Processing

Please wait...

Settings

Settings

Goto Application

1. WO2020117507 - TRAINING SPEECH RECOGNITION SYSTEMS USING WORD SEQUENCES

Publication Number WO/2020/117507
Publication Date 11.06.2020
International Application No. PCT/US2019/062872
International Filing Date 22.11.2019
IPC
G10L 15/197 2013.01
GPHYSICS
10MUSICAL INSTRUMENTS; ACOUSTICS
LSPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
15Speech recognition
08Speech classification or search
18using natural language modelling
183using context dependencies, e.g. language models
19Grammatical context, e.g. disambiguation of recognition hypotheses based on word sequence rules
197Probabilistic grammars, e.g. word n-grams
G06F 21/62 2013.01
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
FELECTRIC DIGITAL DATA PROCESSING
21Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
60Protecting data
62Protecting access to data via a platform, e.g. using keys or access control rules
G06F 40/44 2020.01
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
FELECTRIC DIGITAL DATA PROCESSING
40Handling natural language data
40Processing or translation of natural language
42Data-driven translation
44Statistical methods, e.g. probability models
G10L 15/26 2006.01
GPHYSICS
10MUSICAL INSTRUMENTS; ACOUSTICS
LSPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
15Speech recognition
26Speech to text systems
G10L 15/06 2013.01
GPHYSICS
10MUSICAL INSTRUMENTS; ACOUSTICS
LSPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
15Speech recognition
06Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice
CPC
G06F 21/6254
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
FELECTRIC DIGITAL DATA PROCESSING
21Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
60Protecting data
62Protecting access to data via a platform, e.g. using keys or access control rules
6218to a system of files or objects, e.g. local or distributed file system or database
6245Protecting personal data, e.g. for financial or medical purposes
6254by anonymising data, e.g. decorrelating personal data from the owner's identification
G06F 40/279
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
FELECTRIC DIGITAL DATA PROCESSING
40Handling natural language data
20Natural language analysis
279Recognition of textual entities
G06F 40/30
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
FELECTRIC DIGITAL DATA PROCESSING
40Handling natural language data
30Semantic analysis
G06F 40/44
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
FELECTRIC DIGITAL DATA PROCESSING
40Handling natural language data
40Processing or translation of natural language
42Data-driven translation
44Statistical methods, e.g. probability models
G10L 15/063
GPHYSICS
10MUSICAL INSTRUMENTS; ACOUSTICS
LSPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
15Speech recognition
06Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice
063Training
G10L 15/183
GPHYSICS
10MUSICAL INSTRUMENTS; ACOUSTICS
LSPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
15Speech recognition
08Speech classification or search
18using natural language modelling
183using context dependencies, e.g. language models
Applicants
  • SORENSON IP HOLDINGS, LLC [US]/[US]
Inventors
  • THOMSON, David
  • ADAMS, Jadie
  • BOEHME, Kenneth
Agents
  • PARKE, Brian
  • BENNS, Jonathan, M.
  • ATZET, Ian, A.
  • BARBER, Daniel, R.
  • BRAITHWAITE, Jared, J.
Priority Data
16/209,64004.12.2018US
Publication Language English (EN)
Filing Language English (EN)
Designated States
Title
(EN) TRAINING SPEECH RECOGNITION SYSTEMS USING WORD SEQUENCES
(FR) APPRENTISSAGE POUR DES SYSTÈMES DE RECONNAISSANCE VOCALE À L'AIDE DE SÉQUENCES DE MOTS
Abstract
(EN)
A method may include obtaining first audio data of a communication session between a first device and a second device, obtaining a text string that is a transcription of the first audio data, and selecting a contiguous sequence of words from the text string as a first word sequence. The method may further include comparing the first word sequence to multiple word sequences obtained before the communication session and in response to the first word sequence corresponding to one of the multiple word sequences, incrementing a counter of multiple counters associated with the one of the multiple word sequences. The method may also include deleting the text string and the first word sequence and training and after deleting the text string and the first word sequence, training a language model of an automatic transcription system using the multiple word sequences and the multiple counters. The method is for estimating n-gram statistics based on the communication and provides data protection mechanisms for private data.
(FR)
La présente invention concerne un procédé pouvant comprendre les étapes consistant à obtenir des premières données audio d'une session de communication entre un premier dispositif et un second dispositif, obtenir une chaîne de texte qui est une transcription des premières données audio, et sélectionner une séquence contiguë de mots à partir de la chaîne de texte comme première séquence de mots. Le procédé peut en outre comprendre les étapes consistant à comparer la première séquence de mots à de multiples séquences de mots obtenues avant la session de communication et, en réponse au fait que la première séquence de mots correspond à l'une des multiples séquences de mots, incrémenter un compteur parmi de multiples compteurs associés à l'une des multiples séquences de mots. Le procédé peut également comprendre les étapes consistant à supprimer la chaîne de texte et la première séquence de mots et à effectuer un apprentissage, puis après avoir supprimé la chaîne de texte et la première séquence de mots, à procéder à l'entraînement d'un modèle de langue d'un système de transcription automatique à l'aide des multiples séquences de mots et des multiples compteurs. Le procédé est conçu pour estimer des statistiques de n-gramme sur la base de la communication et fournit des mécanismes de protection de données pour des données privées.
Also published as
Latest bibliographic data on file with the International Bureau