter and outputs a corresponding activity type. The model's parameters are based at least on statistics associated with the user's contextual history but not based on the complete contents of the user's contextual history.

Generating an activity inference model from contextual data

A computer-implemented system and method for providing multi-locational curbside valet parking service is provided. A multiplicity of motor vehicle parking spaces are managed by a server. A plurality of smart parking devices and vehicle occupancy sensors are interfaced with the server. Use of a plurality of the parking spaces is reserved for a valet service. A reservation request for the use of the plurality of the parking spaces is accepted for use by the valet service for a period of time. Occupancy of the reserved parking spaces by a motor vehicle is sensed through the vehicle occupancy sensor. A valet drop-off location of at least one of the parking spaces that is reserved for the use of the valet service and is not occupied by a motor vehicle is identified. The valet drop-off location and a location of a lock box are provided to the motorist through a user interface.

One embodiment of the present invention provides a system for facilitating presentation of activity-based advertising. During operation, the system receives a set of contextual data of a customer and makes a prediction of the customer's future activity. The system then receives a number of advertisements from the advertisers. Based on the prediction, the system chooses a received advertisement to present to the customer. The system further determines the customer's subsequent activity and confirms the prediction of the customer's activity. The system then receives payments from the advertisers whose advertisement is presented based on whether the prediction is confirmed.

Advertising payment based on confirmed activity prediction

In one example, a method includes outputting, for display at a presence-sensitive display, a graphical user interface including a text display region and a graphical keyboard. The method further includes receiving an indication of a first gesture detected at a location within the text display region, and selecting, based at least in part on the location within the text display region, at least a portion of a character string included in the text display region. The method further includes receiving, while the first gesture is detected at the presence-sensitive display, an indication of at least a portion of a second gesture to select at least one key of the graphical keyboard, and outputting, responsive to receiving an indication of a termination of at least the second gesture, a replacement character string for display within the text display region to replace the portion of the character string.

Character string replacement

Techniques are described for character deletion on a computing device that utilizes a gesture-based keyboard. The computing device includes a processor and at least one module operable by the processor to output, for display at a presence-sensitive display, a graphical keyboard comprising a plurality of keys and a text editor field. In response to receiving an indication of a gesture that comprises a first path, the module outputs a first character, based at least in part on the first path, associated with a first key and a second character associated with a second key. In response to determining that the gesture further comprises a second path that retraces at least a portion of the first path, the module removes the second character from the text editor field of the presence-sensitive display.

Character deletion during keyboard gesture

In one example, a method includes outputting, for display at a presence-sensitive display, a graphical user interface comprising a graphical keyboard that includes a group of keys, wherein each key in the group of keys is associated with a respective, different region of the presence-sensitive display. The method further includes receiving an indication of a gesture to select a sequence of one or more keys in the group of keys of the graphical keyboard. The method further includes determining, based at least in part on the indication of the gesture, that the gesture includes at least one feature associated with a cancellation of the selected sequence of one or more keys. The method further includes, in response to determining that the gesture includes the at least one feature associated with the cancellation, omitting the selected sequence of one or more keys from entry to a text input field.

Gesture keyboard with gesture cancellation

Aspects of the present disclosure are directed to techniques for outputting a graphical keyboard comprising a group of keys, wherein each key in the group of keys is associated with a respective, different display region, receiving an indication of a gesture to select a sequence of keys that are each included in the group of keys of the graphical keyboard, determining that the selected sequence of keys corresponds to a character string that is identifiable by at least one format source, wherein the format source is associated with a syntax, determining, based at least in part on the syntax, that at least one separator character is associated with the character string, and in response to determining that the at least one separator character is associated with the character string, outputting the character string and the at least one separator character at a location proximal to the character string.

Contextually-specific automatic separators

Computer-implemented system and method for providing multi-locational curbside valet parking services

A recommender system uses an activity decision tree to model the changes in a user's behavior according to a plurality of time-related features. The system determines historical activities for the user, and generates a decision tree for the user's historical activities. Each leaf node of the decision tree is associated with an activity-prediction model that computes a probability for a corresponding activity. The system selects a path of the decision tree from a root node to a leaf node of the decision tree based on a target time. The selected path traverses two or more non-leaf nodes that are each associated with a temporal decision model that compares the target time against a temporal classifier. The system then determines a probability for a user activity based on an activity-prediction model of the selected path.

Performance-efficient system for predicting user activities based on time-related features

A system facilitates automatically determining the hypothetical context information or the distribution of hypothetical contexts. During operation, the system receives a request from a user for one or more recommendations. The system also receives a current context substantially associated with the request. The system then determines a hypothetical context for the request, wherein the hypothetical context may be determined by considering several sources of information, including but not limited to the current context, past contexts, and relationships between the current context and past contexts. Next, the system determines one or more recommendations for the user based on the hypothetical context. Finally, the system returns the one or more recommendations to the user.

Method and apparatus for automatically incorporating hypothetical context information into recommendation queries

Growing community of inventors

Palo Alto, CA, United States of America

Kurt Edward Partridge