Microsoft Technology Licensing, LLC

Apple Inc.

Hsiao-Wuen Hon

Kuansan Wang

Xuedong D Huang

Stephen Frederick Potter

Francisco Manuel Galanes

Yen-Lu Chow

James D Jacoby

Wei-Ying Ma

Ming Zhou

Zheng Chen

Jianfeng Gao

Benyu Zhang

Hua-Jun Zeng

Daniel B Cook

Yun-Cheng Ju

Li Jiang

Susan M Warren

Renaud Julien Lecoeuche

Karen E Fries

Adam B Fineberg

Gabor Hirschler

Kurt Samuelson

Renaud J Lecoueche

Shipeng Li

Alejandro Acero

Dong Li

Joshua T Goodman

Lie Lu

Yan Lu

Zhiwei Li

William B Dolan

Ji-Rong Wen

Milind V Mahajan

Mei-Yuh Hwang

Zaiqing Nie

Huifeng Shen

Peter V De Souza

Michael J Rozak

Sonja Sabina Knoll

Cheng Niu

Asela J R Gunawardana

Kai-Fu Lee

Adwait Ratnaparkhi

Peter Vincent deSouza

Daniel S Venolia

Eric A Hulteen

Xianghui Zhang

Chen Harr

Yun-Chen Ju

Bao-Sheng Yuan

Gerald T Beauregard

Baosheng Yuan

A home cloud computing system employs a virtualization system to virtualize data of a device and adaptively transform type or format of the virtualized data for one or more other devices, thus leveraging resources of the device for the one or more other devices. Through data virtualization and adaptive transformation, devices of heterogeneous types are seamlessly connected to one another and can act as input or output devices for each other to create a home cloud network of devices.

Home cloud with virtualized input and output roaming over network

A classifier is built to rank documents of different languages found in a query based at least in part on similarity to other documents and the relevance of those other documents to the query. A joint ranking model, e.g., based upon a Boltzmann machine, is used to represent the content similarity among documents, and to help determine joint relevance probability for a set of documents. The relevant documents of one language are thus leveraged to improve the relevance estimation for documents of different languages. In one aspect, a hidden layer of units (neurons) represents clusters (corresponding to relevant topics) among the retrieved documents, with an output layer representing the relevant documents and their features, and edges representing a relationship between clusters and documents.

Joint ranking model for multilingual web search

Web server controls are provided for generating client side markups with recognition and/or audible prompting. Three approaches are disclosed for implementation of the controls.

Web server controls for web enabled recognition and/or audible prompting

Controls are provided for a web server to generate client side markups that include recognition and/or audible prompting. The controls comprise elements of a dialog such as a question, answer, confirmation, command or statement. A module forms a dialog by making use of the information carried in the controls.

Application abstraction with dialog purpose

Collocation errors can be automatically proofed using local and network-based corpora, including the Web. For example, according to one illustrative method, one or more collocations from a text sample are compared with a corpus such as the content of the Web. The collocations are identified for whether they are disfavored in the corpus. Indications are provided via an output device of whether the collocations are disfavored in the corpus. Additional steps may then be taken such as searching for and providing potentially proper word collocations via a user output.

Proofing of word collocation errors based on a comparison with collocations in a corpus

A content object indexing process including creating a content object knowledge index, calculating a description vector of a target content object, and indexing the target content object by searching for the description vector in the content object knowledge database. It may be difficult to search for an exact content object such as a music file or academic researcher as a conventional search index may not include related hierarchical information. A content object indexing process may add hierarchical information taken from a content object knowledge index and incorporate the hierarchical information to the index entry for a specific content object. An application of such a content object indexing process may be a world wide web search engine.

Content object indexing using domain knowledge

A method and apparatus determine the likelihood of a sequence of words based in part on a segment model. The segment model includes trajectory expressions formed as the product of a polynomial matrix and a generation matrix. The likelihood of the sequence of words is based in part on a segment probability derived by subtracting the trajectory expressions from a feature vector matrix that contains a sequence of feature vectors for a segment of speech. Aspects of the method and apparatus also include training the segment model using such a segment probability.

Speech recognition method and apparatus utilizing segment models

A markup language for execution on a client device in a client/server system includes extensions for recognition.

Markup language extensions for web enabled recognition

A sentence is accessed and at least one query is generated based on the sentence. At least one query can be compared to text within a collection of documents, for example using a web search engine. Collocation errors in the sentence can be detected and/or corrected based on the comparison of the at least one query and the text within the collection of documents.

Processing collocation mistakes in documents

A server/client system for processing data includes a network having a web server with information accessible remotely. A client device includes a microphone and a rendering component such as a speaker or display. The client device is configured to obtain the information from the web server and record input data associated with fields contained in the information. The client device is adapted to send the input data to a remote location with an indication of a grammar to use for recognition. A recognition server receives the input data and the indication of the grammar. The recognition server returns data indicative of what was recognized to at least one of the client and the web server.

Web enabled recognition architecture

A computer implemented information retrieval system is provided. The system includes a user input configured to receive a user query relative to the corpus. A machine learning classifier is trained with a first set of training data comprising anchor text relative to at least some of the documents in the corpus. A processing unit is adapted to interact with the classifier to obtain search results relative to the query using the machine learning classifier. In some aspects, the classifier is also trained with a second set of training data. A method of integrating a new document into a corpus of documents is also provided. A method of training a machine learning classifier for retrieving documents from a corpus using two distinct types of training data is also provided.

System and method for using anchor text as training data for classifier-based search systems

A markup language for execution on a client device in a client/server system includes instructions to unify at least one of recognition-related events, GUI events and telephony events on non-display, voice input based client device and a multimodal based client for a web server interacting with each of the client devices. A recognition server for receiving data indicative of inputted data provided to a client device and an indication of a grammar to use for recognition is also provided.

Servers for web enabled speech recognition

A method of interacting with a client/server architecture with a 2.5G mobile phone having a data channel for transmitting data and a voice channel for transmitting speech. The method includes receiving a web page from a web server pursuant to an application through the data channel and rendering the web page on the 2.5G phone, where rendering comprises processing the web page to be responsive speech input. Speech is received from the user corresponding to at least one data field on the web page. A call is established from the 2.5G phone to a telephony server over the voice channel. The telephony server is remote from the 2.5G phone and adapted to process speech. A speech-enabled web page is obtained from the web server corresponding to the web page provided to the 2.5G phone. Speech is transmitted from the 2.5G phone to the telephony server. The speech is processed in accordance with the speech-enabled web page to obtain textual data in accordance with the speech. The textual data is transmitted to the web server. A new web page is obtained on the 2.5G phone through the data channel and rendered having the textual data.

Sequential multimodal input

Growing community of inventors

Bellevue, WA, United States of America

Hsiao-Wuen Hon