Microsoft Technology Licensing, LLC

Nuance Communications, Inc.

Simon H Corston

David Jon Kay

Ethan Robert Bradford

Keith Trnka

Donni McCray

William B Dolan

Raman Chandrasekar

Lucy H Vanderwende

Lisa Braden-Harder

Harr Chen

Robert George Atkinson

Eric David Brill

Malcolm Erik Pearson

Michael Gamon

David Richard Reed

Erland Unruh

Eric K Ringger

Richard G Campbell

Claes-Fredrik Mannby

Miguel Cardoso De Campos

Wendy Bannister

Technology for improving the predictive accuracy of input word recognition on a device by dynamically updating the lexicon of recognized words based on the word choices made by similar users. The technology collects users' vocabulary choices (e.g., words that each user uses, or adds to or removes from a word recognition dictionary), associates users who make similar choices, aggregates related vocabulary choices, filters the words, and sends words identified as likely choices for that user to the user's device. Clusters may include, for example, users in a particular location (e.g., sets of people who use words such as 'Puyallup,' 'Gloucester,' or &#x201c;Waiheke&#x201d;), users with a particular professional or hobby vocabulary, or application-specific vocabulary (e.g., word choices in map searches or email messages).

Updating population language models based on changes made by user clusters

The disclosed system provides an efficient method of using a later word or words entered after a previous word along with one or more language models that show which words are likely to occur together to identify a better disambiguated choice for the previous word. To identify the better disambiguated choice for the previous word, the system can evaluate the conditional probability for the later word of various candidate previous words, and select the candidate previous word that has the highest conditional probability. If the conditional probability of the selected candidate previous word exceeds that of the previous word that was entered by at least a factor, then the system can include the selected candidate previous word in a selection list for user selection. The disclosed system also provides an efficient method for using one or more language models and a later word to correct errors in segmenting the word.

Correction of previous words and other user text input errors

System and methods are provided for detecting numerical text strings within a text string and converting those numerical text strings into digit strings. The digit strings may be reflected in real-time, such as when the user is typing a text message. If more than one possible format of the digit string is determined, the system may then provide a selection of the various formats for selection. Once the proper format for the digit string is determined, that digit string may replace the numerical string previously detected in the text string. The text to digit conversion and associated formatting expedites user text entry such that the user is not required to switch keyboard views, (e.g., virtual keyboards). Additionally, converting to digit strings compresses message length, as well as provide other benefits.

Converting text strings into number strings, such as via a touchscreen input

The present invention relates to systems and methods that determine intent for received data (e.g., email, voice, graphics . . . ) and respond to the data based on the intent. The systems and methods employ various combinations of features based on shallow and deep linguistic analysis (e.g., semantic and syntactic) to yield very high accuracy. The systems and methods analyze and categorize received data to locate data that can include intent. This data can be further refined by extracting features related to the intent. The features can be utilized by a classifier to determine the intent. If the intent warrants a response, the data are further scrutinized and reformulated to generate a description that is indicative of the intent. The reformulation can include representing the features in a logical form, transforming the form and generating a description of the intent that can be presented to a user visually and/or audibly.

Systems and methods that determine intent of data and respond to the data based on the intent

A system that facilitates performance of a focused search over a collection of sites comprises a subweb that corresponds to a topic and/or user characteristic(s) that are of interest to the user. The subweb includes a plurality of domains and/or paths (e.g. sites) that are related to the topic and/or the user characteristic(s). Each of the sites within the subweb is assigned a weight that indicates relevance of the site to the desirable topic and/or user characteristic(s). A search engine employs the subweb to facilitate focusing a search over a collection of sites. The search engine receives a query, and utilizes the subweb to focus a search over the selection of sites corresponding to the topic and/or user characteristic(s) represented by the subweb. The results from the search are returned to the user based at least in part upon the relevance weights assigned to the sites within the subweb.

Building and using subwebs for focused search

The present invention relates to a system and methodology that applies automated learning procedures for determining document relevance and assisting information retrieval activities. A system is provided that facilitates a machine-learned approach to determine document relevance. The system includes a storage component that receives a set of human selected items to be employed as positive test cases of highly relevant documents. A training component trains at least one classifier with the human selected items as positive test cases and one or more other items as negative test cases in order to provide a query-independent model, wherein the other items can be selected by a statistical search, for example. Also, the trained classifier can be employed to aid an individual in identifying and selecting new positive cases or utilized to filter or re-rank results from a statistical-based search.

Machine-learned approach to determining document relevance for search over large electronic collections of documents

A system filters documents in a document set retrieved from a document store in response to a query. The system obtains a first set of logical forms based on a selected one of the query and the documents in the document set. The system obtains a second set of logical forms based on another of the query and the documents in the document set. The system then uses natural language processing techniques to modify the first logical forms to obtain a modified set of logical forms. The system filters documents in the document set based on a predetermined relationship between the modified set of logical forms and the second set of logical forms.

System for processing textual inputs using natural language processing techniques

The present invention is directed to recognizing a discourse structure of a body of text. In a preferred embodiment, a discourse structure recognition facility utilizes syntactic information associated with the body of text to generate a discourse structure tree that characterizes the discourse structure of the body of text. The facility first identifies in the body of text a number of clauses. The facility then determines, for each distinct pair of clauses, which of a number of possible discourse relations should be hypothesized between the pair of clauses, based on the syntactic structure and semantic of the body of text relative to the pair of clauses. The facility then applies the hypothesized relations to the clauses in order to produce a discourse structure tree characterizing the discourse structure of the body of text. In certain embodiments, the facility further generates from the produced discourse structure tree a synopsis of the body of text reflecting the primary goals pursued by its author.

Automatically recognizing the discourse structure of a body of text

Apparatus and accompanying methods for an information retrieval system that utilizes natural language processing to process results retrieved by, for example, an information retrieval engine such as a conventional statistical-based search engine, in order to improve overall precision. Specifically, such a search ultimately yields a set of retrieved documents. Each such document is then subjected to natural language processing to produce a set of logical forms. Each such logical form encodes, in a word-relation-word manner, semantic relationships, particularly argument and adjunct structure, between words in a phrase. A user-supplied query is analyzed in the same manner to yield a set of corresponding logical forms therefor. Documents are ranked as a predefined function of the logical forms from the documents and the query. Specifically, the set of logical forms for the query is then compared against a set of logical forms for each of the retrieved documents in order to ascertain a match between any such logical forms in both sets. Each document that has at least one matching logical forms is heuristically scored, with each different relation for a matching logical forms being assigned a different corresponding predefined weight. The score of each such document is, e.g., a predefined function of the weights of its uniquely matching logical forms. Finally, the retained documents are ranked in order of descending score and then presented to a user in that order.

Growing community of inventors

Seattle, WA, United States of America

Simon H Corston