Abstract: Techniques for recommending a skill experience to a user after a user-system dialog session has ended are described. Upon a dialog session ending, the system uses a first machine learning model to determine potential intents to recommend to a user. The system then uses a second machine learning model to determine a particular skill and intent to recommend. The system then prompts the user to accept the recommended skill and intent. If the user accepts, the system calls the recommended skill to execute. As part of calling the skill, the system sends to the skill at least one entity provided in a natural language user input of the ended dialog session. This enables the skill to skip welcome prompts, and initiate processing to output a response based on the intent and the at least one entity of the ended dialog session.

Abstract: Techniques for storing and using multi-session context are described. A system may store context data corresponding to a first interaction, where the context data may include action data, entity data and a profile identifier for a user. Later the stored context data may be retrieved during a second interaction corresponding to the entity of the second interaction. The second interaction may take place at a system different than the first interaction. The system may generate a response during the second interaction using the stored context data of the prior interaction.

Abstract: Techniques for recommending a skill experience to a user after a user-system dialog session has ended are described. Upon a dialog session ending, the system uses a first machine learning model to determine potential intents to recommend to a user. The system then uses a second machine learning model to determine a particular skill and intent to recommend. The system then prompts the user to accept the recommended skill and intent. If the user accepts, the system calls the recommended skill to execute. As part of calling the skill, the system sends to the skill at least one entity provided in a natural language user input of the ended dialog session. This enables the skill to skip welcome prompts, and initiate processing to output a response based on the intent and the at least one entity of the ended dialog session.

Abstract: Techniques for recommending a skill experience to a user after a user-system dialog session has ended are described. Upon a dialog session ending, the system uses a first machine learning model to determine potential intents to recommend to a user. The system then uses a second machine learning model to determine a particular skill and intent to recommend. The system then prompts the user to accept the recommended skill and intent. If the user accepts, the system calls the recommended skill to execute. As part of calling the skill, the system sends to the skill at least one entity provided in a natural language user input of the ended dialog session. This enables the skill to skip welcome prompts, and initiate processing to output a response based on the intent and the at least one entity of the ended dialog session.

Abstract: Exemplary embodiments relate to methods, mediums, and systems for managing a conversation. In an embodiment, a computer-implemented input interface is provided to receive an input comprising information in natural language. A dialog manager is configured to determine an intent of the input, determine information to fulfill the intent, and identify one or both of information available to the dialog manager or information that is unavailable to the dialog manager. A conversational understanding document documents the intent and the identified information. An output interface forwards the conversational understanding document towards a task completion handler separate and distinct from the dialog manager. Other embodiments are described and claimed.

Abstract: Stream-based programming models allow subscriber to observe a stream of stream items received from a source, such as event notifications and updates to observed data. Presented herein is a stream-based reactive programming platform that allows observers to discover sources and streams, and to specify queries applicable to sources that results in new streams, including conditions that apply to the properties of a federation of sources. Variations of the reactive programming platform include generating a new stream from a non-reactive data source; generating a graph of sources and observers; and providing a directory of discoverable elements respectively identified by uniform resource identifiers.

Abstract: Exemplary embodiments relate to methods, mediums, and systems for managing a conversation. In an embodiment, a computer-implemented input interface is provided to receive an input comprising information in natural language. A dialog manager is configured to determine an intent of the input, determine information to fulfill the intent, and identify one or both of information available to the dialog manager or information that is unavailable to the dialog manager. A conversational understanding document documents the intent and the identified information. An output interface forwards the conversational understanding document towards a task completion handler separate and distinct from the dialog manager. Other embodiments are described and claimed.

Abstract: Computation can be encoded in a lightweight and optionally typed data representation. The data representation can be specified in a prefix-based notation potentially including nested function-argument pairs. Further, the data representation can comprise optional static type information associated with at least a portion of computation. Still further, the data representation can be programming language and platform independent or surfaced in a particular programming language or platform.

Abstract: Stream-based programming models allow subscriber to observe a stream of stream items received from a source, such as event notifications and updates to observed data. Presented herein is a stream-based reactive programming platform that allows observers to discover sources and streams, and to specify queries applicable to sources that results in new streams, including conditions that apply to the properties of a federation of sources. Variations of the reactive programming platform include generating a new stream from a non-reactive data source; generating a graph of sources and observers; and providing a directory of discoverable elements respectively identified by uniform resource identifiers.

Abstract: A technique for protecting the privacy of a query is provided using permissions that may be derived from an analysis of the context of the query. A monitoring component can be provided to receive or intercept queries directed at a datastore, and a privacy component is provided that analyzes permissions associated with the queries. The privacy component can also determine access levels of the queries based on the analysis of the permissions. A response component can then be provided to answer the queries in accordance with the access levels of the queries.

Abstract: Computation can be encoded in a lightweight and optionally typed data representation. The data representation can be specified in a prefix-based notation potentially including nested function-argument pairs. Further, the data representation can comprise optional static type information associated with at least a portion of computation. Still further, the data representation can be programming language and platform independent or surfaced in a particular programming language or platform.

Abstract: Techniques for defocusing queries over big datasets and dynamic datasets are provided to broaden search results and incorporate all potentially relevant data and avoid overly narrowing queries. An analytic component can receive queries directed at one region of a dataset and analyze the queries to generate inferences about the queries. The queries can then be defocused by a defocusing component and incorporate a larger dataset than originally searched to broaden the queries. The larger dataset can incorporate all, or a part of the original dataset and can also be disparate from the original dataset. Clusters of queries can also be merged and unified to deal with ‘local minima’ issues and broaden the understanding of the dataset. In other embodiments, dynamic data can be monitored and changes tracked, to ensure that all portions of the dataset are being searched by the queries.

Abstract: Permission based query processing techniques are provided for protecting privacy. A monitoring component can be provided to receive or intercept queries directed at a datastore, and a privacy component is provided that analyzes permissions associated with the queries. The privacy component can also determine access levels of the queries based on the analysis of the permissions. A response component can then be provided to answer the queries in accordance with the access levels of the queries. While the response component is answering the queries, it can also mask the data in the datastore through a variety of techniques disclosed herein. In other embodiments, query responses can be filtered based on query contexts and data from two or more datastores can be compared and similarities identified without exposing the data from either datastore.

Abstract: Techniques for defocusing queries over big datasets and dynamic datasets are provided to broaden search results and incorporate all potentially relevant data and avoid overly narrowing queries. An analytic component can receive queries directed at one region of a dataset and analyze the queries to generate inferences about the queries. The queries can then be defocused by a defocusing component and incorporate a larger dataset than originally searched to broaden the queries. The larger dataset can incorporate all, or a part of the original dataset and can also be disparate from the original dataset. Clusters of queries can also be merged and unified to deal with ‘local minima’ issues and broaden the understanding of the dataset. In other embodiments, dynamic data can be monitored and changes tracked, to ensure that all portions of the dataset are being searched by the queries.

Abstract: The subject disclosure relates to using structured query language constructs in non-structured query language domains. For example, through mathematical and logical transformation of concepts from a key, value pair domain associated with structured query language data structures to graphical-related data structures, the value originating in the structured query language domain can be modified for use in non-structured query language domains. This can open up options in analytics and can solve some of the problems associated with liner algebra.