Abstract: A method of populating an image repository is disclosed. The method includes identifying a keyword from a dictionary and querying an image source using the keyword, thereby yielding a set of images associated with the keyword. The method also includes performing a second query on the image source using a combination of the keyword and one or more additional terms, thereby yielding a plurality of images associated with the combination. Then the plurality of images associated with the combination is subtracted from the set of images associated with the keyword, thereby yielding a difference set of images which are added to the image repository if the difference set of images satisfies predefined criteria.

Abstract: A computer-implemented method of providing personalized route information involves gathering a plurality of past location indicators over time for a wireless client device, determining a future driving objective using the plurality of previously-gathered location indicators, obtaining real-time traffic data for an area proximate to the determined driving objective, and generating a suggested route for the driving objective using the near real-time traffic data.

Abstract: The subject matter of this specification can be embodied in, among other things, a method that includes generating content-based keywords based on content generated by users of a social network. The method includes labeling nodes comprising user nodes, which are representations of the users, with advertising labels comprising content-based keywords that coincide with advertiser-selected keywords that are based on one or more terms specified by an advertiser. The method also includes outputting, for each node, weights for the advertising labels based on weights of advertising labels associated with neighboring nodes, which are related to the node by a relationship.

Abstract: Systems and methods for allocating space for advertisements in a network of electronic display devices are provided. Attribute information indicating retailer and categories of products available for purchase in the vicinity of a display device is maintained for each device in a database. Advertisers may upload advertisement messages to a server specifying information such as budget, price per impression, preferred billboards and/or other constraints. One or more keywords or other descriptors are specified for each advertisement message. The system then generates an advertising campaign specifying where the advertisement message is to be output and send the messages to the specified displays. The output may consist of various forms including video, audio, printed incentive, interactive data transfers and/or combinations of these.

Abstract: A human interaction proof may restrict access to computer systems by testing the image orientation detection capabilities of the user making the access attempt. The human interaction proof includes a challenge where a plurality of images is presented. The images in the challenge may be drawn from image databases or the Internet, and may be part of a set of images that have been screened for difficulty with respect to computers. Some of the presented images are rotated away from their proper orientations. A user presented with the challenge is asked to differentiate the rotated images from the non-rotated images. A correct response leads to a grant of access, and an incorrect answer leads to a denial of access. The images in the challenge may be further modified with regard to color, lighting, cropping, etc. before presentation; the modifications increase the difficulty of the challenge for computers.

Abstract: The subject matter of this specification can be embodied in, among other things, a method that includes inferring labels for videos, users, advertisements, groups of users, and other entities included in a social network system. The inferred labels can be used to generate recommendations such as videos or advertisements in which a user may be interested. Inferred labels can be generated based on social or other relationships derived from, for example, profiles or activities of social network users. Inferred labels can be advantageous when explicit information about these entities is not available. For example, a particular user may not have clicked on any online advertisements, so the user is not explicitly linked to any advertisements.

Abstract: A method including training a plurality of learning systems, each learning system implementing a learning function and having an input and producing an output, initializing one or more data structures, and evaluating a target sample is described. Also described are methods that include initializing one or more data structures and evaluating a target sample for a best match.

Abstract: A computer-implemented method of displaying information on a mobile device is discussed. The method includes displaying on the mobile device a first view having a first search result element in an expanded format and a plurality of additional search result elements in a collapsed format, receiving a user input that identifies a selected search result element from the plurality of additional search result elements, and in response to the user input, displaying on the mobile device a second view having one of the plurality of additional search result elements in an expanded format, and the remainder of the plurality of additional search result elements in a collapsed format.

Abstract: A system and method for recommending video programs to a user comprising determining a first video program that is of interest to a user and then determining a second video program to recommend to the user, the second video program being determined from a recommendation database assembled by analyzing access logs from one or more search engines or online video content providers.

Abstract: A method for generating a speech recognition model includes accessing a baseline speech recognition model, obtaining information related to recent language usage from search queries, and modifying the speech recognition model to revise probabilities of a portion of a sound occurrence based on the information. The portion of a sound may include a word. Also, a method for generating a speech recognition model, includes receiving at a search engine from a remote device an audio recording and a transcript that substantially represents at least a portion of the audio recording, synchronizing the transcript with the audio recording, extracting one or more letters from the transcript and extracting the associated pronunciation of the one or more letters from the audio recording, and generating a dictionary entry in a pronunciation dictionary.

Abstract: Additional documents are automatically located that are relevant to an original document, such as a document being read by a user, and also potentially relevant to personal information of the user. The additional documents may be located based on descriptive information that includes personal information of the user and content information of the document being read. The additional documents, or links to the additional documents, may be incorporated into the document being read. In some implementations, the additional documents may be presented in-line with the document being read, such as through an in-link link or text snippet. The user can thus be efficiently presented with additional information that is relevant to the original document being read.

Abstract: Systems and methods for socially-based correction of tilted images. In an embodiment, the present invention levels tilted images by (1) receiving an image; (2) rotating the image by a random angle; (3) requesting at least one user rotate the image to level the image; (4) determining a collective user-suggested angle based on the rotations of each user; and (5) transmitting, over a communication network, a correction angle, where the correction angle is based on the random angle and the collective user-suggested angle.

Abstract: Additional documents are automatically located that are relevant to an original document, such as a document being read by a user, and also potentially relevant to personal information of the user. The additional documents may be located based on descriptive information that includes personal information of the user and content information of the document being read. The additional documents, or links to the additional documents, may be incorporated into the document being read. In some implementations, the additional documents may be presented in-line with the document being read, such as through an in-link link or text snippet. The user can thus be efficiently presented with additional information that is relevant to the original document being read.

Abstract: A system analyzes multiple images to identify similar images using histograms, image intensities, edge detectors, or wavelets. The system retrieves labels assigned to the identified similar images and selectively concatenates the extracted labels. The system assigns the concatenated labels to each of the identified similar images and uses the concatenated labels when performing a keyword search of the plurality of images.

Abstract: A first indication of a portion of an image presented on a display device associated with a first user is received in response to a prompt to identify an object. A second indication of a portion of the image presented on a display device associated with a second user is received in response to a prompt to identify the object. A region-of-interest in the image is identified based on the first indication and the second indication. The region-of-interest is associated with an identifier of the object. A designator is associated with the region-of-interest in the image, the designator being configured to present information related to the object. Presentation of the designator associated with the region-of-interest in the image is enabled in subsequent presentations of the image.

Abstract: Messages of a user are ranked based on metrics derived from user actions with regard to the messages. Groupings of messages, such as conversations or a set of messages that are determined to be similar, may also be ranked. When new messages are received, their interest to the user is estimated based on rankings of messages and message groupings to which the new message is related. The new messages may be made more or less prominent based on the estimated interest to the user.

Abstract: A request to select sponsored content to be presented with a first content item including video is received. A second content item including video is identified. The video included in the second content item has been viewed by at least one viewer of the video included in the first content item. Sponsored content that has been presented with the second content item is identified and selected to be presented with the first content item is selected.

Abstract: A system and method detects matches between portions of video content. A matching module receives an input video fingerprint representing an input video and a set of reference fingerprints representing reference videos in a reference database. The matching module compares the reference fingerprints and input fingerprints to generate a list of candidate segments from the reference video set. Each candidate segment comprises a time-localized portion of a reference video that potentially matches the input video. A classifier is applied to each of the candidate segments to classify the segment as a matching segment or a non-matching segment. A result is then outputted identifying a matching portion of a reference video from the reference video set based on the segments classified as matches.

Abstract: A system analyzes multiple images to identify similar images using histograms, image intensities, edge detectors, or wavelets. The system retrieves labels assigned to the identified similar images and selectively concatenates the extracted labels. The system assigns the concatenated labels to each of the identified similar images and uses the concatenated labels when performing a keyword search of the plurality of images.

Abstract: The subject matter of this specification can be embodied in, among other things, a method that includes determining, for a portion of users of a social network, label values each comprising an inferred interest level of a user in a subject indicated by a label, associating a first user with one or more second users based on one or more relationships specified by the first user, and outputting a first label value for the first user based on one or more second label values of the one or more second users.