Embodiments are directed towards managing data using modeling platform. A principal data set and match data set may be provided. The principal data set may include principal objects and the match data set may include other model objects. Blocking may associate each of the plurality of principal objects with potential match objects. Principal objects from the plurality of principal objects may be selected based on heuristics, their associated potential match objects, or the like. Potential match objects may be labeled as a true match or a non-match with respect to their associated principal object. A ranker may be trained to identify top-ranked matches based on the labeled potential match objects. The trained ranker may be employed to indicate matches where each matching other model object is a potential match object that is ranked higher than other potential match objects that are associated with its matched principal object.

Systems and methods are described to address shortcomings in a conventional conversation system via a novel technique utilizing artificial neural networks to train the conversation system whether or not to continue context. In some aspects, an interactive media guidance application determines a type of conversation continuity in a natural language conversation comprising first and second queries. The interactive media guidance application determines a first token in the first query and a second token in the second query. The interactive media guidance application identifies entity data for the first and second tokens. The interactive media guidance application retrieves, from a knowledge graph, graph connections between the entity data for the first and second tokens. The interactive media guidance application applies this data as inputs to an artificial neural network. The interactive media guidance application determines an output that indicates the type of conversation continuity between the first and second queries.

Methods, apparatus, systems, and computer-readable media are provided for storing incomplete dialog sessions between a user and an automated assistant in order that the dialog sessions can be completed in furtherance of certain actions. While interacting with an automated assistant, a user can become distracted and not complete the interaction to the point of the automated assistant performing some action. In response, the automated assistant can store the interaction as a dialog session. Subsequently, the user may express interest, directly or indirectly, in completing the dialog session, and the automated assistant can provide the user with a selectable element that, when selected, causes the dialog session to be reopened. The user can then continue the dialog session with the automated assistant in order that the originally intended action can be performed by the automated assistant.

Disclosed are systems and methods for context modification of queries to refine search results. A user may make a force-interval selection of one or more display elements displayed on the screen of a mobile computing device. The selected display elements may be analyzed to determine context. One or more context options may be displayed onscreen for user selection. A selected context option can cause the mobile computing device to modify the original query to generate a context-modified query. The context-modified query can be used to retrieve refined search results.

A method for extracting entities from a text document includes, for at least a section of a text document, providing a first set of entities extracted from the at least a section, clustering at least a subset of the extracted entities in the first set into clusters, based on locations of the entities in the document. Complete ones of the clusters of entities are identified. Patterns for extracting new entities are learned based on the complete clusters. New entities are extracted from incomplete clusters based on the learned patterns.

A method of disambiguating user queries in a multi-turn dialogue including a set of user utterances. The method includes using a predefined language model to recognize an ambiguous entity in an unresolved user utterance from the multi-turn dialogue, and using the predefined language model to recognize entity constraints of the ambiguous entity. The method further includes, in a computer-accessible conversation history of the multi-turn dialogue, searching a set of previously-resolved entities for a candidate entity having entity properties with a highest confidence correspondence to the entity constraints of the ambiguous entity. The unresolved user utterance is rewritten as a rewritten utterance that replaces the ambiguous entity with the candidate entity. The rewritten utterance is output to one or more query answering machines.

A method and system for improving an intent based search is provided. The method includes analyzing a search phase entered by a user with respect to a Website level search query for specified subject matter. In response, a subject based intent classification is determined to be associated with a confidence factor of less than 100 percent confidence with respect to the subject based intent classification being correct. The subject based intent classification is compared to search results data and a subset of search results correlating to the subject based intent classification is determined and presented to the user. An interaction of the user with respect to the subset of search results is determined and in response, a value of the confidence factor is derived.

A method for retrieving information includes receiving a search query within an information corpus. Search results for the search query may be identified. A score for each of a plurality of data items identified in the search results may be generated. The score for a corresponding one of the plurality of data items may be based on a score dependent on the search query within the information corpus. The score may be also based on at least one score independent of the search query. The at least one score independent of the search query may include a ranking signal associated with a World Wide Web (WWW) search of the corresponding one of the plurality of data items using a second information corpus. The search results may be ranked based on the generated score.

An apparatus includes a memory and a hardware processor that receives a query from a device. The query includes first search parameters. The processor also retrieves, from a database and based on the first search parameters, a plurality of previously issued queries and applies a machine learning algorithm on the plurality of previously issued queries to determine second search parameters. The processor further adds the second search parameters to the query to form an enhanced query and communicates the enhanced query to a plurality of response systems. The processor then receives, from the plurality of response systems, a plurality of responses to the enhanced query, constructs, based on the plurality of responses to the enhanced query, an enhanced response to the query, and communicates the enhanced response to the device for selection of a response from the plurality of responses.

A selected target location is used as a basis for selecting an alternate target location for distribution of content. The alternate target location is selected based on historical information about user activity involving the selected target location and may include an analysis of time records associated with locations included in the historical data.