Speech Technology Group Cambridge Research Lab Toshiba Research Europe Ltd
description
Transcript of Speech Technology Group Cambridge Research Lab Toshiba Research Europe Ltd
Copyright 2009, Toshiba Corporation.
12 January 2010
Speech Technology GroupCambridge Research LabToshiba Research Europe Ltd
Kate KnillManager, Interaction [email protected]
2
Toshiba
• World leader in high technology
• 3 key areas:– Digital media
– Electronic devices and components
– Social infrastructure systems
• 197,000 employees worldwide
• Sales over US$70billion
• Strong ecological commitment
3
Toshiba R&D: Toward the Innovation Driven Company
• Subline und Fliesstexte in Helvetica Neue 24 Light
Ein Aufzählungszeichen ist auch möglich
Toshiba Corporate R&D Center
Toshiba Corporate R&D Center
Toshiba China R&D CenterPeking
Toshiba China R&D CenterPeking
Toshiba Research Europe Limited
◆Cambridge Research Laboratory (CRL)
◆Telecommunications Research Laboratory Bristol
Toshiba Research Europe Limited
◆Cambridge Research Laboratory (CRL)
◆Telecommunications Research Laboratory Bristol
TARI Branch Officein Silicon ValleySan Jose
TARI Branch Officein Silicon ValleySan Jose
Toshiba America Research, Inc.Piscataway, New Jersey
Toshiba America Research, Inc.Piscataway, New Jersey
4
Toshiba Cambridge Research Lab
Established 1991 –
Semiconductor Physics for the 21st Century– Quantum Information
– Nano-biotechnology
Speech Technology Group added 2002
Computer Vision Group added 2006
5
Toshiba Speech and Language R&D
Toshiba China R&D, Beijing
Toshiba Corporate R&D Center, Kawasaki
Toshiba Research Europe Ltd, Cambridge
6
CRL Speech Technology Group
Toshiba China R&D, Beijing
Toshiba Corporate R&D Center, Kawasaki
• Focus on embedded ASR and TTS– Core technology research and development
• Noise and speaker robustness
• LVCSR
• HMM-TTS
– European and North American languages
• Approx 15 researchers– Multinational team
– Mix of engineers, computer scientists and linguists
7
Vision of Toshiba Speech Research
• Enhance the human-machine interface Interact with devices how, when and where you want
• Create a paradigm shift Input/output communication
8
Speech Recognition Challenges
Speaker Robustness Noise Robustness
Task Robustness
• Current ASR engines still suffer from lack of robustness– Major limitation in deploying speech recognition systems
9
Text-to-Speech Synthesis Challenges
• Increase in naturalness of synthesis– Same or even smaller footprint!
• Increase in voice variety– Faster, cheaper addition
– Non-professional voices
neutral friendly expressive emotional
large corpus professional
voice
small corpus professional
voice
small corpus amateur voices
10
Toshiba in SCALE: Second Supervisor• Recognition
– Kate Knill
– KK Chin
• Projects:– RS-3 Hierarchical Trajectory Models for Speech Recognition, Heyun
Huang, Lou Boves– AHSR-2 Data Association Multisource Acoustic Models, Liang Lu,
Steve Renals
• Synthesis– Heiga Zen
– Projects:• RS-1 Trajectory HMMs for Reactive Speech Synthesis, Cassia Valentini,
Simon King• RS-4 Speech Synthesis by Analysis, Mauro Nicalao, Roger Moore
11