Apple Inc.

Vobile, Inc.

Topsy Labs, Inc.

Lun Cui

Rishab Aiyer Ghosh

Vipul Ved Prakash

Vivek Sehgal

Thomas James Emerson

John Martin Hörnkvist

Hon Yuk Chan

Julien F Freudiger

Robert A Walker

Trungtin Tran

Jared A Crawford

Nikrouz Ghotbi

Anubhav Malhotra

Jian Lu

Marek Rastislav Telgarsky

Kirill Novitchenko

Stanley N Hung

Kyle W Lu

Briana C Chapman

Matias Hernan Sulzberger

Kevin Siu Wai Li

Steven J Falkenburg

Christopher S Linn

Alexandre F Carlhian

John M Hornkvist

Keith Stattenfield

Thomas Deniau

Sumit Wattal

Nicolas Zeitlin

Saravana Kumar Siva Kumaran

David Salim

Gennaro Frazzingaro

Karolina Buchner

Pulah J Shah

Raffael Hannemann

Ian M Priestley

Luca Lupo

Nilesh Karia

Harsh Fatepuria

Charles Henry Salzenberg, Jr

João Pedro De Almeida Forjaz De Lacerda

Raphaeal Mor

Robert A Walker, Ii

A device implementing a system for determining an answer to a question includes a processor configured to receive a question from a client device, and send the question to a first information domain comprising plural search subdomains, each of the plural search subdomains being configured to provide a respective candidate answer to the question in a format that differs from those provided by the other search subdomains of the plural search subdomains. The processor is further configured to receive, in response to sending the question, the candidate answer from each of the plural search subdomains, and provide at least one of the candidate answers to the client device, for providing an answer to the question based on the at least one of the candidate answers.

Accessing multiple domains across multiple devices for candidate responses

Systems and processes for providing personalized suggestions indicating that a task may be performed using a digital assistant of an electronic device are provided. An example method includes, at an electronic device with a display, receiving context data associated with the electronic device; determining, based on the context data, a task that may be performed by a digital assistant of the electronic device in response to a natural-language expression; determining, based on the context data, whether suggestion criteria associated with the determined task are satisfied; and in accordance with a determination that the suggestion criteria are satisfied, providing a suggestion indicating that the determined task may be performed using the digital assistant of the electronic device.

Voice assistant discoverability through on-device targeting and personalization

A device implementing a system for anonymizing user data provided for server-side operations includes a processor configured to receive user input including a search term, wherein first and second data structures are stored on the device, the first data structure including user interest data items that correspond to prior user activity, the second data structure including topic data items mapped to the user interest data items, the topic data items being broader than the user interest data items. The processor accesses the second data structure to obtain one or more topic data items. The processor transmits, to a server, the search term and one or more topic data items for obtaining a query suggestion or search result. The processor receives, from the server, the query suggestion or search result, the query suggestion or search result having been obtained based on the search term and one or more topic data items.

Anonymizing user data provided for server-side operations

A method and apparatus of a device that re-rank a plurality of search results received from a search server on the device is described. In an exemplary embodiment, the device receives, with a browser of the device, a search query from a user. The device further sends the search query to a search server. In addition, the device receives the plurality of search results from the search server, wherein the plurality of search results is ranked according to a first ranking. The device also retrieves browser history of the user, wherein the browser history is collected on the device without revealing the browser history to the search server. The device further re-ranks the plurality of search results according to a second ranking based on at least the browser history. In addition, the device presents the plurality of search results using the second ranking.

Client-side search result re-ranking

A method and apparatus of a device that generates a re-ranking model used to re-rank a plurality of search results on a client device is described. In an exemplary embodiment, the device receives a crowd-sourced intra-domain model from a server, where the intra-domain model is a search result re-ranking model generated based on at least device interactions of a plurality of users interacting with a plurality of other devices. The device further generates a re-ranking model from the crowd-sourced intra-domain model and a local model, where the local model includes private data representing a device user's interaction with that device and the re-ranking model is used to re-rank a plurality of search results.

Blending learning models for search support

A method and apparatus of a device that performs a multi-domain query search is described. In an exemplary embodiment, the device receives a query prefix from a client of a user. The device further determines a plurality of search completions across the plurality of separate search domains. In addition, the device ranks the plurality of search completions based on a score calculated for each of the plurality of search completions determined by a corresponding search domain, where at least one of the plurality of search completions is used to generate a plurality of search results without an indication from the user and in response to receiving the query prefix.

Multi-domain query completion

A new approach is proposed that contemplates systems and methods to examine and determine quality of objects cited by citations in a search result based on a citation graph that includes citing subjects, citations, and cited objects. First, influence scores of a plurality of subjects/sources that compose the citations of the objects in the search result are calculated. The quality of the objects cited by the subjects can then be determined by examining the influence scores for the subjects of the citations. Finally, the cited objects selected can be presented to a user or provided to a thirty party for further processing together with the relevant citations and citing subjects.

System and method for determining quality of cited objects in search results based on the influence of citing subjects

A method and apparatus of a device that re-rank a plurality of search results is described. In an exemplary embodiment, the device receives a search query from a user and generates the plurality of search results over a plurality of search domains, wherein the plurality of search results is ranked according to a first ranking. The device additionally generates a re-ranking model, where the re-ranking model includes a plurality of intra-domain models that are generated based on at least based on-device interactions of a plurality of users interacting with a plurality of other devices and each of the plurality of search domains corresponds to one of the plurality of intra-domain models. The device further re-ranks the plurality of search results using the re-ranking model and presents the plurality of search results using the second ranking.

Re-ranking search results using blended learning models

A device implementing a system for providing query completion suggestions includes at least one processor configured to receive a partial query, and derive one or more query completion suggestions for the partial query from content items that are locally stored on the device, each respective query completion suggestion being associated with metadata extracted from one of the content items from which the respective query completion suggestion was derived. The at least one processor is further configured to rank the one or more query completion suggestions based at least in part on the metadata associated with each of the one or more query completion suggestions. The at least one processor is further configured to provide, for display, at least one of the one or more ranked query completion suggestions.

Query completion suggestions

A method and apparatus of a device that selects an application state for use in a multi-device search is described. In this embodiment, the device detects, on the device, that the application state has been selected as a query result for a device-level search on that device. The device further transmits the application state to a server, wherein the application state is to be indexed with other application states from other devices.

Crowdsourcing application history search

A new approach is proposed that contemplates systems and methods to determine temporality of a query in order to generate a search result including a list of objects that are not only based on matching of the objects to the query but also based on temporality analysis of the query. Here, the temporality of the query can be defined as the distribution over time of the objects matching the query, i.e., the chronology histogram of the query. Such distribution can be analyzed to provide a classification of the intent of the query. Classification of the intent of the query can result either in discrete classification of the query into categories, or in continuous classification of the query which may be a scalar or vector value resulting from transformations of the chronology histogram.

System and method for query temporality analysis

A new approach is proposed that contemplates systems and methods to provide a ranking of cited objects and citing subjects identified as results of a search, where the relative expertise of subjects or sources of citations to said targets or objects is considered. The relative expertise is a function of the share of the subject's citations matching the query term or search criteria relative to the share of all subjects' citations matching the query term, weighted by the influence of the subjects. This allows the identification of 'experts' on 'topics' without any pre-defined categorization of topics or pre-computation of expertise. Under this novel approach, expertise can be determined on any query term in real-time.

Search of sources and targets based on relative expertise of the sources

A new approach is proposed that contemplates systems and methods to generate customized subjective search results from the perspective of a user who conducts the search or any other subject entity of chosen by the user. A scored subject list is created from the user's network of sources/subjects/contacts, where each element on the list is a subject/source and the score reflects the subject's potential influence or closeness of its connection/relation with the user. Once created, the subject list is then used as a bias filter on the list of citations from search results. With such influence-weighted citation scores, objects and/or subjects from citations of subjects that have big influence on or enjoy high respect from the user will be ranked prominently in the search result presented to the user, thus biasing the search results from the user's perspective.

System and method for customizing search results from user's perspective

Growing community of inventors

Fremont, CA, United States of America

Lun Cui