Abstract: Methods and apparatus related to determining a triggering event of a user, selecting media relevant to the triggering event, and providing the selected media to the user. Some implementations are directed to methods and apparatus for determining a past event of the user that is indicative of past interaction of the user with one or more past entities and the triggering event may be determined to be associated with the past event. The media selected to provide to the user may contain media that includes the one or more past entities associated with the past event and the media may be provided to the user in response to the triggering event.

Abstract: Methods and apparatus related to determining a triggering event of a user, selecting media relevant to the triggering event, and providing the selected media to the user. Some implementations are directed to methods and apparatus for determining a past event of the user that is indicative of past interaction of the user with one or more past entities and the triggering event may be determined to be associated with the past event. The media selected to provide to the user may contain media that includes the one or more past entities associated with the past event and the media may be provided to the user in response to the triggering event.

Abstract: The technology relates to allocating the resource costs of a sensor network. In particular, an auction model is used to control the demand-side of sensor network utilization. Users bid for fractional sensor network utilization over particular variables such as time, geography, and data type. During the bidding process, the available sensor network resources may be unknown. Therefore, users bid on the predicted fractional sensor network utilization, which can be estimated using historical sensor network trends and models. Throughout this process, device performance is preserved while providing for user experience, privacy, and security.

Abstract: Methods for identifying and obtaining a suitable application for interacting with a selected content item by providing secure access to a limited portion of information on the selected content item are provided. In one aspect, a method includes receiving a selection from a user in an initial application running on a device to interact with a content item accessible by the initial application, and generating a hash value from the limited portion of information on the content item. The method further includes requesting a copy of the at least one suitable application when the comparison of the hash value for the content item with the hash value for the suitable application match, executing the suitable application on the device, and providing the limited portion of information on the content item to the suitable application for interaction by the user. Systems and machine-readable media are also provided.

Abstract: The technology relates to allocating the resource costs of a sensor network. In particular, an auction model is used to control the demand-side of sensor network utilization. Users bid for fractional sensor network utilization over particular variables such as time, geography, and data type. During the bidding process, the available sensor network resources may be unknown. Therefore, users bid on the predicted fractional sensor network utilization, which can be estimated using historical sensor network trends and models. Throughout this process, device performance is preserved while providing for user experience, privacy, and security.

Abstract: In one aspect, an automated endorsement prompt systems includes an endorsement prompt module comprising an endorsement signal module for retrieving an endorsement signal from an endorsement server; a search result module for retrieving search results from a search engine; a web history module for retrieving a web history for a user; and combiner logic for providing search results and an endorsement prompt. The combiner logic generates the endorsement prompt from the endorsement signal and the web history. A method for automatically generating endorsement prompts including the steps of: receiving a query from a user; obtaining additional information signals; obtaining a search result using the query; determining whether prompt behavior exists using the additional information; generating a prompt for an endorsement if the prompt behavior exists; and providing the search result and the prompt for presentation.

Abstract: A method and system for using recent search queries for inserting relevant search results for navigational queries. At a server, multiple first queries including one or more search terms are received from a first user of a client. At least a subset of the first queries is stored in memory. A navigational query is received from the first user. The navigational query includes a first identifier of a first web site. An alternate query including one or more of the search terms is performed at the server. The search results for the alternate query are transmitted to the client for display to the first user.

Abstract: The technology relates to allocating the resource costs of a sensor network. In particular, an auction model is used to control the demand-side of sensor network utilization. Users bid for fractional sensor network utilization over particular variables such as time, geography, and data type. During the bidding process, the available sensor network resources may be unknown. Therefore, users bid on the predicted fractional sensor network utilization, which can be estimated using historical sensor network trends and models. Throughout this process, device performance is preserved while providing for user experience, privacy, and security.

Abstract: Methods for identifying and obtaining a suitable application for interacting with a selected content item by providing secure access to a limited portion of information on the selected content item are provided. In one aspect, a method includes receiving a selection from a user in an initial application running on a device to interact with a content item accessible by the initial application, and generating a hash value from the limited portion of information on the content item. The method further includes requesting a copy of the at least one suitable application when the comparison of the hash value for the content item with the hash value for the suitable application match, executing the suitable application on the device, and providing the limited portion of information on the content item to the suitable application for interaction by the user. Systems and machine-readable media are also provided.

Abstract: The technology relates to allocating the resource costs of a sensor network. In particular, an auction model is used to control the demand-side of sensor network utilization. Users bid for fractional sensor network utilization over particular variables such as time, geography, and data type. During the bidding process, the available sensor network resources may be unknown. Therefore, users bid on the predicted fractional sensor network utilization, which can be estimated using historical sensor network trends and models. Throughout this process, device performance is preserved while providing for user experience, privacy, and security.

Abstract: Methods and apparatus related to determining a triggering event of a user, selecting media relevant to the triggering event, and providing the selected media to the user. Some implementations are directed to methods and apparatus for determining a past event of the user that is indicative of past interaction of the user with one or more past entities and the triggering event may be determined to be associated with the past event. The media selected to provide to the user may contain media that includes the one or more past entities associated with the past event and the media may be provided to the user in response to the triggering event.

Abstract: One or more devices may receive multiple images; determine one or more classifications for each of the multiple images; generate a confidence score for each classification of the one or more classifications for each of the multiple images. The confidence score may include a value between two numbers and may identify the likelihood that that a particular classification, of the one or more classifications, correctly corresponds to a particular image, of the multiple images. The confidence score may be based on information, associated with a user, and pixel-based classification information. The one or more devices may sort the multiple images by the one or more classifications and by corresponding confidence scores to form sorted multiple images; and provide the sorted multiple images.

Abstract: Methods and apparatus related to determining a triggering event of a user, selecting media relevant to the triggering event, and providing the selected media to the user. Some implementations are directed to methods and apparatus for determining a past event of the user that is indicative of past interaction of the user with one or more past entities and the triggering event may be determined to be associated with the past event. The media selected to provide to the user may contain media that includes the one or more past entities associated with the past event and the media may be provided to the user in response to the triggering event.

Abstract: Implementations generally relate to improving metadata. In some implementations, a method includes determining timestamps, where the timestamps correspond to images. The method also includes determining one or more reference times based on predetermined metadata. The method also includes detecting a timestamp error in one or more timestamps based on the one or more reference times. The method also includes adjusting each timestamp having a timestamp error based on the one or more reference times.

Abstract: Methods and apparatus related to determining a triggering event of a user, selecting media relevant to the triggering event, and providing the selected media to the user. Some implementations are directed to methods and apparatus for determining a past event of the user that is indicative of past interaction of the user with one or more past entities and the triggering event may be determined to be associated with the past event. The media selected to provide to the user may contain media that includes the one or more past entities associated with the past event and the media may be provided to the user in response to the triggering event.

Abstract: One or more devices may receive multiple images; determine one or more classifications for each of the multiple images; generate a confidence score for each classification of the one or more classifications for each of the multiple images. The confidence score may include a value between two numbers and may identify the likelihood that that a particular classification, of the one or more classifications, correctly corresponds to a particular image, of the multiple images. The confidence score may be based on information, associated with a user, and pixel-based classification information. The one or more devices may sort the multiple images by the one or more classifications and by corresponding confidence scores to form sorted multiple images; and provide the sorted multiple images.

Abstract: One or more computer devices may receive a document from private content that is accessible by a set of users, and may extract metadata from the document. The metadata may provide information relating to one or more aspects of the document. The one or more computer devices may determine, based on the metadata, whether a portion of the document includes information that is only to be accessible by a particular user from the set of users, and may append a restrict indicator to the metadata when the portion of the document includes information that is only to be accessible by the particular user. The restrict indicator may identify the particular user and the portion of the document. The one or more computer devices may store the metadata and the restrict indicator in a personal search index.

Abstract: A method uses a computing device that includes a first processor in a first, inactive state operatively coupled to a second processor in an active state. While the first processor is in the first state (301) the second processor uses sensors to determine (303) that a first condition occurred. In response to determining that the first condition has occurred, the first processor initializes (305) to a second, active state. If the first or second processor determines (311) that a second condition has occurred within a given time period (307 and 309) after the first condition occurred, the first processor performs (315) a first action such as launching a software application, capturing a digital photograph, or placing a phone call.