Microsoft Technology Licensing, LLC

Arungunram Chandrasekaran Surendran

Tarek Najm

Viswanath Vadlamani

Abhinai Srivastava

John Carlton Platt

Gregory Ray Smith

Nuria M Oliver

Phani Vaddadi

Munirathnam Srikanth

Chintan S Thakkar

Rajeev Prasad

Jack W Stokes

Milind V Mahajan

Erin L Renshaw

Patrick Nguyen

Ying Li

Jack Wilson Stokes, Iii

Chao He

Christopher J C Burges

Raman Chandrasekar

Qin Li

Flavia Silvana Moser

Michael K Forney

Dean A Slawson

Piali Choudhury

Asif Iqbal Mohammad

Siva Mohan

Somsak Sukittanon

Prasanth Pulavarthi

Mahbubul Alam Ali

Suvrit Sra

John Midgley

Martin Ester

Embodiments of a computer-implemented advertisement display system are disclosed. In one embodiment, the system includes a speech recognition component that processes a video clip and produces a corresponding collection of speech recognition data indicative of an audio portion of the video clip. The system also includes a collection of advertising material. An advertisement selection component selects an advertisement from the collection of advertising material based on the corresponding collection of speech recognition data. The system also includes a display. An advertisement presentation component displays an indication of the selected advertisement on the display during a simultaneous display of the video clip.

Video contextual advertisements using speech recognition

A framework is provided for obtaining window information. The window information can be applied to different assignment models to assign windows to different groups. A group may correspond to a task being performed by a user. The window information can be semantic or temporal information captured as window events and properties of windows whose events are captured. Temporal information can be information about switches between windows. Semantic information can be window titles. Temporal information, semantic information, or both, can be used to assign windows to groups.

Window grouping

Search result re-ranking is provided by employing a concept graph from a metabase. When a search query is received, a query context of the search query is analyzed to identify dominant concepts for the search query. The dominant concepts are expanded by identifying the dominant concepts within a concept graph and identifying additional concepts having a strong relationship with the dominant concepts within the concept graph. A set of search results for the search query is analyzed to determine strength of relationship of each search result to the expanded concepts. The search results are re-ranked based on the strength of relationship of each search result to the expanded concepts and the strength of relationship of each expanded concept to the dominant concepts.

Re-ranking search results based on lexical and ontological concepts

A procedural inference system is described herein that infers procedural knowledge from various data sources to help a user complete one or more tasks for which the data sources provide information. The system understands users' queries, identifies a task at hand, provides recommendations on the steps to take and the agents to use based on a knowledge base of tasks and agents, and provides the fabric to determine which different agents can work together to help the user accomplish a task. Tasks can be started on one device and completed on another seamlessly. Users are able to finish complex, multi-step tasks efficiently, without trial and error or data reentry. Thus, the procedural inference system provides a generalized framework that helps users to complete tasks using already available data and does not ask each data provider to invest in infrastructure to build dedicated task information systems.

Inferring procedural knowledge from data sources

Systems, methods, and computer-storage media for generating lateral concepts are provided. The system includes a search engine to receive user queries, a storage to store content and its associated categories, and a lateral concept generator. The lateral concept generator is connected to both the search engine and storage. The lateral concept generator selects lateral concepts from categories associated with the content based on similarity scores for the stored content.

Generating and presenting lateral concepts

Systems, methods, and computer-storage media for generating and providing knowledge content to users utilizing a web architecture that integrates information across data silos through a common, flexible data storage schema, such as a star or snowflake schema, are provided. Data from a content graph, a user activity graph, a social graph, and temporal data as it relates to each of the content graph, the social graph and the user activity graph, is stored in a knowledge content database utilizing the star schema. In this way, data from each of these formerly disparate sources may be accessed from a common, extensible application platform utilizing ontologies and pivot table functionality, thus providing smarter, more comprehensive knowledge in response to received user queries.

Providing knowledge content to users

Methods, systems, and computer-readable media for identifying intentions related to a user's present intention are provided. The user's present intention may be determined by mapping the user's current actions to a present intent. A related intent may be determined using an intent chain. The intent chain may be generated by analyzing patterns of user activities and mapping the patterns of activities to the underlying intents.

Anticipating interests of an online user

A regression-based residual echo suppression (RES) system and process for suppressing the portion of the microphone signal corresponding to a playback of a speaker audio signal that was not suppressed by an acoustic echo canceller (AEC). In general, a prescribed regression technique is used between a prescribed spectral attribute of multiple past and present, fixed-length, periods (e.g., frames) of the speaker signal and the same spectral attribute of a current period (e.g., frame) of the echo residual in the output of the AEC. This automatically takes into consideration the correlation between the time periods of the speaker signal. The parameters of the regression can be easily tracked using adaptive methods. Multiple applications of RES can be used to produce better results and this system and process can be applied to stereo-RES as well.

System and process for regression-based residual acoustic echo suppression

Electronic documents are analyzed to identify assertions, which are inverted to generate questions that may be answered by the assertions. A document or a corpus of electronic documents may be analyzed to identify entities and relationships among entities within the text of the document(s). Assertions are identified based on the entities and relationships among the entities. Each assertion represents a fact about an entity, and a group of assertions represents a summary of the document or document corpus. The assertions are inverted to generate questions that may be answered by the assertions. The questions may be further analyzed to identify relevant concepts and topics and to cluster the questions around the concepts and topics. A combined graph may also be generated that facilitates traversal among topics, concepts, questions, assertions, document summaries, and documents.

Ontological categorization of question concepts from document summaries

A table of contents is generated for search results in response to a search query by identifying semantic concepts as topics for inclusion within the table of contents. When a search query is received, a search is performed to identify search results. The search results are compared to an ontology of topics to identify relevant topics. Additionally, search results are compared to an ontology of partial topics to identify relevant partial topics, which are named. Independent key-phrases are further generated from search results and identified as key-phrase topics. The identified topics, named partial topics, and key-phrase topics are ranked, and topics are selected for inclusion within a table of contents. A search results page is returned in response to the search query that includes search results and the generated table of contents.

Semantic table of contents for search results

An interactive voice advertisement exchange presents an interactive voice advertisement to a user of a voice network who is placed on hold. After the user sends a signal response from a user device to a receiving system and is placed on hold, the hold time and a parameter associated with the signal response are received from the receiving system. Using the hold time and the parameter, an interactive voice advertisement is selected from one or more interactive voice advertisements sent by advertisers. The selected interactive voice advertisement includes a voice message and an application that dictates how the user can interact with the voice message. The voice message and the application of the selected voice interactive voice advertisement are sent to the receiving system. The receiving system sends the voice message to the user device and executes the application during a portion of the hold time.

Interactive voice advertisement exchange

Advertisements are selected for presentation on search result pages and web pages based on phrases generated from lateral concepts and topics identified for the search result pages and web pages. A search query or an indication of a web page is received for which advertisements are to be provided. Lateral concepts and topics are identified based on the search query or content of the web page. The lateral concepts and topics are used as phrases for selecting advertisements from an advertisement inventory. Selected advertisements are provided for presentation on a search results page in response to a search query or on a web page initially identified.

Growing community of inventors

Sammamish, WA, United States of America

Arungunram Chandrasekaran Surendran