Abstract: A system and method provide a soundtrack recommendation service for recommending one or more soundtrack for a video (i.e., a probe video). A feature extractor of the recommendation service extracts a set of content features of the probe video and generates a set of semantic features represented by a signature vector of the probe video. A video search module of the recommendation service is configured to search for a number of video candidates, each of which is semantically similar to the probe video and has an associated soundtrack. A video outlier identification module of the recommendation service identifies video candidates having an atypical use of their soundtracks and ranks the video candidates based on the typicality of their soundtrack usage. A soundtrack recommendation module selects the soundtracks of the top ranked video candidates as the soundtrack recommendations to the probe video.

Abstract: A system, computer readable storage medium, and computer-implemented method presents video search results responsive to a user keyword query. The video hosting system uses a machine learning process to learn a feature-keyword model associating features of media content from a labeled training dataset with keywords descriptive of their content. The system uses the learned model to provide video search results relevant to a keyword query based on features found in the videos. Furthermore, the system determines and presents one or more thumbnail images representative of the video using the learned model.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for identifying similar images. In some implementations, a method is provided that includes receiving a collection of images and data associated with each image in the collection of images; generating a sparse feature representation for each image in the collection of images; and training an image similarity function using image triplets sampled from the collection of images and corresponding sparse feature representations.

Abstract: Methods and systems for providing rich content with an answer to a question query. A method includes receiving a query determined to be a question query and a corresponding answer generated in response to the question query, generating a contextual query that includes an element relating to the question query and an element relating to the answer; submitting the contextual query to a rich content search process and receiving data specifying a first set of rich content items responsive to the contextual query, determining first rich content item in the first set of rich content items that meet a context condition that is indicative of a rich content item providing contextual information of both elements of the question query and the answer query; and preferentially selecting from the first content items relative to the second rich content items to be provided as one or more answer rich content items.

Abstract: Systems and methods for recommending an outdoor activity geographic location are provided. In some aspects, multiple data items that are associated with a recurring time period and a destination geographic location are selected from one or more data repositories. That an outdoor activity for the user is available in the destination geographic location during the recurring time period is determined using information in the selected multiple data items and at least one characteristic of a user of a client computing device. An indication that the outdoor activity is available in the destination geographic location is provided to the client computing device.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for identifying similar images. In some implementations, a method is provided that includes receiving a collection of images and data associated with each image in the collection of images; generating a sparse feature representation for each image in the collection of images; and training an image similarity function using image triplets sampled from the collection of images and corresponding sparse feature representations.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for processing drawing-based search queries. In one aspect, a method includes receiving search request data specifying a search request, the search request comprising a drawing represented by a set of line strokes, each line stroke representing a trace of a moving input point; identifying line segments from the line strokes; comparing the identified line segments to reference line segments, each of the reference line segments representing a portion of a corresponding reference drawing; identifying a candidate reference drawing based on the comparison of the identified line segments with reference line segments; identifying a keyword for the candidate reference drawing, the keyword being a term determined to be relevant to a subject matter of the candidate reference drawing; and in response to receiving the search request, providing search results data specifying search results responsive to the keyword.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for retrieving images based on a seasonal time period. In one aspect, a method includes receiving a query including image data defining an image depicting a subject. A location at which the image was captured by an image capturing device is determined. One or more additional images of the subject that were captured by an image capturing device at the identified location are identified. At least a portion of the additional images are provided in response to receiving the query.

Abstract: In one aspect, a method for automatically adjusting temporal creation data for a plurality of images is disclosed. The method may generally include receiving a plurality of image files associated with the plurality of images, wherein each of the plurality of image files includes an initial creation time. In addition, the method may include determining a baseline creation time by analyzing the plurality of image files based on their association with at least one of daytime hours or nighttime hours and adjusting the initial creation time for at least one of the plurality of image files based on the baseline creation time.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training scoring models. One method includes storing data identifying a plurality of positive and a plurality of negative training images for a query. The method further includes selecting a first image from either the positive group of images or the negative group of images, and applying a scoring model to the first image. The method further includes selecting a plurality of candidate images from the other group of images, applying the scoring model to each of the candidate images, and then selecting a second image from the candidate images according to scores for the images. The method further includes determining that the scores for the first image and the second image fail to satisfy a criterion, updating the scoring model, and storing the updated scoring model.

Abstract: Systems and methods are provided to select a most typical pronunciation of a location name on a map from a plurality of user pronunciations. A server generates a reference speech model based on user pronunciations, compares the user pronunciations with the speech model and selects a pronunciation based on comparison. Alternatively, the server compares the distance between one the user pronunciations and every other user pronunciations and selects a pronunciation based on comparison. The server then annotates the map with the selected pronunciation and provides the audio output of the location name to a user device upon a user's request.

Abstract: This disclosure is directed to providing audio playback to a mobile device user. According to one aspect of this disclosure, a mobile device may be to modify audio playback in response to detecting an inclination of the mobile device (and thereby a user) with respect to a reference plane. According to another aspect of this disclosure, a mobile device may be configured to automatically identify an audible sound that may be motivational to a user, and store an indication of the audible sound in response to the identification. According to another aspect of this disclosure, a mobile device may automatically play back a previously identified motivational song in response to detection of user movement.

Abstract: A device is configured to receive a translation query that requests a translation of terms from a source language to a target language; determine translation features associated with the translation query; assign a feature value to each of the translation features to form feature values; apply a feature weight to each of the feature values to generate a final value; and determine whether to provide a dialog translation user interface or a non-dialog translation user interface based on whether the final value satisfies a threshold. The dialog translation user interface may facilitate translation of a conversation, the non-dialog translation user interface may provide translation search results, and the non-dialog translation user interface may be different than the dialog translation user interface. The device also configured to provide the dialog translation user interface for display when the final value satisfies the threshold.

Abstract: A network of apparatuses that characterizes items is presented. A self-updating apparatus includes a processing unit that has a memory storing parameters that are useful for characterizing different items, and a processing module configured to automatically select sources from which to receive data, modify the parameters based on the data that is received, and to select recipients of modified parameters. Selection of sources and recipients is based on comparison of parameters between the processing module and the sources, and between the processing module and the recipients, respectively. The processing unit may include an artificial intelligence program (e.g., a neural network such as a machine learning program). When used in a network, the processing units may “train” other processing units in the network such that the characterization accuracy and range of each processing unit improves over time.

Abstract: Methods and systems for providing rich content with an answer to a question query. A method includes receiving a query determined to be a question query and a corresponding answer generated in response to the question query, generating a contextual query that includes an element relating to the question query and an element relating to the answer; submitting the contextual query to a rich content search process and receiving data specifying a first set of rich content items responsive to the contextual query, determining first rich content item in the first set of rich content items that meet a context condition that is indicative of a rich content item providing contextual information of both elements of the question query and the answer query; and preferentially selecting from the first content items relative to the second rich content items to be provided as one or more answer rich content items.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for identifying similar images. In some implementations, a method is provided that includes receiving a collection of images and data associated with each image in the collection of images; generating a sparse feature representation for each image in the collection of images; and training an image similarity function using image triplets sampled from the collection of images and corresponding sparse feature representations.

Abstract: A system and method provide a soundtrack recommendation service for recommending one or more soundtrack for a video (i.e., a probe video). A feature extractor of the recommendation service extracts a set of content features of the probe video and generates a set of semantic features represented by a signature vector of the probe video. A video search module of the recommendation service is configured to search for a number of video candidates, each of which is semantically similar to the probe video and has an associated soundtrack. A video outlier identification module of the recommendation service identifies video candidates having an atypical use of their soundtracks and ranks the video candidates based on the typicality of their soundtrack usage. A soundtrack recommendation module selects the soundtracks of the top ranked video candidates as the soundtrack recommendations to the probe video.

Abstract: Disclosed herein are methods and apparatuses for compressing a video signal. In one embodiment, the method includes storing a function derived from a set of human ratings in a memory, identifying within at least a portion of the video signal at least one content-based feature, inputting the at least one identified content-based feature into the stored function, determining a compression ratio based on the function using a processor and generating a compressed video signal at the determined compression ratio.

Abstract: Systems and methods are described herein for comparing products in a marketplace. An image or video of the products may be captured using a camera associated with a mobile device. User input may be received to select two or more products within the image. Machine vision techniques may be applied to specifically identify the selected products. Product features associated with each of the identified products may be retrieved and formatted into a comparison of product features. The comparison may be presented to the user.

Abstract: A system and method provide a soundtrack recommendation service for recommending one or more soundtrack for a video (i.e., a probe video). A feature extractor of the recommendation service extracts a set of content features of the probe video and generates a set of semantic features represented by a signature vector of the probe video. A video search module of the recommendation service is configured to search for a number of video candidates, each of which is semantically similar to the probe video and has an associated soundtrack. A video outlier identification module of the recommendation service identifies video candidates having an atypical use of their soundtracks and ranks the video candidates based on the typicality of their soundtrack usage. A soundtrack recommendation module selects the soundtracks of the top ranked video candidates as the soundtrack recommendations to the probe video.