[Elsnet-list] job announcement

Henk van den Heuvel H.vandenHeuvel at let.ru.nl
Sat May 20 00:46:55 CEST 2006


Would you be so kind as to put the job announcement below on the ELSNET
website?

Thanks!
Henk

------------------------------------------------------------------------
Henk van den Heuvel
------------------------------------------------------------------------
CLST / SPEX
Radboud University Nijmegen
Erasmusplein 1
NL-6525 HT Nijmegen
Netherlands
------------------------------------------------------------------------
Tel   : +31 24 3611686               Fax : +31 24 3612907
e-mail: H.vandenHeuvel at let.ru.nl
http://www.ru.nl/clst
http://www.spex.nl
------------------------------------------------------------------------


Job Announcement Radboud University Nijmegen

PhD candidate: Automatic speech recognition in the presence of noise (1.0
fte)
Vacancynumber: 23.85.06
Closing date: 4th June 2006

Job description
Robustness to noise in automatic speech recognition is essential for the
development of successful applications. Noise reduction techniques have
been applied with some success in the past, but there remains a large
performance gap between the best ASR implementations and human
recognition, especially when the noise is non-stationary.

The MIDAS project tackles the noise robustness problem in ASR through
missing data techniques (MDT). For a proper recognition it is not only
important to know how speech sounds map to written words. It is equally
important to know when external sound sources may have obscured some
characteristics of the speech sound, in other words when input data are
not related to speech.

The PhD student in Nijmegen will focus on the issue of how to detect that
specific spectro-temporal components of an incoming sound are not
speech-related. He/she will build detectors of missing data that
incorporate as much (language-specific) knowledge about speech as
possible, while making minimal assumptions about the noise.

Research area: Improving automatic speech recognition in noisy
circumstances.
The main task of the candidate will be to develop and improve techniques
in which knowledge about the structure of speech signals can be used to
separate speech from noise.

Requirements:
A successful candidate must have a strong background in informatics,
applied mathematics, signal processing or pattern recognition.

Organization
The Faculty of Arts consists of eleven departments in the area of language
and culture, history, history of arts,  linguistics and business
communication, who together cater to about 2,800 students and work closely
together in teaching and research.

The work of the PhD student is part of a collaborative STEVIN  project
entitled Missing Data Solutions (MIDAS). This project is carried out by a
consortium consisting of the ESAT Speech group (Katholieke Universiteit
Leuven), Center for Language and Speech Technology (Radboud Universiteit
Nijmegen) and Nuance Communication International.

Conditions of employment
Maximum salary per month, based on full-time employment: € 2.472.
The PhD students will be assigned to the function profile 'PhD Student'
(UFO system).
The starting gross salary is € 1.933 euro per month and will increase to €
2.472 in the fourth year.
Duration of contract: 1 year with a possibility of prolongation for
another 2,5 years, from 1st July 2006 or ASAP after this date.










Additional information
Informal enquiries should be made to dr. B. Cranen (B.Cranen at let.ru.nl) or
prof. dr. L. Boves (L.Boves at let.ru.nl)

An elaborate description of the project plan can be found at
http://lands.let.ru.nl/projects/midas_proposal.pdf

Application
Application (with vacancy number: 23.85.06), including curriculum vitae,
can be sent to:

Radboud Universiteit Nijmegen, Faculteit der Letteren,
Afdeling PZ, De heer B. Treep
Postbus 9103, 6500 HT Nijmegen
E-mail: pzlet at let.ru.nl





More information about the Elsnet-list mailing list