NATURAL SPEECH TECHNOLOGY
http://www.natural-speech-technology.org
Centre for Speech Technology Research, University of Edinburgh Speech Research Group, University of Cambridge Speech and Hearing Group, University of Sheffield
Natural Speech Technology (NST) is an EPSRC Programme Grant, involving the Universities of Edinburgh, Cambridge and Sheffield. Its objective is to significantly advance the state-of-the-art in speech technology by making it more natural, approaching human levels of reliability, adaptability and conversational richness. NST starts in May 2011, and has a duration of 5 years.
8 research posts in speech recognition and speech synthesis are available on the project.
The overall aim of NST is to develop new speech technologies and theoretical frameworks which are adaptable, personalised, and expressive. These technologies will be closely linked to exemplar applications and the needs of the project User Group. We have four main technical objectives:
1. Learning and Adaptation: Models and algorithms for synthesis and recognition that can learn from continuous streams of data, can compactly represent and adapt to new scenarios and speaking styles, and seamlessly adapt to new situations and contexts almost instantaneously.
2. Natural Transcription: Speech recognisers that can detect “who spoke what, when, and how” in any acoustic environment and for any task domain.
3. Natural Synthesis: Controllable speech synthesisers that automatically learn from data, and are capable of generating the full expressive diversity of natural speech.
4. Exemplar Applications: Deployment of these advances in novel applications, with an emphasis on the health/social domain, personal listeners, and the needs of the User Group stakeholders.
The available positions are listed below, and at http://www.natural-speech-technology.org/jobs.html
To apply: For each position for which you wish to be considered, please apply to the host university using the corresponding web link in the list below (which also gives further information about the position). In addition, please also send an email including which jobs you are applying for and containing your CV as an attachment to:
Closing date: 10 June 2011
AVAILABLE POSITIONS:
1. University of Edinburgh. Science Manager in Speech Technology.
http://www.jobs.ed.ac.uk/vacancies/index.cfm?fuseaction=vacancies.detail&vacancy_ref=3014291
Reporting to the Programme Director, Prof Steve Renals, the Science Manager will ensure synchronisation, motivation and communications between the individual projects that make up the NST Programme Grant, and will also be responsible for developing and maintaining the interactions with the User Group. The Science Manager will work closely with lead researchers of each individual project, ensuring results are communicated across the consortium, and building liaisons with any User Group members for whom the project has relevance.
In addition, the Science Manager is expected to take a leading research role in one or more the individual projects in NST, working on statistical parametric speech synthesis and/or large vocabulary speech recognition.
The successful candidate should have or a PhD in speech processing, computer science, cognitive science, linguistics, engineering, mathematics, or a related discipline. He or she must have excellent programming skills, a background in statistical modelling using Hidden Markov Models, research experience in speech recognition and/or speech synthesis, and a strong publications record in international journals and conferences. He or she will also have experience and evidence of effective independent contribution to collaborative research teams. More broadly, he or she will have a demonstrated ability to lead, design and complete research projects, to solve problems independently and make original contributions to research.
Informal inquiries can be made by email to Prof Steve Renals (s.renals@ed.ac.uk) or Prof Simon King (Simon.King(at)ed.ac.uk)
This post is fixed-term for 2 years, with the possibility of extension.
2. University of Edinburgh. Postdoctoral Research Associate in Speech Synthesis.
http://www.jobs.ed.ac.uk/vacancies/index.cfm?fuseaction=vacancies.detail&vacancy_ref=3014313
This position is concerned with research in statistical parametric speech synthesis. The work will have a particular focus on the development of structured acoustic models which take account of factors such as accent and speaking style, and on the development of machine learning techniques for vocoding. You will have (or be near completion of) a PhD in speech processing, computer science, cognitive science, linguistics, engineering, mathematics, or a related discipline. You will have the necessary programming ability to conduct research in this area, a background in statistical modelling using Hidden Markov Models, speech signal processing, and research experience in speech synthesis.
A background in one or more of the following areas is also desirable: statistical parametric text-to-speech synthesis using HMMs and HSMMs; glottal source modelling; speech signal modelling; speaker adaptation using the MLLR or MAP family of techniques; familiarity with software tools including HTK, HTS, Festival; and familiarity with modern machine learning.
Informal inquiries can be made by email to Prof Steve Renals (s.renals@ed.ac.uk) or Prof Simon King (Simon.King(at)ed.ac.uk)
This post is fixed-term for 2 years, with the possibility of extension.
3. University of Edinburgh. Postdoctoral Research Associate in Speech Recognition.
http://www.jobs.ed.ac.uk/vacancies/index.cfm?fuseaction=vacancies.detail&vacancy_ref=3014315
This position is concerned with research in large vocabulary speech recognition, with a focus on models and algorithms that enable coverage of a wide set of domains. The work will have a particular focus on acoustic model structures factored in terms of acoustic environment, channel condition, band- width, speaker, and speaker style, and on factorised model building that can include supervised and unsupervised sources. You will have (or be near completion of) a PhD in speech processing, computer science, cognitive science, linguistics, engineering, mathematics, or a related discipline. You will have the necessary programming ability to conduct research in this area, a background in statistical modelling using Hidden Markov Models and research experience in speech recognition.
A background in one or more of the following areas is also desirable: subspace Gaussian mixture models; joint factor analysis; speaker adaptation using the MLLR or MAP family of techniques; familiarity with software tools including HTK and Kaldi; experience of the design, construction and evaluation of large vocabulary speech recognition systems; distant speech recognition; and multilingual acoustic modelling.
Informal inquiries can be made by email to Prof Steve Renals (s.renals@ed.ac.uk) or Prof Simon King (Simon.King(at)ed.ac.uk)
This post is fixed-term for 2 years, with the possibility of extension.
4. University of Sheffield. Senior Research Fellow in Speech Transcription.
http://www.jobs.ac.uk/job/ACP385/senior-research-fellow-in-natural-speech-technology-speech-transcription/
You will work on the Natural Transcription theme, addressing wide domain coverage, environment models and canonical acoustic models. Wide domain coverage in speech transcription is concerned with addressing the poor generalisation of recognition systems to new domains. Environment modelling means the automatic learning of acoustic scenarios for the benefit of improved far field recognition performance. Both of these rely on improved adaptability in recognition systems, addressed by the project on canonical acoustic modelling. You are required to work well in research teams and actively contribute to research advancement of the NST programme, the SpandH research group and the Department of Computer Science. All team members are expected to publish or contribute to publication in international conferences and journals at the forefront of the field.
Applicants should have a PhD (or have equivalent experience) in a related subject area. Solid knowledge of Unix type operating systems and programming in C/C++ is required. Applicants should have experience in one or more of the following areas: Acoustic modelling for automatic speech recognition, language modelling for automatic speech recognition, statistical pattern processing and the Hidden Markov Model Toolkit (HTK). Applicants are required to have an excellent track record in research of speech recognition and/or machine learning topics. Experience in research management is essential for this position as candidates are expected to take a leading role in site scientific management.
Informal inquiries can be made by email to Dr Thomas Hain (t.hain@dcs.shef.ac.uk) or Prof Phil Green (p.green(at)dcs.shef.ac.uk).
This post is fixed-term for 2 years, with the possibility of extension.
5. University of Sheffield. Research Associate in Speech Transcription
http://www.jobs.ac.uk/job/ACP392/research-associate-in-natural-speech-technology-speech-transcription/
You will be expected to work on NST projects on wide domain coverage, environment models and canonical acoustic models. Wide domain coverage in speech transcription is concerned with addressing the poor generalisation of recognition systems to new domains. Environment modelling means the automatic learning of acoustic scenarios for the benefit of improved far field recognition performance. Both of these rely on improved adaptability in recognition systems, addressed by the project on canonical acoustic modelling.
Applicants should have a PhD (or have equivalent experience) in a related subject area. Solid knowledge of Unix type operating systems and programming in C/C++ is required. Applicants should have experience in one or more of the following areas: Acoustic modelling for automatic speech recognition, language modelling for automatic speech recognition, statistical pattern processing and the Hidden Markov Model Toolkit (HTK). Applicants are required to have a good track record in research of speech recognition and/or machine learning topics demonstrated by publications in international journals and conferences. Candidates are required to work well in research teams and actively contribute to research advancement of the NST programme, the SpandH research group and the Department of Computer Science. All team members are expected to publish or contribute to publication in international conferences and journals at the forefront of the field.
Informal inquiries can be made by email to Dr Thomas Hain (t.hain@dcs.shef.ac.uk) or Prof Phil Green (p.green(at)dcs.shef.ac.uk).
This post is fixed-term for 2 years, with the possibility of extension.
6. University of Sheffield. Research Associate in Clinical Applications of Speech Technology.
http://www.jobs.ac.uk/job/ACP387/research-associate-in-natural-speech-technology-clinical-applications/
You will work on the NST project homeService, which will develop personalised, adaptive speech technology allowing users (who may be disabled and may have speech disorders) to interact with environmental control systems and home monitoring devices. The goals are to help people who cannot use conventional (keyboard/ mouse/screen) interfaces, people who prefer not to use such interfaces, in circumstances where conventional interfaces are impractical and for people who cannot communicate with others verbally, because their speech is disordered or they cannot speak at all. To meet these challenges our overarching goal is spoken language technology which adapts to the voice and the needs of an individual, a “personal adaptive listener”.
Applicants should have a PhD (or have equivalent experience) in a related subject area. Solid knowledge of Unix type operating systems and programming in C/C++ is required. Applicants should have experience in one or more of the following areas: Acoustic modelling for automatic speech recognition, language modelling for automatic speech recognition, statistical pattern processing and the Hidden Markov Model Toolkit (HTK). Applicants are required to have a track record in research of speech recognition, but must also demonstrate ‘user-facing’ skills, since homeService will be based on a longitudinal study in which NST technology will be deployed in user’s homes. Candidates are required to work well in research teams and actively contribute to research advancement of the NST programme, the SpandH research group and the Department of Computer Science. All team members are expected to publish or contribute to publication in international conferences and journals at the forefr!
ont of the field.
Informal inquiries can be made by email to Prof Phil Green (p.green@dcs.shef.ac.uk) or Dr Thomas Hain (t.hain(at)dcs.shef.ac.uk).
This post is fixed-term for 3 years, with the possibility of extension.
7. University of Cambridge. Two Research Associates in Speech Technology.
http://www.admin.cam.ac.uk/offices/hr/jobs/vacancies.cgi?job=8189&wantsFull=1&adv=1
Two positions for Research Associates will be available to work on NST at Cambridge. The five year programme is a collaboration between the University of Cambridge and the Universities of Edinburgh and Sheffield. It aims to significantly improve the state-of-the-art in large vocabulary speech recognition and speech synthesis. Appointments will be made for up to two years with the possibility of extension.
Applicants must have a very good first degree in a relevant discipline and would normally have a PhD degree in an area related to speech technology. It is expected that candidates will have a good knowledge of software tools including HTK and HTS. A good knowledge of C/C++ is required. In addition, experience in one or more of the following technical areas is necessary: machine learning; acoustic modelling techniques including methods for training and adaptation; HMM-based speech synthesis; language modelling for large vocabulary speech recognition.
Informal inquiries can be made by email to Prof Phil Woodland (pcw(at)eng.cam.ac.uk), Dr Mark Gales (mjfg(at)eng.cam.ac.uk), or Dr Bill Byrne (bill.byrne(at)eng.cam.ac.uk).
This post is fixed-term for 2 years, with the possibility of extension.
