Long D Nguyen

US Patent:

7295978, Nov 13, 2007

Filed:

Sep 5, 2000

Appl. No.:

09/655325

Inventors:

Richard Mark Schwartz - Arlington MA, US
Jason Charles Davenport - Walpole MA, US
James Donald Van Sciver - Wayland MA, US
Long Nguyen - Stoneham MA, US

Assignee:

Verizon Corporate Services Group Inc. - New York NY
BBN Technologies Corp. - Cambridge MA

International Classification:

G01L 15/00

US Classification:

704240, 704238, 704230, 704250, 704243

Abstract:

A system for recognizing speech receives an input speech vector and identifies a Gaussian distribution. The system determines an address from the input speech vector () and uses the address to retrieve a distance value for the Gaussian distribution from a table (). The system then determines the probability of the Gaussian distribution using the distance value () and recognizes the input speech vector based on the determined probability ().

US Patent:

7401023, Jul 15, 2008

Filed:

Sep 6, 2000

Appl. No.:

09/656264

Inventors:

Richard Mark Schwartz - Arlington MA, US
Han Shu - Billerica MA, US
John Makhoul - Winchester MA, US
Long Nguyen - Stoneham MA, US

Assignee:

Verizon Corporate Services Group Inc. - New York NY
BBN Technologies Corp. - Cambridge MA

International Classification:

G10L 21/00
H04M 3/42

US Classification:

704275, 704277, 37921801

Abstract:

An automated directory assistance () includes a training system () and a directory assistance system (). The training system () trains and maintains the directory assistance system (). The training system () includes a transcription module (), a speech grammar estimation module (), a listings statistics estimation module (), and a required words determination module (). The transcription module () obtains transcripts relating to directory service requests. The speech grammar estimation module () creates an n-gram grammar for multiple telephone numbers from the transcripts. The listings statistics estimation module () identifies words used to refer to each of the telephone numbers from the transcripts. The required words determination module () identifies at least one word that is required to request each of the telephone numbers from the transcripts. The directory assistance system () includes a speech recognition module (), a listing retrieval module (), and an accept/reject module ().

US Patent:

7447636, Nov 4, 2008

Filed:

May 12, 2005

Appl. No.:

11/129270

Inventors:

Richard Mark Schwartz - Arlington MA, US
Han Shu - Billerica MA, US
John Makhoul - Winchester MA, US
Long Nguyen - Stoneham MA, US

Assignee:

Verizon Corporate Services Group Inc. - New York NY
BBN Technologies Corp. - Cambridge MA

International Classification:

G10L 11/00

US Classification:

704275, 379 8801, 379 8803

Abstract:

An automated directory assistance () includes a training system (). and a directory assistance system (). The training system () trains and maintains the directory assistance system (). The training system () includes a transcription module (), a speech grammar estimation module (), a listings statistics estimation module (), and a required words determination module (). The transcription module () obtains transcripts relating to directory service requests. The speech grammar estimation module () creates an n-gram grammar for multiple telephone numbers from the transcripts. The listings statistics estimation module () identifies words used to refer to each of the telephone numbers from the transcripts. The required words determination module () identifies at least one word that is required to request each of the telephone numbers from the transcripts. The directory assistance system () includes a speech recognition module (), a listing retrieval module (), and an accept/reject module ().

US Patent:

56218591, Apr 15, 1997

Filed:

Jan 19, 1994

Appl. No.:

8/183719

Inventors:

Richard M. Schwartz - Sudbury MA
Long Nguyen - Medford MA

Assignee:

BBN Corporation - Cambridge MA

International Classification:

G10L 506
G10L 900

US Classification:

395 265

Abstract:

The invention provides a method of large vocabulary speech recognition that employs a single tree-structured phonetic hidden Markov model (HMM) at each frame of a time-synchronous process. A grammar probability is utilized upon recognition of each phoneme of a word, before recognition of the entire word is complete. Thus, grammar probabilities are exploited as early as possible during recognition of a word. At each frame of the recognition process, a grammar probability is determined for the transition from the most likely preceding grammar state to a set of words that share at least one common phoneme. The grammar probability is combined with accumulating phonetic evidence to provide a measure of the likelihood that a state in the HMM will lead to the word most likely to have been spoken. In a preferred embodiment, phonetic context information is exploited, even before the complete context of a phoneme is known. Instead of an exact triphone model, wherein the phonemes previous and subsequent to a phoneme are considered, a composite triphone model is used that exploits partial phonetic context information to provide a phonetic model that is more accurate than aphonetic model that ignores context.