Processing

Please wait...

Settings

Settings

Goto Application

1. WO2021038364 - SOFT-FORGETTING FOR CONNECTIONIST TEMPORAL CLASSIFICATION BASED AUTOMATIC SPEECH RECOGNITION

Publication Number WO/2021/038364
Publication Date 04.03.2021
International Application No. PCT/IB2020/057719
International Filing Date 17.08.2020
IPC
G10L 15/16 2006.1
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
16using artificial neural networks
G06N 3/08 2006.1
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
08Learning methods
CPC
G06N 3/0445
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
04Architectures, e.g. interconnection topology
0445Feedback networks, e.g. hopfield nets, associative networks
G06N 3/0454
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
04Architectures, e.g. interconnection topology
0454using a combination of multiple neural nets
G06N 3/082
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
08Learning methods
082modifying the architecture, e.g. adding or deleting nodes or connections, pruning
G06N 3/084
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
08Learning methods
084Back-propagation
G10L 15/05
GPHYSICS
10MUSICAL INSTRUMENTS; ACOUSTICS
LSPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
15Speech recognition
04Segmentation; Word boundary detection
05Word boundary detection
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
Applicants
  • INTERNATIONAL BUSINESS MACHINES CORPORATION [US]/[US]
  • IBM UNITED KINGDOM LIMITED [GB]/[GB] (MG)
  • IBM (CHINA) INVESTMENT COMPANY LIMITED [CN]/[CN] (MG)
Inventors
  • AUDHKHASI, Kartik
  • SAON, George, Andrei
  • TUESKE, Zoltan
  • KINGSBURY, Brian
  • PICHENY, Michael, Alan
Agents
  • PYECROFT, Justine
Priority Data
16/551,91527.08.2019US
Publication Language English (EN)
Filing Language English (EN)
Designated States
Title
(EN) SOFT-FORGETTING FOR CONNECTIONIST TEMPORAL CLASSIFICATION BASED AUTOMATIC SPEECH RECOGNITION
(FR) OUBLI EN DOUCEUR POUR UNE RECONNAISSANCE AUTOMATIQUE DE LA PAROLE BASÉE SUR LA CLASSIFICATION TEMPORELLE CONNEXIONNISTE
Abstract
(EN)
In an approach to soft-forgetting training, one or more computer processors train a first model utilizing one or more training batches wherein each training batch of the one or more training batches comprises one or more blocks of information. The one or more computer processors, responsive to a completion of the training of the first model, initiate a training of a second model utilizing the one or more training batches. The one or more computer processors jitter a random block size for each block of information for each of the one or more training batches for the second model. The one or more computer processors unroll the second model over one or more non-overlapping contiguous jittered blocks of information. The one or more computer processors, responsive to the unrolling of the second model, reduce overfitting for the second model by applying twin regularization.
(FR)
Dans une approche d'apprentissage par oubli en douceur, un ou plusieurs processeurs informatiques entraînent un premier modèle en utilisant un ou plusieurs lots d'apprentissage, chaque lot d'apprentissage du ou des lots d'apprentissage comprenant un ou plusieurs blocs d'informations. Le ou les processeurs informatiques, en réponse à un achèvement de l'apprentissage du premier modèle, initient un apprentissage d'un second modèle en utilisant le ou les lots d'apprentissage. Le ou les processeurs informatiques rendent instable une taille de bloc aléatoire pour chaque bloc d'informations pour chacun du ou des lots d'apprentissage pour le second modèle. Le ou les processeurs informatiques déroulent le second modèle sur un ou plusieurs blocs d'informations rendus instables contigus ne se chevauchant pas. Le ou les processeurs informatiques, en réponse au déroulement du second modèle, réduisent le surajustement pour le second modèle par application d'une double régularisation.
Related patent documents
Latest bibliographic data on file with the International Bureau