Abstract: Systems and techniques for sequence to sequence conversational query understanding are described herein. A query may be received that includes multiple words. It may be identified that the query is to be reformulated based on an attention value for an attention word in the query. Relationships may be determined among words of the query and words in a previously submitted query and words in results from the previously submitted query. The query may be reformulated based on the relationships. The reformulated query may be employed to retrieve query results.

Abstract: Searching is assisted by recognizing a selection of text from a document as an indication that a user wishes to initiate a search based on the selected text. The user is provided with query suggestions based on the selected text and the query suggestions are ranked based on a context provided by the document. The user may select the text by using a mouse, drawing a circle around the text on a touch screen, or by other input techniques. The query suggestions may be based on query reformulation or query expansion techniques applied to the selected text. Context provided by the document is used by a language model and/or an artificial intelligence system to rank the query suggestions in predicted order of relevance based on the selected text and the context.

Abstract: A system and method for generating modifiers for updated search queries are provided. In example embodiments, metadata is accessed, the metadata corresponds to search results of an input query and comprising a plurality of candidate modifiers. A score is calculated for each candidate based on a relevance value that indicates the correlation between a candidate modifier and the input query. A list of top number of candidate modifiers is generated based on the score of the candidate modifier transgressing a first threshold. A uniqueness score is calculated for combination pairs of candidate modifiers within the list, the uniqueness score being used to eliminate candidate modifiers. The list of top number of candidate modifiers is presented, at a user interface, according to a ranked order based on the score.

Abstract: Systems and techniques for sequence to sequence conversational query understanding are described herein. A query may be received that includes multiple words. It may be identified that the query is to be reformulated based on an attention value for an attention word in the query. Relationships may be determined among words of the query and words in a previously submitted query and words in results from the previously submitted query. The query may be reformulated based on the relationships. The reformulated query may be employed to retrieve query results.

Abstract: A system and method for generating modifiers for updated search queries are provided. In example embodiments, metadata is accessed, the metadata corresponds to search results of an input query and comprising a plurality of candidate modifiers. A score is calculated for each candidate based on a relevance value that indicates the correlation between a candidate modifier and the input query. A list of top number of candidate modifiers is generated based on the score of the candidate modifier transgressing a first threshold. A uniqueness score is calculated for combination pairs of candidate modifiers within the list, the uniqueness score being used to eliminate candidate modifiers. The list of top number of candidate modifiers is presented, at a user interface, according to a ranked order based on the score.

Abstract: Techniques for utilizing data mining technology to extract universal topics with multilingual representations from a multilingual database, and to organize existing or new documents in different languages by analyzing their respective topic distributions.

Abstract: A computer-readable storage medium having stored thereon computer-executable instructions which, when executed by a computer, cause the computer to implement an opinion search engine. The instructions to implement an opinion search engine cause the computer to collect opinion data about one or more objects from the Internet, extract metadata about the opinion data from the opinion data, remove duplicate metadata from the metadata to generate a resulting metadata, categorize the resulting metadata for similar objects according to one or more taxonomies from one or more websites on the Internet and rank the similar objects based on the categorized metadata.

Abstract: Techniques and technologies for expanding a gallery by seeding the gallery with seed query results. A template is identified which is associated with the subjects of the seed queries and content identifiers are identified which include the template. These content identifiers are used to expand the gallery to include more content identifiers than before the expansion of the gallery.

Abstract: Embodiments of the present invention relate to systems, methods, and computer-storage media for a method of contextually analyzing terms within a search query. In one embodiment, a received search query is classified into a domain category. Additionally, information is assigned to a schema associated with the domain by analyzing the search query. Further, at least one search result that helps a user complete a task within the domain is provided based on the information in the schema.

Abstract: Documents such as web pages may be regarded as offering various actions; e.g., a website for a movie theater may offer options for viewing movie listings and purchasing tickets. A user may wish to view the set of actions available for a particular document, and/or the performance of an action. However, it may be difficult to identify available actions with acceptable accuracy in an automated manner, and the set of documents (such as the entire worldwide web) may be too voluminous for human identification. In order to identify available actions, the document may be searched for menus containing options, and identifying the actions associated with each option according to an option score. Additionally, documents may be grouped into document categories (e.g., websites for movie theaters and websites for musicians) to facilitate the association options in similar documents with similar sets of actions that are often provided for such documents.

Abstract: Techniques for utilizing data mining technology to extract universal topics with multilingual representations from a multilingual database, and to organize existing or new documents in different languages by analyzing their respective topic distributions.

Abstract: Providing sentiment classification of out of domain data are disclosed herein. In some aspects, a source domain having a trained classifier is matched to a target domain having a target classifier. The trained classifier may include identifiers that may be used to predict the sentiment of opinion data for the source domain. The target classifier may use the identifiers of the trained classifier to determine the sentiment of opinion data for the target domain.

Abstract: A method for identifying transactional queries includes associating user queries with forms clicks on by users who employ a search engine to place the queries during query sessions. A score is assigned to each user query. The score reflects a likelihood that the respective query is a transactional query. The query is classified as a transactional query if the score exceeds a threshold value.

Abstract: Techniques for utilizing data mining technology to extract universal topics with multilingual representations from a multilingual database, and to organize existing or new documents in different languages by analyzing their respective topic distributions.

Abstract: An invisible control may be implemented in a client device or in an application of the client device. A user may activate the invisible control by applying a gesture on a predetermined region of the client device or the application. In response to receiving the user gesture, a predetermined action associated with the invisible control may be activated. The predetermined action may be applied to the application or some or all of the content associated with the application. An Application Programming Interface may further be provided to allow the user, an application vendor or a content provider to customize the invisible control or operating modes associated with activation of the invisible control.

Abstract: Providing sentiment classification of out of domain data are disclosed herein. In some aspects, a source domain having a trained classifier is matched to a target domain having a target classifier. The trained classifier may include identifiers that may be used to predict the sentiment of opinion data for the source domain. The target classifier may use the identifiers of the trained classifier to determine the sentiment of opinion data for the target domain.

Abstract: Providing sentiment classification of out of domain data are disclosed herein. In some aspects, a source domain having a trained classifier is matched to a target domain having a target classifier. The trained classifier may include identifiers that may be used to predict the sentiment of opinion data for the source domain. The target classifier may use the identifiers of the trained classifier to determine the sentiment of opinion data for the target domain.

Abstract: Searching is assisted by recognizing a selection of text from a document as an indication that a user wishes to initiate a search based on the selected text. The user is provided with query suggestions based on the selected text and the query suggestions are ranked based on a context provided by the document. The user may select the text by using a mouse, drawing a circle around the text on a touch screen, or by other input techniques. The query suggestions may be based on query reformulation or query expansion techniques applied to the selected text. Context provided by the document is used by a language model and/or an artificial intelligence system to rank the query suggestions in predicted order of relevance based on the selected text and the context.

Abstract: The techniques describe recommending mobile query suggestions by integrating time and location in response to a query input submitted on a mobile computing device. A process constructs a bipartite graph by extracting users that submitted queries from mobile search logs and associating time and location with the submitted queries. The process determines the users are similar having submitted similar queries at similar times and at similar locations. The process receives a query input associated with a current time and a current location of a mobile computing device. Next, the process computes a relatedness of candidate queries to the query input based on a similarity between the user groups having submitted both the candidate queries and the query input, and distances of times and locations at which the user previously issued the query input and the candidate queries.

Abstract: Embodiments of the present invention relate to systems, methods, and computer-storage media for a method of contextually analyzing terms within a search query. In one embodiment, a received search query is classified into a domain category. Additionally, information is assigned to a schema associated with the domain by analyzing the search query. Further, at least one search result that helps a user complete a task within the domain is provided based on the information in the schema.