The ISCA board is always interested in improving its activities and the membership services it provides. To help us with this, could you please send us your ideas/comments/suggestions/impressions? We would be most grateful if you could take a moment to complete the form on the ISCA website : http://www.isca-speech.org/index.php and send us your feedback.

Your message will be sent to the ISCA secretariat : secretariat@isca-speech.org

Please enter ideas/comments/suggestions/impressions you may have on any new (or old) activities and membership services.

Please note: you can send us your comments anonymously, if you so wish.

Eva Hajicova - Membership Services 

 Emmanuelle Foxonet - ISCA Secretariat  for the ISCA board 

Thanks  to Ebru's efforts the online grant application system has been modified accordingly to Alan Black's suggestions and is now reachable for all isca-students.org registered users from the main menu and through the page:

http://www.isca-students.org/grants
 
You might want to link it from the relevant pages on IS08 and ISCA's websites.

Use it when registering for Interspeech 2008 (Brisbane)!

We remain at your disposal for any further need.

Marco A. Piccolino-Boniforti
ISCA-SAC web coordinator

map55@cam.ac.uk

Kindly be reminded that due to ISCA copyright issues, proceedings of the Interspeech conferences should not be put online for public access. The complete ISCA electronic archive, including over 100 proceedings of Interspeech (and former Eurospeech and ICSLP) conferences and ISCAworkshops is available to all members.   Thank you for your attention. 

Prof  Helen Meng,  ISCA publications.

         Dear Prof. Dr. Wellekens,

As you may already be aware, fraudulent e-mail solicitations for scholarly papers have been circulating which claim to originate from Elsevier, Inc. and are directed to prospective authors and editors. 

We are concerned about these emails and want to alert our community to this. We also want to protect our community as well as helping you to recognize fraudulent and/or phishing emails.

The fraudulent e-mail messages currently in circulation, generally contain "Manuscript Submission" or "Call for Papers" in the subject line and are typically sent using e-mail accounts supported by Gmail, Hotmail or by other free e-mail providers. Typically, the body of these messages contain a "Call for Papers," requesting that authors submit scholarly articles via e-mail for publication by Elsevier in various Elsevier journals and other publications.  Ultimately, these fraudulent e-mails involve a request for the victims to send "handling fees" to cover the processing of the article that has been submitted. 

Please be assured that Elsevier, Inc. is in no way associated with this fraudulent e-mail campaign.  Elsevier is currently investigating this fraud to identify the persons responsible and to bring them to justice.

In addition, please be advised that Elsevier does not solicit intellectual property from authors in this fashion, and does not utilize Gmail, Hotmail, or any other free third-party e-mail providers in communications with authors and editors. 

If you receive any e-mail messages that appear to be a part of this fraudulent solicitation, DO NOT respond to the message and do not open any attachments contained in the message.  Rather, please forward the message to Elsevier's Fraud Department at emailabuse@elsevier.com. We will use the information included in the message to aid in our investigation. If you know of someone who has received this message, please pass along the above information and ask them also to forward the message to the Elsevier's Fraud Department.

Thank you for your understanding and your cooperation. 

Kind regards,

Gail Rodney 
G.Rodney@elsevier.com

 AVISA has been involved in organizing the upcoming International Conference on Auditory-Visual Speech Processing (AVSP). This conference is to be held at the Tangalooma Wild Dolphin Resort on Moreton Island from the 26-29 September 2008. AVSP 2008 is a satellite conference to Interspeech 2008, being held in Brisbane from the 22-26 September 2008. Tangalooma is located at close distance from Brisbane, so that attendance at AVSP 2008 can easily be combined with participation in Interspeech 2008. This conference is bringing together speech scientists, cognitive psychologists, phoneticians, computational engineers and researchers in language learning studies from all around the world to talk about matters that incorporate the audio-visual speech production and perception by humans and machine alike.  Since the inaugural workshop in Bonas in 1995, Auditory-Visual Speech Processing workshops have been organized on a regular basis. In line with previous meetings, this conference will consist of a mixture of regular presentations (both posters and oral), and lectures by invited speakers. This year we Prof. Jeffrey F. Cohn (CMU), Prof Eric Vatikiotis-Bateson (UBC) and Dr Iain Matthews (Weta Digital) as our invited speakers.

INTERSPEECH 2008 incorporating SST 08

September 22-26, 2008

Brisbane Convention & Exhibition Centre

Brisbane, Australia

http://www.interspeech2008.org/

INTERSPEECH ’08 PROGRAM RELEASE – REGISTER NOW, AVOID LATE FEES

1.      Register Now for INTERSPEECH ‘08 in Brisbane, Australia

INTERSPEECH ’08 is being held in sunny Brisbane, Australia, September 22 – 26.  The end of the standard registration is fast approaching (August 31).  To allow for international time zones please register before midnight (UTC) on August 30.  To register please click on the following to register - http://www.interspeech2008.org/registration.html

2.      Program

The varied and comprehensive INTERSPEECH ’08 program including all conference sessions, tutorials, social events and a schedule of poster sessions is appended below (see also http://www.interspeech2008.org/docs/IS2008_Program_at_a_glance.pdf)

3.      But Wait, there’s More!

In addition to the scientific program a number of tours to the Australia Zoo, the Lone Pine Koala Park (photos with Koalas included) etc are being organised for delegates and accompanying persons.  Visit http://www.interspeech2008.org and review the Tours and Visitor Info section.  This provides a host of details on airport transfers, the venues, visa and quarantine regulations.

 Interspeech is the world's largest and most comprehensive conference on Speech

Science and Speech Technology. We invite original papers in any related area,

including (but not limited to):

Human Speech Production, Perception and Communication;

Speech and Language Technology;

Spoken Language Systems;

Applications, Resources, Standardisation and Evaluation

Important Dates

Tutorial Day: Monday, 22 September 2008

Main conference: 23-26 September 2008

You will need a VISA

All INTERSPEECH international delegates (with the sole exception of citizens of New Zealand travelling on a New Zealand passport) will need a visa to enter Australia and this must be applied for in advance. Delegates from many countries can apply for an electronic visa, called an ETA (Electronic Travel Authority) for a modest fee of $20 and your application is likely to be approved instantly.  Currently citizens of the following countries can apply electronically at:  www.eta.immi.gov.au

Andorra, Austria , Belgium , Brunei , Canada , Denmark ,France, Finland, Germany , Greece, Hong Kong (SAR), Iceland, Ireland ,  Italy, Japan, Liechtenstein , Luxembourg , Malaysia , Malta, Monaco, Netherlands ,Norway , Portugal , San Marino , Singapore , South Korea, Spain,  Sweden , Switzerland,    Taiwan * UK **,  USA,  Vatican City. 

 If your country is not listed above, you will not be able to apply using the ETA system.  You will need to apply for a visa (a 456 Business Short Stay visa) at an Australian overseas office – for a list of offices go to: http://www.immi.gov.au/contacts/overseas/index.htm).  It is likely that you will be required to provide an INTERSPEECH conference registration confirmation letter with your visa application.  If you do not have one of these already, please write to megan@ccm.com.au and ask for a letter t be sent. The Australian Government’s Department of Immigration and Citizenship website provides full details on visas – please visit:  http://www.immi.gov.au/.

INTERSPEECH is being held in September, which is spring time in Australia! It’s a great time to visit sunny Brisbane and all of the attractions for which Australia is famous:  the Great Barrier Reef, the Gold and Sunshine Coasts, the Outback – they are all at their very best during spring.

See the INTERSPECH 2008 website   www.interspeech2008.org  Information about Brisbane and Australia and links to useful resources.

Start your travels to Australia right here: www.australia.com

BOOK YOUR INTERSPEECH ACCOMMODATION

INTERSPEECH 2008 has negotiated special discount accommodation rates at a variety of hotels and apartment hotels in Brisbane and you can make your booking when you complete the INTERSPEECH registration form.  Be warned, September is a busy time of year and we urge you to book your accommodation as soon as possible.

For a list of accommodation options, please go to the “Accommodation” sidebar on the INTERSPEECH 2008 website:  www.interspeech2008.org .

Chairman: Denis Burnham, MARCS, University of West Sydney. 

September 6-10, 2009, Brighton, UK,
Conference Website
Chairman: Prof. Roger Moore, University of Sheffield.

Chiba, Japan
Conference Website
ISCA is pleased to announce that INTERSPEECH 2010 will take place in Makuhari-Messe, Chiba, Japan, September 26-30, 2010. The event will be chaired by Keikichi Hirose (Univ. Tokyo), and will have as a theme "Towards Spoken Language Processing for All - Regardless of Age, Health Conditions, Native Languages, Environment, etc."

Interspeech 2011

Palazzo dei Congressi,  Italy, August 27-31, 2011.

Organizing committee

Piero Cosi (General Chair),

Renato di Mori (General Co-Chair),

Claudia Manfredi (Local Chair),

Roberto Pieraccini (Technical Program Chair),

Maurizio Omologo (Tutorials),

Giuseppe Riccardi (Plenary Sessions).

More information www.interspeech2011.org

 This section shows recent books whose titles been have communicated by the authors or editors.

Also some advertisement for recent books in speech are included.

Book presentation is written by the authors and not by this newsletter editor or any  voluntary reviewer.

Music and speech share numerous aspects (language, structural, acoustics, cognitive), as long in their production, that in their representation and their perception. This list has for object to warn its users, various events dealing with the study of the links between music and speech. It thus intends to connect several communities, their allowing each to take advantage of a stimulating interaction.

As a member of the speech or music community, you are invited to
subscribe to musicspeech group. The group will be moderated and
maintained by IRCAM.

Group details:
* Name: musicspeech
* Home page: http://listes.ircam.fr/wws/info/musicspeech
* Email address: musicspeech@ircam.fr

Greg Beller, IRCAM,
moderator, musicspeech list

AT&T - Labs Research is seeking exceptional candidates for Research Staff positions. AT&T is the premiere broadband, IP, entertainment, and wireless communications company in the U.S. and one of the largest in the world. Our researchers are dedicated to solving real problems in speech and language processing, and are involved in inventing, creating and deploying innovative services. We also explore fundamental research problems in these areas. Outstanding Ph.D.-level candidates at all levels of experience are encouraged to apply. Candidates must demonstrate excellence in research, a collaborative spirit and strong communication and software skills. Areas of particular interest are               

• Large-vocabulary automatic speech recognition
• Acoustic and language modeling
• Robust speech recognition
• Signal processing
• Speaker recognition
• Speech data mining
• Natural language understanding and dialog
• Text and web mining
• Voice and multimodal search

AT&T Companies are Equal Opportunity Employers. All qualified candidates will receive full and fair consideration for employment. More information and application instructions are available on our website at http://www.research.att.com/. Click on "Join us". For more information, contact Mazin Gilbert (mazin at research dot att dot com).

Nagoya Institute of Technology is seeking a researcher for a

post-doctoral position in a new European Commission-funded project

EMIME ("Efficient multilingual interaction in mobile environment")

involving Nagoya Institute of Technology and other five European

partners, starting in March 2008 (see the project summary below).

The earliest starting date of the position is March 2007. The initial

duration of the contract will be one year, with a possibility for

prolongation (year-by-year basis, maximum of three years). The

position provides opportunities to collaborate with other researchers

in a variety of national and international projects. The competitive

salary is calculated according to qualifications based on NIT scales.

The candidate should have a strong background in speech signal

processing and some experience with speech synthesis and recognition.

Desired skills include familiarity with latest spectrum of technology

including HTK, HTS, and Festival at the source code level.

For more information, please contact Keiichi Tokuda

(http://www.sp.nitech.ac.jp/~tokuda/).

About us

Nagoya Institute of Technology (NIT), founded on 1905, is situated in

the world-quality manufacturing area of Central Japan (about one hour

and 40 minetes from Tokyo, and 36 minites from Kyoto by Shinkansen).

NIT is a highest-level educational institution of technology and is

one of the leaders of such institutions in Japan. EMIME will be

carried at the Speech Processing Laboratory (SPL) in the Department of

Computer Science and Engineering of NIT. SPL is known for its

outstanding, continuous contribution of developing high-performance,

high-quality opensource software: the HMM-based Speech Synthesis

System "HTS" (http://hts.sp.nitech.ac.jp/), the large vocabulary

continuous speech recognition engine "Julius"

(http://julius.sourceforge.jp/), and the Speech Signal Processing

Toolkit "SPTK" (http://sp-tk.sourceforge.net/). The laboratory is

involved in numerous national and international collaborative

projects. SPL also has close partnerships with many industrial

companies, in order to transfer its research into commercial

applications, including Toyota, Nissan, Panasonic, Brother Inc.,

Funai, Asahi-Kasei, ATR.

Project summary of EMIME

The EMIME project will help to overcome the language barrier by

developing a mobile device that performs personalized speech-to-speech

translation, such that a user's spoken input in one language is used

to produce spoken output in another language, while continuing to

sound like the user's voice. Personalization of systems for

cross-lingual spoken communication is an important, but little

explored, topic. It is essential for providing more natural

interaction and making the computing device a less obtrusive element

when assisting human-human interactions.

We will build on recent developments in speech synthesis using hidden

Markov models, which is the same technology used for automatic speech

recognition. Using a common statistical modeling framework for

automatic speech recognition and speech synthesis will enable the use

of common techniques for adaptation and multilinguality.

Significant progress will be made towards a unified approach for

speech recognition and speech synthesis: this is a very powerful

concept, and will open up many new areas of research. In this

project, we will explore the use of speaker adaptation across

languages so that, by performing automatic speech recognition, we can

learn the characteristics of an individual speaker, and then use those

characteristics when producing output speech in another language.

Our objectives are to:

1. Personalize speech processing systems by learning individual

characteristics of a user's speech and reproducing them in

synthesized speech.

2. Introduce a cross-lingual capability such that personal

characteristics can be reproduced in a second language not spoken

by the user.

3. Develop and better understand the mathematical and theoretical

relationship between speech recognition and synthesis.

4. Eliminate the need for human intervention in the process of

cross-lingual personalization.

5. Evaluate our research against state-of-the art techniques and in a

practical mobile application.

Title: Solutions for non-linear acoustic echo.

Background:

Acoustic echo is an annoying disturbance due to the sound feedback between the loudspeaker and the microphone of terminals. Acoustic echo canceller and residual echo cancellation are widely used to reduce the echo signal. The performance of existing echo reduction systems strongly relies on the assumption that the echo path between transducers is linear. However, today’s competitive audio consumer market may favour sacrificing linear performance for the integration of low cost analogue components. The assumption of linearity is not hold anymore, due to the nonlinear distortions introduced by the loudspeakers and the small housing where transducers are placed.

Task:

The PhD thesis will lead research in the field of non-linear system applied for acoustic echo reduction. The foreseen tasks deal first with proper modelling of mobile phone transducers presenting non-linearity, to get a better understanding in which environment echo reduction works. Using this model as a basis, study of performance of linear system will permit to get a good understanding on the problems brought by non-linearity. In a further step, the PhD student will develop and test non-linear algorithms coping with echo cancellation in non-linear environment.

About the Company:

Sophia-Antipolis site is one of the main Infineon Technologies research and development centers worldwide. Located in the high-tech valley of Sophia-Antipolis, near Nice in the south of France, a team of 140 experienced research and development engineers specialized in Mobile Solutions, Embedded SRAM, and Design-Flow Software. The PhD will take place within the Mobile Solution group which is responsible for specifying and designing baseband integrated circuits for cellular phones. The team is specialized in innovative algorithm development, especially in audio, system specification and validation, circuit design and embedded software. Its work makes a significant contribution to the Infineon Technologies wireless chipset portfolio.

Required skills:

-        Master degree

-        Strong background in signal processing.

-        Background in speech signal or non-linear system processing is a plus.

-        Programming: Matlab, C.

-        Knowledge in C-fixed point / DSP implementation is a plus.

-        Language: English

Length of the PhD: 3 years

Place: Infineon Technologies France, Sophia-Antipolis

Contact:

Christophe Beaugeant

Phone: +33 (0)4 92 38 36 30

E-mail : christophe.beaugeant@infineon.com

At the Institut für Theoretische Informatik, Lehrstuhl Prof. Waibel Universität Karlsruhe (TH) a

Ph.D. position

in the field of

Software System Integration of Automatic Speech Recognition and Machine Translation for Speech based Multimedia Indexing

has to be filled immediately with a salary according to TV-L, E13.

The responsibilities include integration, fusion and development of core technologies in the area of automatic speech recognition, simultaneous machine translation, in the context of speech based indexing of multimedia documents within application targeted research projects in the area of multimodal Human-machine interaction.  Set in a framework of internationally and industry funded research programs, the successful candidate is expected to contribute to showcases for state-of-the art of modern recognition and translation systems.

We are an internationally renowned research group with an excellent infrastructure. Examples of our projects for improving Human-Machine and Human-to-Human interaction are: JANUS - one of the first speech translation systems proposed, simultaneous translation of lectures, portable speech translators, meeting browser and lecture tracker.

Within the framework of the International Center for Advanced Communication Technology (interACT), our institute operates in two locations, Universität Karlsruhe (TH), Germany and at Carnegie Mellon University, Pittsburgh, USA.  International joint and collaborative research at and between our centers is common and encouraged, and offers great international exposure and activity. 

Applicants are expected to have:

• an excellent university degree (M.S, Diploma or Ph.D.) in Computer Science, Electrical Engineering, Mathematics, or related fields
• excellent programming skills 
• advanced knowledge in at least one of the fields of Machine Learning, Pattern Recognition, Statistics, or System Integration

For candidates with Bachelor or Master’s degrees, the position offers the opportunity to work toward a Ph.D. degree.

In line with the university's policy of equal opportunities, applications from qualified women are particularly encouraged. Handicapped applicants will be preferred in case of the same qualification.

Questions may be directed to: Sebastian Stüker, Tel. +49 721 608 6284, E-Mail: stueker@ira.uka.de,  http://isl.ira.uka.de

The application should be sent to Professor Waibel, Institut für Theoretische Informatik, Universität Karlsruhe (TH), Adenauerring 4, 76131 Karlsruhe, Germany

----------------------------------------------------------------------------------------------------------------------------------

Ph.D. position

in the field of

Multimodal Dialog Systems

Is to be filled immediately with a salary according to TV-L, E13.

The responsibilities include basic research in the area of multimodal dialog systems, especially multimodal human-robot interaction and learning robots, within application targeted research projects in the area of multimodal Human-machine interaction.  Set in a framework of internationally and industry funded research programs, the successful candidate(s) are expected to contribute to the state-of-the art of modern spoken dialog systems, improving natural interaction with robots.

We are an internationally renowned research group with an excellent infrastructure. Current research projects for improving Human-Machine and Human-to-Human interaction are focus on dialog management for Human-Robot interaction.

Within the framework of the International Center for Advanced Communication Technology (interACT), our institute operates in two locations, Universität Karlsruhe (TH), Germany and at Carnegie Mellon University, Pittsburgh, USA.  International joint and collaborative research at and between our centers is common and encouraged, and offers great international exposure and activity. 

Applicants are expected to have:

• an excellent university degree (M.S, Diploma or Ph.D.) in Computer Science, Computational Linguistics, or related fields
• excellent programming skills 
• advanced knowledge in at least one of the fields of Speech and Language Processing, Pattern Recognition, or Machine Learning

For candidates with Bachelor or Master’s degrees, the position offers the opportunity to work toward a Ph.D. degree.

In line with the university's policy of equal opportunities, applications from qualified women are particularly encouraged. Handicapped applicants will be preferred in case of the same qualification.

Questions may be directed to: Hartwig Holzapfel, Tel. +49 721 608 4057, E-Mail: hartwig@ira.uka.de,  http://isl.ira.uka.de

The application should be sent to Professor Waibel, Institut für Theoretische Informatik, Universität Karlsruhe (TH), Adenauerring 4, 76131 Karlsruhe, Germany

Job opening at TFH Berlin University of Applied Sciences, Department of Computer Sciences and Media, Germany: Post-graduate position (part-time) for a computer scientist or engineer with a background in ASR and/or TTS in a three-year project in Computer-Aided Language Learning funded by the German Ministry of Education and Research. Start: 1 July 2008. The task will be the development and evaluation of a software system for teaching Mandarin pronunciation to Germans, as well as administrative duties with the funding body. Knowledge of E-Learning applications, German and/or Mandarin are welcome, good English skills mandatory. Candidates will have the opportunity to pursue a PhD degree and should be preferably EU citizens. The position is paid according to BAT 2a/2 (German pay scale for federal employees), about €28.000/year depending on age and marital status. Please direct further enquiries to Prof. Dr. Hansjörg Mixdorff at mixdorff@tfh-berlin.de. 

Offre d’Allocation de Recherche – Rentrée Universitaire 2008

Les cartes sensorimotrices de la parole: Corrélats neuroanatomiques des systèmes de perception et de production des voyelles et consonnes du Français.

Marc Sato, Chargé de Recherche CNRS

Jean-Luc Schwartz, Directeur de Recherche CNRS

GIPSA-Lab, UMR CNRS 5216, Département Parole et Cognition, Equipe "Perception, Multimodalité, Développement", Grenoble France (http://gipsa-lab.inpg.fr).

Pour la rentrée universitaire 2008, nous proposons une Allocation Doctorale de Recherche dans le cadre de l’Ecole Doctorale Ingénierie pour la santé, la Cognition et l’Environnement, habilitée par les universités Pierre Mendès France, Joseph Fourier et l’Institut National Polytechnique de Grenoble (EDISCE – ED216, http://www-sante.ujf-grenoble.fr/edisce/).

Dans le cadre théorique d’un possible couplage fonctionnel entre systèmes de perception et de production de la parole, ce projet de recherche a pour but de tester l’existence de connectivités dynamiques fonctionnelles spécifiques entre régions sensorielles et motrices lors de la perception et de la production des voyelles et consonnes du Français. Dans ce but, une série d’expériences en imagerie par résonance magnétique fonctionnelle (IRMf) et en électro-encéphalographie (EEG) devra permettre une description spatiale et temporelle précise des activités cérébrales impliquées lors de la production et la perception des phonèmes du français ainsi que de la connectivité dynamique entre ces régions. Les expériences menées devraient permettre une compréhension approfondie des processus d’analyse et de construction des représentations verbales par la mise en évidence d’une co-structuration et dépendance des régions sensorielles et motrices.

Outre une recherche bibliographique approfondie du corpus de la littérature, multidisciplinaire, dans les domaines de la phonétique et de la phonologie, de la neuropsycholinguistique et des neurosciences cognitives, ce projet comprendra l’élaboration de protocoles expérimentaux en IRMf et EEG, la passation des sujets et le recueil des données, enfin l’analyse statistique des données et leur interprétation.

Le(a) candidat(e) aura de préférence un M2R en neurosciences, sciences cognitives, psychologie cognitive ou neuropsychologie. Le candidat devra être familier avec l’expérimentation comportementale classique ainsi que les tests et analyses statistiques appliqués en psychologie cognitive. La pratique de l’anglais et une première expérience avec les techniques d’IRMf et/ou d’EEG est souhaitable.

Ce travail de thèse sera inscrit dans le contexte d’un projet de recherche grenoblois portant sur la mise en œuvre de nouvelles techniques d’analyses et de modélisation non-linéaire de mesure de la connectivité cérébrale. Il s’appuiera également sur des collaborations nationales et internationales, notamment avec le Centre de Recherche sur le Langage, l’Esprit et le Cerveau de l’Université McGill (Canada). Le financement est pour une période de trois ans, à compter d’Octobre 2008.

Contact pour cette annonce:

Marc Sato

GIPSA-Lab, UMR CNRS 5216

Université Stendhal

BP 25 - 38040 Grenoble cedex 9

Tel: (+33) (0)476 827 784

Fax: (+33) (0)476 824 335

E-mail: marc.sato [at] gipsa-lab.inpg.fr

La date limite pour envoyer un CV détaillé est fixée au 18 juin. Dès que possible, envoyez également les notes finales et classement au M2R, et éventuellement une lettre de recommandation.

Thèse : Caractérisation et l'identification automatique de dialectes These DeadLine: 10/06/2008 jerome.farinas@irit.fr http://www.irit.fr/recherches/SAMOVA/these-caracterisation-et-lidentification-automatique-de-dialectes.html Description du sujet Le recherche en traitement automatique de la parole s'intéresse de plus en plus au traitement de grandes collections de données, dans des conditions de parole spontanée et conversationnelle. Les performances sont dépendantes de toutes les variabilités de la parole. Une de ces variabilités concerne l'appartenance dialectale du locuteur, qui induit de la variabilité tant au niveau de la prononciation phonétique, mais également au niveau de la morphologie des mots et de la prosodie. Nous proposons de réaliser un sujet de recherche sur la caractérisation automatique dialectale des locuteurs, en vue de diriger l'adaptation des systèmes de reconnaissance de la parole : la sélection de modèles acoustiques et prosodiques adaptées permettront d'améliorer des performances dans des conditions de reconnaissance indépendante du locuteur. La réalisation d'un tel système devra s'appuyer sur les avancées récentes en identification de la langue au niveau de la modélisation acoustique par exp! loration de réseaux de phonèmes et proposer une modélisation fine basée sur la micro et macro prosodie. Les bases de données disponibles au sein du projet sur la phonologie du français contemporain (http://www.projet-pfc.net/) permettront de disposer d'un large éventail de données sur les variances de prononciation. Le système final sera évalué lors des campagnes internationales organisées par le NIST sur la vérification de la langue, qui prennent maintenant en compte les variances dialectales (mandarin, anglais, espagnol et hindi) : http://www.nist.gov/speech/tests/lre/. English version The research in automatic speech processing is increasingly concerned at the treatment of large data collections, with spontaneous and conversational speech. The variability of speech alter the general performances. The dialect of the speaker is of these variability sources, who leads in alterations in terms of both the phonetic pronunciation, but also in terms of the morphology of words and prosody. We propose to conduct a research on the automatic characterization of the dialects, in order to adapt automatic speech recognition systems: by selection of acoustic and prosodic models suited to improve performance in speaker independent recognition conditions. The realization of such a system should be based on recent advances in the identification of the language in the exploration of phonemes modeling lattices and propose a fine modelling based on micro and macroprosody. The databases available within the PFC project (http://www.projet-pfc.net/) will provide a wide range of d! ata on variances pronunciation. The final system will be evaluated during international campaigns conducted by the NIST on language verification, which now take into account the dialect identification (Mandarin, English, Spanish and Hindi): http://www.nist.gov/speech/tests/lre/. Connaissances et compétences requises * compétences en informatique (en particulier traitement automatique de la parole) * compétences en linguistique (phonologie, prosodie) Contact Un financement sera attribué aux meilleurs candidats de thèse de l'Ecole Doctorale, il faut donc nous contacter avant le 10 juin pour pouvoir participer à ce classement.

Thèse de l’école doctorale MITT, Université Paul Sabatier Toulouse III

DeadLine: 10/06/2008

Contacts: jerome.farinas@irit.fr

http://www.irit.fr/-Equipe-SAMoVA-

DESCRIPTION DU SUJET :

Le recherche en traitement automatique de la parole s'intéresse de plus en plus au traitement de grandes collections de données, dans des conditions de parole spontanée et conversationnelle. Les performances sont dépendantes de toutes les variabilités de la parole. Une de ces variabilités concerne l'appartenance dialectale du locuteur, qui induit de la variabilité tant au niveau de la prononciation phonétique, mais également au niveau de la morphologie des mots et de la prosodie. Nous proposons de réaliser un sujet de recherche sur la caractérisation automatique dialectale des locuteurs, en vue de diriger l'adaptation des systèmes de reconnaissance de la parole : la sélection de modèles acoustiques et prosodiques adaptées permettront d'améliorer des performances dans des conditions de reconnaissance indépendante du locuteur. La réalisation d'un tel système devra s'appuyer sur les avancées récentes en identification de la langue au niveau de la modélisation acoustique par exploration de réseaux de phonèmes et proposer une modélisation fine basée sur la micro et macro prosodie. Les bases de données disponibles au sein du projet sur la phonologie du français contemporain (http://www.projet-pfc.net/) permettront de disposer d'un large éventail de données sur les variances de prononciation. Le système final sera évalué lors des campagnes internationales organisées par le NIST sur la vérification de la langue, qui prennent maintenant en compte les variances dialectales (mandarin, anglais, espagnol et hindi) : http://www.nist.gov/speech/tests/lre/.

ENGLISH VERSION:

The research in automatic speech processing is increasingly concerned at the treatment of large data collections, with spontaneous and conversational speech. The variability of speech alter the general performances. The dialect of the speaker is of these variability sources, who leads in alterations in terms of both the phonetic pronunciation, but also in terms of the morphology of words and prosody. We propose to conduct a research on the automatic characterization of the dialects, in order to adapt automatic speech recognition systems: by selection of acoustic and prosodic models suited to improve performance in speaker independent recognition conditions. The realization of such a system should be based on recent advances in the identification of the language in the exploration of phonemes modeling lattices and propose a fine modelling based on micro and macroprosody. The databases available within the PFC project (http://www.projet-pfc.net/) will provide a wide range of data on variances pronunciation. The final system will be evaluated during international campaigns conducted by the NIST on language verification, which now take into account the dialect identification (Mandarin, English, Spanish and Hindi): http://www.nist.gov/speech/tests/lre/.

CONNAISSANCES ET COMPETENCES REQUISES

• compétences en informatique (en particulier traitement automatique de la parole)

• compétences en linguistique (phonologie, prosodie)

Cambridge University: Research Position in Speech Synthesis and Recognition / Machine Translation


A position exists for a Research Associate to work on the EMIME ("Efficient multilingual interaction in mobile environment") project. This project is funded by the European Commission within the FP7 programme. The project aims to develop a mobile device that performs personalized speech-to-speech translation such that a user's spoken input in one language is used to produce spoken output in another language, while continuing to sound like the user's voice. We will build on recent developments in speech synthesis using hidden Markov models, which is the same technology used for automatic speech recognition. Using a common statistical modelling framework for automatic speech recognition and speech synthesis will enable the use of common techniques for adaptation and multilinguality. The project objectives are to

1. Personalise speech processing systems by learning individual characteristics of a user's speech and reproducing them in synthesised speech.
2. Introduce a cross-lingual capability such that personal characteristics can be reproduced in a second language not spoken by the user.
3. Develop and better understand the mathematical and theoretical relationship between speech recognition and synthesis.
4. Eliminate the need for human intervention in the process of cross-lingual personalisation.
5. Evaluate our research against state-of-the art techniques and in a practical mobile application.
See the EMIME website for more information: http://www.emime.org/

This is an opportunity to work in a research group with a world-leading reputation in speech recognition and statistical machine translation research. There are excellent opportunities for publications, travel and conference visits. The group has outstanding research facilities. For suitably qualified candidates there may also be the chance to contribute to the MPhil in Computer Speech, Text and Internet Technology (http://mi.eng.cam.ac.uk/cstit/).

The successful candidate must have a very good first degree in a relevant discipline and preferably have a higher degree as well as experience in acoustic modeling for speech synthesis and/or recognition. Expertise in one or more of the following technical areas is also a distinct advantage:
- speech recognition with the HTK toolkit (http://htk.eng.cam.ac.uk)
- speech synthesis with the HTS HMM-based Speech Synthesis System (http://hts.sp.nitech.ac.jp)
- weighted finite state transducers for speech and language processing
The project focus is acoustic modeling but experience in statistical machine translation is also an advantage.

The cover sheet for applications, PD18 is available from http://www.admin.cam.ac.uk/offices/personnel/forms/pd18/ Part I and Part III only, should be sent, with a letter and CV to Dr Bill Byrne, Department of Engineering, Trumpington Street, Cambridge, CB2 1PZ, (Fax +44 01223 332662, email wjb31@cam.ac.uk).
Quote Reference: NA03547, Closing Date: 30 June 2008

The University values diversity and is committed to equality of opportunity.

Head of NLP :

 We are now looking for a very experienced Computational Linguist

 to lead our efforts in the natural language processing area. This

 is a hugely challenging, but also a very rewarding role; the

 opportunities for applying linguistic techniques are virtually

 limitless and even small improvements in the algorithms for

 detecting and correcting potential conversion errors translate

 into serious cost savings for the company. This is a senior position

 leading a team and having the autonomy to build a strategic way

 forward for this department.

 Experience Needed:

 Grammars and parsing for spontaneous speech.

 Statistical methods.

 At least basic programming ability (shell scripts, Perl, awk).

 Spell-checkers, grammar checkers, auto correcting tools and predictive typing.

 Experience with Automatic Speech Recognition technology.

 Probabilistic Language Modelling

 Phonetics

 Multi-lingual

info@voxid.co.uk 

Nuance is the leading provider of speech and imaging solutions for businesses and consumers around the world.  Every day, millions of users and thousands of businesses experience Nuance by calling directory assistance, requesting account information, dictating patient records, telling a navigation system their destination, controlling their mobile phone or digitally reproducing documents that can be shared and searched.  With more than 2000 employees worldwide, we are committed to make the user experience more enjoyable by transforming the way people interact with information and how they create, share and use documents. Making each of those experiences productive and compelling is what Nuance is about. To strengthen our international team we are currently looking for a

Junior Research Engineer for Embedded Automatic Speech Recognition

Work Environment

·         You will work in the Embedded ASR research and production team in Merelbeke, Belgium or Aachen, Germany, working with state-of -he-art speech technology, tools and runtime software. Both Gent and Aachen are nice, historical European university cities.

·         You will work in an international company and cooperate with people and research teams on various locations across the globe. You may occasionally be asked to travel.

·         You will work  with our natural language understanding and dialogue research teams as well support our professional services teams.

·         You will work on the development of cutting edge speech recognition products for automotive platforms and mobile devices. You will help the engine cope with multi-lingual speech in various noise conditions, and this while respecting strong limitations on the usage of memory and processing power.

Key Responsibilities

·         Design, implementation, evaluation, optimization and testing of new algorithms and tools, with a strong focus on speech signal processing and acoustic modeling in adverse, noisy environments.

·         Activities are targeted at the creation of commercial products for resource limited platforms.

·         Focus on creating efficient production and development processes to bring the technology to marketable products in a wide range of languages.

·         Occasional application of the developed algorithms and tools for producing systems for a specific language.

·         Specification and follow-up of projects to make the system work with third party components, such as beam formers, echo cancellers or content data providers.

Your Profile

• You have a University degree in engineering, mathematics or physics.
• A PhD degree in speech processing or equivalent relevant experience is a strong asset.
• Experience in speech recognition research, especially acoustic modeling or signal processing, is required.
• Experience in speech processing, machine learning techniques or statistical modeling is required.
• Knowledge about small platforms and experience in developing software for them is a plus.
• Strong software skills are required, especially C/C++ and a scripting language like Perl or Python in a Linux/Unix environment. Knowledge of Matlab is a plus.
• Additional background in computational linguistics is a plus.
• You are a team player, willing to take initiative, and are goal oriented.
• You have a strong desire to make things “really work” in practice, on hardware platforms with limited memory and processing power.
• You are fluent in English and at least one other language, and you can write high quality English documentation.  

Interested?

Please send your CV to Deanna Roe at deanna.roe@nuance.com. If you have any questions, please contact her at +44 207 922 5757.

We are looking forward to receiving your application!

In order to strengthen our Embedded ASR Research team, we are looking for a:

       SOFTWARE ENGINEER SPEECH DIALOGUE TOOLS

As part of our team, you will be creating solutions for voice user interfaces for embedded applications on mobile and automotive platforms.

OVERVIEW:

- You will work in Nuance's Embedded ASR (automatic speech recognition) research and development team, developing technology, tools, and run-time software to enable our customers to develop and test embedded speech applications. Together with our team of speech and language experts, you will work on natural language dialogue systems for our customers in the Automotive and Mobile sector.

- You will work on fascinating technology that has now reached the maturity to enable new generations of powerful and natural user interfaces. Your code is crucial to the research in speech and language technology that defines the state of the art in this field It is equally important for the products that you will find in the market, in speech-enabled cars, navigation devices, and cell phones.

- You will work in a large international software company that is the leading provider of speech and imaging solutions for businesses and consumers around the world. You will cooperate with people on various locations including in Europe, America and Asia. You may occasionally be asked to travel.

RESPONSIBILITIES:

- You will work on the development of tools and solutions for cutting edge speech and language understanding technologies for automotive and mobile devices.

- You will work on enhancing various aspects of our advanced natural language dialogue system, such as the layer of connected applications, the configuration setup, inter-module communication, etc.

- In particular, you will be responsible for the design, implementation, evaluation, optimization and testing, and documentation of tools such as GUI and XML applications that are used to develop, configure, and fine-tune advanced dialogue systems.

QUALIFICATIONS:

- You have a university degree in computer science, engineering, mathematics, physics, computational linguistics, or a related field.

- You have very strong software and programming skills, especially in C/C++, ideally also for embedded applications.

- You have experience with Python or other scripting languages.

- GUI programming experience is an asset.

The following skills are a plus:

- Understanding of communication protocols

- Understanding of databases

- A background in (computational) linguistics, dialogue systems, speech processing, grammars, and parsing techniques, statistics, pattern recognition, and machine learning, especially as related to natural language processing, dialogue, and representation of information

- Understanding of computational agents and related frameworks (such as OAA).

- You can work both as a team player and as goal-oriented independent software engineer.

- You can work in a multi-national team and communicate effectively with people of different cultures.

- You have a strong desire to make things really work in practice, on hardware platforms with limited memory and processing power.

- You are fluent in English and you can write high quality documentation.

- Knowledge of other languages is a plus.

CONTACT:

Please send your applications, including cover letter, CV, and related documents (maximum 5MB total for all documents, please) to

Benjamin Campued       Benjamin.Campued@nuance.com

Please make sure to document to us your excellent software engineering skills.

ABOUT US:

Nuance is the leading provider of speech and imaging solutions for businesses and consumers around the world.  Every day, millions of users and thousands of businesses experience Nuance by calling directory assistance, requesting account information, dictating patient records, telling a navigation system their destination, or digitally reproducing documents that can be shared and searched.  With more than 3500 employees worldwide, we are committed to make the user experience more enjoyable by transforming the way people interact with information and how they create, share and use documents. Making each of those experiences productive and compelling is what Nuance is about.

Nuance is the leading provider of speech and imaging solutions for businesses and consumers around the world.  Every day, millions of users and thousands of businesses experience Nuance by calling directory assistance, requesting account information, dictating patient records, telling a navigation system their destination, controlling their mobile phone or digitally reproducing documents that can be shared and searched.  With more than 2000 employees worldwide, we are committed to make the user experience more enjoyable by transforming the way people interact with information and how they create, share and use documents. Making each of those experiences productive and compelling is what Nuance is about. To strengthen our international team we are currently looking for a

 Speech Scientist for Embedded Automatic Speech Recognition

Work Environment

·          You will work in the Embedded ASR research and production team in Merelbeke, Belgium or Aachen, Germany, working with state-of-the-art speech technology, tools and runtime software. Both Gent and Aachen are nice, historical European university cities.

·          You will work in an international company and cooperate with people on various locations, from the USA up to Japan. You may occasionally be asked to travel.

·          You will work on the localization and production of language variants for our cutting edge speech recognition products targeted at automotive platforms and mobile devices. You will help the engine cope with multi-lingual speech in various noise conditions.

·          Initially, you will work on the production of language variants of our acoustic models, later extending your knowledge towards production of statistical language models and natural language dialogue systems.

Key Responsibilities

·          Training of  acoustic models or statistical language models for new languages.

·          Localizing natural language dialogue systems towards a specific market.

·          Contributing to the improvement, design, implementation, evaluation, optimization and testing of new algorithms, tools and processes.

·          Supporting our professional services teams to contribute to customer project success.

·          Assisting senior team members in research tasks.

Your Profile

• You have a University degree in linguistics, engineering, mathematics or physics.
• A PhD or similar experience in a relevant field is a plus.
• Experience in acoustic modeling, NLU or statistical language modeling is recommended.
• Additional background in computational linguistics is a plus.
• Working in Windows and Linux environments comes naturally to you. Experience with computing farms and grid software is welcome.
• You are knowledgable about small, embedded platforms and requirements of software applications designed for them.
• Good software skills are required, especially scripting language like Perl or Python in a Linux/Unix environment, and knowledge of C/C++.
• Experience in speech processing or machine learning techniques is an asset.
• You are a team player, willing to take initiative, and are goal oriented.
• You have a strong sense of precision and quality in your daily job.
• You are fluent in English and you can write high quality documentation.  
• You illustrate your interest in languages by speaking at least two other languages.

Interested?

Please send your CV to Deanna Roe at deanna.roe@nuance.com. If you have any questions, please contact her at +44 207 922 5757.

We are looking forward to receiving your application!

The experience speaks for itself™

Nuance is the leading provider of speech and imaging solutions for businesses and consumers around the world.  Every day, millions of users and thousands of businesses experience Nuance by calling directory assistance, requesting account information, dictating patient records, telling a navigation system their destination, controlling their mobile phone or digitally reproducing documents that can be shared and searched.  With more than 2000 employees worldwide, we are committed to make the user experience more enjoyable by transforming the way people interact with information and how they create, share and use documents. Making each of those experiences productive and compelling is what Nuance is about. To strengthen our international team we are currently looking for a

Senior Research Engineer for Embedded Automatic Speech Recognition

Work Environment

·          You will work in the Embedded ASR research and production team in Merelbeke, Belgium or Aachen, Germany, working with state-of -he-art speech technology, tools and runtime software. Both Gent and Aachen are nice, historical European university cities.

·          You will work in an international company and cooperate with people and research teams on various locations across the globe. You may occasionally be asked to travel.

·          You will work  with our natural language understanding and dialogue research teams as well support our professional services teams.

·          You will work on the development of cutting edge speech recognition products for automotive platforms and mobile devices. You will help the engine cope with multi-lingual speech in various noise conditions, and this while respecting strong limitations on the usage of memory and processing power.

Key Responsibilities

·          Design, implementation, evaluation, optimization and testing of new algorithms and tools, with a strong focus on speech signal processing and acoustic modeling in adverse, noisy environments.

·          Activities are targeted at the creation of commercial products for resource limited platforms.

·          Focus on creating efficient production and development processes to bring the technology to marketable products in a wide range of languages.

·          Occasional application of the developed algorithms and tools for producing systems for a specific language.

·          Specification and follow-up of projects to make the system work with third party components, such as beam formers, echo cancellers or content data providers.

Your Profile

• You have a University degree in engineering, mathematics or physics.
• A PhD degree in speech processing or equivalent relevant experience is a strong asset.
• Experience in speech recognition research, especially acoustic modeling or signal processing, is required.
• Experience in speech processing, machine learning techniques or statistical modeling is required.
• Knowledge about small platforms and experience in developing software for them is a plus.
• Strong software skills are required, especially C/C++ and a scripting language like Perl or Python in a Linux/Unix environment. Knowledge of Matlab is a plus.
• Additional background in computational linguistics is a plus.
• You are a team player, willing to take initiative, and are goal oriented.
• You have a strong desire to make things “really work” in practice, on hardware platforms with limited memory and processing power.
• You are fluent in English and at least one other language, and you can write high quality English documentation. 

Interested?

Please send your CV to Deanna Roe at deanna.roe@nuance.com. If you have any questions, please contact her at +44 207 922 5757.

We are looking forward to receiving your application!

The experience speaks for itself™

Title: jr. Speech Scientist

Location: Aachen, Germany

Type: Permanent

Job:  

Overview:

Nuance is the leading provider of speech and imaging solutions for businesses and consumers around the World. Our technologies, applications and services make the user experience more compelling by transforming the way people interact with information and how they create, share and use documents. Every day, millions of users and thousands of businesses, experience Nuance by calling directory assistance, getting account information, dictating patient records, telling a navigation system their destination, or digitally reproducing documents that can be shared and searched. Making each of those experiences productive and compelling is what Nuance is all about.

Responsibilities:

Nuance is seeking a jr. Speech Scientist who possesses a solid background in natural language technology and computational linguistics.

Candidates should enjoy working in a fast-paced, collaborative atmosphere that applies speech science in a commercial, result driven and customer oriented setting.

As a jr. Speech Scientist in the Embedded Professional Services group, you will work on speech recognition grammars, statistical language models, prompts and custom voice development for leading edge automotive applications across the world, covering a broad range of activities in all project phases, including the design, development, and optimization of the system.

Representative duties include:

• Develop rule based grammars, train statistical language models for speech recognition and natural language understanding in commercial products in a variety of languages, according to UI Design specifications
• Identify or gather suitable text for training language models and custom voices
• Design, develop, and test semantic classifier rules and models
• Develop custom voices for use with Nuance’s leading text to speech products
• Direct voice talents for prompt recordings
• Organize and conduct usability tests
• Localization of speech resources for embedded speech applications
• Optimize accuracy of applications by analyzing performance and tuning statistical language models, grammars, and pronunciations within CPU and memory constraints of embedded platforms
• Contribute to the generation and presentation of client-facing reports

Qualifications:

• University degree in computational linguistics or Software design or similar degree
• Strong analytical and problem solving skills and ability to troubleshoot issues
• Good judgment and quick-thinking
• Strong programming skills, preferably Perl or Python
• Excellent written and verbal communications skills
• Ability to scope work taking technical, business and time-frame constraints into consideration
• Ability and willingness to travel abroad
• Works well independently and collaboratively in team settings in fast-paced environment
• Mastering Office applications

Beneficial Skills

• Additional language skills, eg. French, German, Spanish or other
• Strong programming skills in either Perl, Python, C, VB
• Speech recognition knowledge
• Pattern recognition, linguistics, signal processing, or acoustics knowledge

Opened positions/internships at Microsoft: Danish Linguist (M/F)

MLDC – Microsoft Language Development Center, a branch of the Microsoft Product Group that develops Speech Recognition and Synthesis Technologies, situated in Porto Salvo, Portugal (http://www.microsoft.com/portugal/mldc), is seeking a full-time temporary language expert in the Danish language, for a 3-4 month contract, to work in speech technology related development projects. The successful candidate should have the following requirements:

·         Be native or near native Danish speaker

·         Have a university degree in Linguistics or related field (preferably in Danish Linguistics)

·         Have an advanced level of English

·         Have some experience in working with Speech Technology/Natural Language Processing/Linguistics, either in academia or in industry

·         Have some computational ability – no programming is required, but he/she should be comfortable working with MS Windows and MS Office tools

·         Have team work experience

·         Willing to work in Porto Salvo (near Lisbon) for the duration of the contract

·         Willing to start in September 2008

To apply, please submit your resume and a brief statement describing your experience and abilities to Daniela Braga: i-dbraga@microsoft.com

We will only consider electronic submissions. 

Opened positions/internships at Microsoft: Catalan Linguist (M/F)

MLDC – Microsoft Language Development Center, a branch of the Microsoft Product Group that develops Speech Recognition and Synthesis Technologies, situated in Porto Salvo, Portugal (http://www.microsoft.com/portugal/mldc), is seeking a full-time temporary language expert in the Catalan language, for a 3-4 month contract, to work in speech technology related development projects. The successful candidate should have the following requirements:

·         Be native or near native Catalan speaker

·         Have a university degree in Linguistics or related field (preferably in Catalan Linguistics)

·         Have an advanced level of English

·         Have some experience in working with Speech Technology/Natural Language Processing/Linguistics, either in academia or in industry

·         Have some computational ability – no programming is required, but he/she should be comfortable working with MS Windows and MS Office tools

·         Have team work experience

·         Willing to work in Porto Salvo (near Lisbon) for the duration of the contract

·         Willing to start in September 2008

To apply, please submit your resume and a brief statement describing your experience and abilities to Daniela Braga: i-dbraga@microsoft.com

We will only consider electronic submissions. 

titre du sujet : Synthèse vocale de haute qualité

Introduction
=========

Ces dernières années ont vu l'émergence de systèmes de synthèse de la parole construits autour de base de données de parole de taille importante qui correspondent le plus souvent à quelques heures d'enregistrement de parole. A des degrés divers, ces systèmes considèrent qu'il est possible de produire une parole de qualité en allant chercher des fragments de sons dans une base de données enregistrée au préalable par un locuteur. Ce type d'approche pousse à l'extrême l'hypothèse fonctionnelle des systèmes fondés sur la concaténation d'unités acoustiques. Avec une base de données suffisamment importante, il doit être possible de couvrir statistiquement les cas les plus fréquents de coarticulation sonore.

Des systèmes récents comme Festival (Black 1995), CHATR (Campbell 1996), Whistler (Huang 1996), XIMERA (Toda 2006), IBM \citethese{Eid06}, prouvent que cette approche méthodologique permet de construire des systèmes de synthèse de très bonne qualité.

En suivant cette méthodologie, les modèles ne sont plus utilisés pour produire des valeurs de paramètres qui serviront à la génération d'un signal de parole. Ils sont en revanche utilisés pour rechercher dans la base d'exemples sonores un extrait de parole qui sera le plus proche possible des paramètres modélisés et conformes à une élocution humaine. Concernant la problématique de recherche d'une séquence d'unités acoustiques, différentes solutions sont possibles. Les plus connues appliquent des solutions de recherche de meilleurs chemins (Sagisaka 1992) (Hunt 1996) en proposant une hypothèse de programmation dynmique. D'autres travaux (Donovan 1995) ont défini des modèles acoustiques permettant de guider le choix d'une séquence d'unités.

L'enjeu du procédé de sélection est double, (Iwahashi 1992). Il s'agit d'une part de trouver une correspondance entre une sous-séquence de la chaîne phonémique à synthétiser et un exemplaire plausible dans le corpus de référence. On parle alors d'une \emph{discrimination par critères de cible}, (Hune 1996). Une correspondance à la cible ne suffit pas puisque cette décision est prise unité par unité. Il faut un mécanisme supplémentaire garantissant que l'enchaînement du séquencement proposé réponde à des critères de continuité acoustique (de nature segmentale ou supra-segmentale). On parle dans ce cas de critères de concaténation. La difficulté du problème réside dans le fait que les deux critères sont combinés. Le choix d'une sous-séquence en correspondance avec une unité du corpus dépend de son contexte passé (contexte de la séquence à gauche) et à venir (contexte à droite). Il s'agit encore une fois d'un problème de nature combinatoire qui peut formellement être posé comme un problème de recherche d'un meilleur chemin dans un graphe.

La grande majorité des systèmes de synthèse appliquent un algorithme de Viterbi. Cet algorithme, efficace en complexité spatiale et temporelle, tire sa justification du fait que l'expression du coût global d'une séquence d'unités s'écrit, par hypothèse, sous la forme d'une suite récurrente additive. Cette justification est largement partagée par l'ensemble de la communauté pour ce qui est de l'expression des coûts de concaténation et des coûts de proximité à la cible. En revanche pour ce qui concerne la prise en compte de coûts de nature prosodique, une mise en forme recurrente est plus délicate et difficilement justifiable puisque ces phénomènes ont lieu à l'échelle du groupe intonatif et de la phrase.

Nous considérons qu'il est possible de dépasser la qualité des systèmes de synthèse actuels par la prise en compte de critère prosodiques lors de la recherche de la séquence optimale des unités. Tenir compte de ces critières proposidiques n'est pas une chose simple, puiqu'il faut définir de nouveaux modèles de description des coûts acoustiques et prosodiques d'une séquence. Ces nouvelles techniques de sélection devraient être acapables de proposer des voix avec d'une part plus de relief ou d'expressivité tout en maintenant une très bonne qualité sonore.

(Sagisake 1992) : Sagisaka, Y. and Kaiki, N. and Iwahashi, N. and Mimura, K., ATR mu-TALK speech synthesis system, proceedings of the International Conference on Spoken Language Processing (ICSLP'92)", 1992, pp. 483-486.
(Hunt 1996) :  Hunt, A. and Black, A.W., Unit selection in a concatenative speech synthesis system using a large speech database, proceedings of the IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP'96), 1996, pp. 373-376.
(Donovan 1995) :  Donovan, R. and P. Woodland, P., Automatic speech synthesizer parameter estimation using HMMs, proceedings of the IEEE International Conference on Acoustics and Signal Processing (ICASSP'95), 1995, pp. 640-643.
(Iwahashi 1992) : Iwahashi, N. and Kaiki, N. and Sagisaka, Y., Concatenative speech synthesis by minimum distortion criteria, proceedings of the IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP'92), 1992, pp. 65-68.
(Black 1995) :  Alan W. Black and Nick Campbell, Optimizing selection of units from speech databases for concatenative sysnthesis, proceedings of the IEEE International Conference on Acoustics and Signal Processing (ICASSP'95), vol. 1, pp. 581-584.
(Campbell 1996) :  Campbell, N. and Black, A., CHART: A High-definition speech re-sequencing system, in Progress in Speech Synthesis,
eds. van Santen, J. and Sproat, R. and Olive, J. and Hirschberg, J., 1996, pp. 365-381,
(Huang 1996) : Huang, X. and Acero, A. and Adcock, J. and Hon, H.-W. and Goldsmith, J. and Liu, J. and Plumpe, M., Whistler: A trainable text-to-speech system, proceedings of the International Conference on Spoken Language Processing (ICSLP'96), 1996, pp. 2397-2390.
(Toda 2006) :  Tomoki Toda and Hisashi Kawai and Toshio Hirai and Jinfu Ni and Nobuyuki Nishizawa and Junichi Yamagishi and Minoru Tsuzaki and Keiichi Tokuda and Satoshi Nakamura, Developing a test bed of English text-to-speech system XIMERA for the Blizzard challenge 2006,  Blizzard Challenge, 2006.
(Eide 2006) :  Ellen Eide and Raul Fernandez and Ron Hoory and Wael Hamza and Zvi Kons and Michael Picheny and Ariel Sagi and Slava Shechtman and Zhi Wei Shuang, The IBM submission to the 2006 Blizzard text-to-speech challenge, Blizzard Challenge, 2006.

Proposition d'un travail de thèse
======================

Nous proposons de nous intéresser à de nouvelles méthodologies de sélection d'unités acoustiques pour la synthèse de la parole à partir du texte. La proposition de thèse comporte deux volets: un axe de propositions scientifiques permettant de lever certains verrous notamment dans la formulation du coût d'une séquence d'unités, et un axe expérimental par la proposition d'une évolution du système de synthèse du groupe Cordial permettant de mettre en place des évaluations perceptuelles qui permettront de valider ou d'invalider les hypothèses de travail qui auront été choisies. Le travail expérimental sera réalisé sur le français. Nous souhaitons doubler les expérimentations sur l'anglais et participer ainsi au challenge Blizzard qui est une compétition internationale en synthèse de la parole.

Le travail de thèse prendra comme point d'appui la proposition suivante:
  * Mise en place et évaluation d'un premier système reposant sur l'état de l'art actuel en synthèse de la parole par corpus de parole continue. Prise en compte des niveaux acoustiques. Utilisation d'une base de parole expressive,  "chronic",  issue du projet ANR Vivos.
  * Proposition de modèles de sélection de nature prosodique.
  * Propositions algorithmiques, définition d'heuristiques pour une solution acceptable en temps de calcul.
  * Intégration des propositions prosodiques au système de synthèse de référence et évaluation.

Contexte du travail de thèse
===================

L'étudiant sera accueilli au sein de l'équipe Cordial de l'irisa : http://www.irisa.fr/cordial dont les
principaux travaux concernent le traitement de la parole : synthèse, transformation de parole, annotation de corpus.
L'équipe de recherche est hébergée dans les locaux de l'Ecole Nationale Supérieure des Sciences Appliquées et de Technologie,
http://www.enssat.fr, à Lannion. La thèse est financée sur trois ans par une bourse du conseil général des Côtes d'Armor.

__________________________________________________________________________________________________________

Opened positions/internships at Microsoft: Norwegian Linguist (M/F)

MLDC – Microsoft Language Development Center, a branch of the Microsoft Product Group that develops Speech Recognition and Synthesis Technologies, situated in Porto Salvo, Portugal (http://www.microsoft.com/portugal/mldc), is seeking a full-time temporary language expert in the Norwegian language (Bokmal), for a 4-6 month contract, to work in speech technology related development projects. The successful candidate should have the following requirements:

·         Be native or near native Norwegian Bokmal speaker

·         Have a university degree in Linguistics or related field (preferably in Norwegian Linguistics)

·         Have an advanced level of English

·         Have some experience in working with Speech Technology/Natural Language Processing/Linguistics, either in academia or in industry

·         Have some computational ability – no programming is required, but he/she should be comfortable working with MS Windows and MS Office tools

·         Have team work experience

·         Willing to work in Porto Salvo (near Lisbon) for the duration of the contract

·         Willing to start in October 2008

To apply, please submit your resume and a brief statement describing your experience and abilities to Daniela Braga: i-dbraga@microsoft.com

We will only consider electronic submissions.

Deadline for submissions: August 10, 2008 

Opened positions/internships at Microsoft: Finnish Linguist (M/F)

MLDC – Microsoft Language Development Center, a branch of the Microsoft Product Group that develops Speech Recognition and Synthesis Technologies, situated in Porto Salvo, Portugal (http://www.microsoft.com/portugal/mldc), is seeking a full-time temporary language expert in the Finnish language, for a 6 month contract, to work in speech technology related development projects. The successful candidate should have the following requirements:

·         Be native or near native Finnish speaker

·         Have a university degree in Linguistics or related field (preferably in Norwegian Linguistics)

·         Have an advanced level of English (oral and written)

·         Have some experience in working with Speech Technology/Natural Language Processing/Linguistics, either in academia or in industry

·         Have some computational ability – no programming is required, but he/she should be comfortable working with MS Windows and MS Office tools

·         Have team work experience

·         Willing to work in Porto Salvo (near Lisbon) for the duration of the contract

·         Willing to work in a multicultural and multinational team across the globe

·         Willing to start in September 1, 2008

To apply, please submit your resume and a brief statement describing your experience and abilities to Daniela Braga: i-dbraga@microsoft.com

We will only consider electronic submissions.

Deadline for submissions: August 10, 2008

Opened positions/internships at Microsoft: Norwegian, Finnish, Italian, French, English, Polish Linguists (M/F)

MLDC – Microsoft Language Development Center, a branch of the Microsoft Product Group that develops Speech Recognition and Synthesis Technologies, situated in Porto Salvo, Portugal (http://www.microsoft.com/portugal/mldc), is seeking full-time temporary language experts in the following languages:

-          Norwegian language (Bokmal variety)

-          Finnish

-          Italian

-          English (UK)

-          French (France)

-          Polish

The contracts range from 2-6 months and the scope of them is to work in speech technology related development projects. The successful candidate should have the following requirements:

·         Be a native or near native speaker (for each of the required language)

·         Have a university degree in Linguistics or related field

·         Have an advanced level of English (oral and written)

·         Have some experience in working with Speech Technology/Natural Language Processing/Linguistics, either in academia or in industry

·         Have some computational ability – no programming is required, but he/she should be comfortable working with MS Windows and MS Office tools

·         Have team work experience

·         Willing to work in Porto Salvo (near Lisbon) for the duration of the contract

·         Willing to work in a multicultural and multinational team across the globe

·         Willing to start immediately

To apply, please submit your resume and a brief statement describing your experience and abilities to Daniela Braga: i-dbraga@microsoft.com

We will only consider electronic submissions.

Deadline for submissions: Opened until filled.

Title:      Assistant Professor Position at EURECOM
            in Multimedia content analysis and processing
Department: Multimedia Communications
URL:        http://www.eurecom.fr/research/
Start date: September 2008

Description:
Research in the Department currently focuses on several aspects of the processing of Multimedia Content:
• Video analysis and information filtering,
• Image Processing with application to 3D Face Cloning, watermarking and
biometrics,
• Speech and sound processing.
The new faculty is expected to undertake research in close collaboration with the existing activities and to participate in the teaching program for Master students. Extensions of the current research areas are encouraged.

Requirements:
The candidates must have a Ph.D. in computer science or electrical engineering with a solid background in signal processing and statistical analysis. The ideal candidate will have an established record of conducting research activities at the international level, and a proven record of collaboration with academic and industrial partners in national and European programs or equivalent. Excellence in research is a constant requirement for EURECOM. A strong commitment to excellence in research and teaching is essential.

Applications:
Send, by email, a letter of motivation, a resume including a list of publications, the names of 3 references and a copy of the three most important publications.

Contact:        Prof. Bernard Merialdo
Postal address: 2229 route des Crêtes
                BP 193
                06904 Sophia Antipolis cedex
                France
Email address:  Bernard.Merialdo@eurecom.fr
Web address:    http://www.eurecom.fr/main/institute/job.fr.htm
Phone:          +33/0 4 93 00 81 29
Fax:            +33/0 4 93 00 82 00

Located in the heart of Sophia Antipolis technology park, EURECOM is a graduate school and a Research center in Communication Systems, founded in 1991 by TELECOM ParisTech (Ecole Nationale Supérieure des Télécommunications) and EPFL (Swiss federal institute of Lausanne) in a consortium form, combining academic and industrial partners. Teaching and research activities of EURECOM focus on three areas: networking and security, mobile communications and multimedia communications. EURECOM has a strong international scope and strategy. 

IEEE Transactions on Multimedia            
 Special Issue on Integration of Context and Content for Multimedia Management
=====================================================================

Guest Editors:

Alan Hanjalic, Delft University of Technology, The Netherlands
Alejandro Jaimes, IDIAP Research Institute, Switzerland
Jiebo Luo, Kodak Research Laboratories, USA
        Qi Tian, University of Texas at San Antonio, USA

---------------------------------------------------
URL: http://www.cs.utsa.edu/~qitian/cfp-TMM-SI.htm
---------------------------------------------------
Important dates:

Manuscript Submission Deadline:       April 1, 2008
        Notification of Acceptance/Rejection: July 1, 2008
        Final Manuscript Due to IEEE:         September 1, 2008
        Expected Publication Date:            January 2009

---------------------
Submission Procedure
---------------------
Submissions should follow the guidelines set out by IEEE Transaction on Multimedia.
Prospective authors should submit high quality, original manuscripts that have not
appeared, nor are under consideration, in any other journals.

-------
Summary
-------
Lower cost hardware and growing communications infrastructure (e.g., Web, cell phones,
etc.) have led to an explosion in the availability of ubiquitous devices to produce,
store, view and exchange multimedia (images, videos, music, text). Almost everyone is
a producer and a consumer of multimedia in a world in which, for the first time,
tremendous amount of contextual information is being automatically recorded by the
various devices we use (e.g., cell ID for the mobile phone location, GPS integrated in
a digital camera, camera parameters, time information, and identity of the producer).

In recent years, researchers have started making progress in effectively integrating
context and content for multimedia mining and management. Integration of content and
context is crucial to human-human communication and human understanding of multimedia:
without context it is difficult for a human to recognize various objects, and we
become easily confused if the audio-visual signals we perceive are mismatched. For the
same reasons, integration of content and context is likely to enable  (semi)automatic
content analysis and indexing methods to become more powerful in managing multimedia
data. It can help narrow part of the semantic and sensory gap that is difficult or
even impossible to bridge using approaches that do not explicitly consider context for
(semi)automatic content-based analysis and indexing.

The goal of this special issue is to collect cutting-edge research work in integrating
content and context to make multimedia content management more effective. The special
issue will unravel the problems generally underlying these integration efforts,
elaborate on the true potential of contextual information to enrich the content
management tools and algorithms, discuss the dilemma of generic versus narrow-scope
solutions that may result from "too much" contextual information, and provide us
vision and insight from leading experts and practitioners on how to best approach the
integration of context and content. The special issue will also present the state of
the art in context and content-based models, algorithms, and applications for
multimedia management.

-----
Scope
-----

The scope of this special issue is to cover all aspects of context and content for
multimedia management.

Topics of interest include (but are not limited to):
        - Contextual metadata extraction
        - Models for temporal context, spatial context, imaging context (e.g., camera
          metadata), social and cultural context and so on
- Web context for online multimedia annotation, browsing, sharing and reuse
- Context tagging systems, e.g., geotagging, voice annotation
- Context-aware inference algorithms
        - Context-aware multi-modal fusion systems (text, document, image, video,
          metadata, etc.)
- Models for combining contextual and content information
        - Context-aware interfaces
- Context-aware collaboration
- Social networks in multimedia indexing
- Novel methods to support and enhance social interaction, including
          innovative ideas integrating context in social, affective computing, and
          experience capture.
- Applications in security, biometrics, medicine, education, personal
          media management, and the arts, among others
- Context-aware mobile media technology and applications
- Context for browsing and navigating large media collections
- Tools for culture-specific content creation, management, and analysis

------------
Organization
------------
Next to the standard open call for papers, we will also invite a limited number of
papers, which will be written by prominent authors and authorities in the field
covered by this Special Issue. While the papers collected through the open call are
expected to sample the research efforts currently invested within the community on
effectively combining contextual and content information for optimal analysis,
indexing and retrieval of multimedia data, the invited papers will be selected to
highlight the main problems and approaches generally underlying these efforts.

All papers will be reviewed by at least 3 independent reviewers. Invited papers will
be solicited first through white papers to ensure the quality and relevance to the
special issue. The accepted invited papers will be reviewed by the guest editors and
expect to account for about one fourth of the papers in the special issue.

---------
Contacts
---------
Please address all correspondences regarding this special issue to the Guest Editors
Dr. Alan Hanjalic (A.Hanjalic@ewi.tudelft.nl), Dr. Alejandro Jaimes
(alex.jaimes@idiap.ch), Dr. Jiebo Luo (jiebo.luo@kodak.com), and Dr. Qi Tian
(qitian@cs.utsa.edu).
-------------------------------------------------------------------------------------

Guest Editors:
Alan Hanjalic, Alejandro Jaimes, Jiebo Luo, and Qi Tian

CALL FOR PAPERS
Special Issue of Speech Communication

on *Spoken Language Technology for Education*


*Guest-editors:*
Maxine Eskenazi, Associate Teaching Professor, Carnegie Mellon University
Abeer Alwan, Professor, University of California at Los Angeles
Helmer Strik, Assistant Professor, University of Nijmegen
 

Language technologies have evolved to the stage where they are reliable
enough, if their strong and weak points are properly dealt with, to be
used for education. The creation of an application for education
presents several challenges: making the language technology sufficiently
reliable (and thus advancing our knowledge in the language
technologies), creating an application that actually enables students to
learn, and engaging the student. Papers in this special issue should
deal with several of these issues. Although language learning is the
primary target of research at present, papers on the use of language
technologies for other education applications are encouraged. The scope
of acceptable topic interests includes but is not limited to:

 

- Use of speech technology for education

- Use of spoken language dialogue for education

- Applications using speech and natural language processing for education

- Intelligent tutoring systems using speech and natural language

- Pedagogical issues in using speech and natural language technologies
for education

- Assessment of tutoring software

- Assessment of student performance

 

*Tentative schedule for paper submissions, review, and revision**: ** *

Deadline for submissions: June 1, 2008.

Deadline for decisions and feedback from reviewers and editors: August
31, 2008.

Deadline for revisions of papers: November 31, 2008.

 

*Submission instructions:*

Authors should consult the "Guide for Authors", available online, at
http://www.elsevier.com/locate/specom for information about the
preparation of their manuscripts. Authors, please submit your paper via
_http://ees.elsevier.com/specom_, choosing *Spoken Language Tech. *as
the Article Type, and  Dr. Gauvain as the handling E-i-C.

Call for Papers

Recent Advances in Biometric Systems: A Signal Processing Perspective

Biometrics a digital recognition technology that relies on highly distinctive physical and physiological characteristics of an individual is potentially a powerful and reliable method for personal authentication. The increasing importance of biometrics is underscored by the rapidly growing number of educational and research activities devoted to this field; and by a large number of annually organized Conferences and Symposia exclusively devoted to biometrics. Biometrics is a multidisciplinary field with researchers from signal processing, pattern recognition, computer vision, and statistics. Recently, a number of new important directions have been identified for biometric research, including processing and encoding of nonideal data, biometrics at a distance, and data quality assessment. Problems in nonideal biometric data include off-angle, occluded, blurred, and noisy images. Biometrics at a distance is concerned with recognition from video or snapshots of a biometric samples captured from a noncooperative moving individual. The goal of this special issue is to focus on recent advances in signal processing of biometric data that allow improved recognition performance through novel restoration, processing, and encoding; matching techniques capable of dealing with complexity and distortions in data acquired from a distance; recognition from biometric data acquired from unconstrained environments or complex experimental set ups; and the characterization of quality and its relationship with performance.

Topics of interest include, but are not limited to:

Biometric-based recognition under unconstrained presentation and/or complex environment using the following:
    o Face
    o Iris
    o Fingerprint
    o Voice
    o Hand
    o Soft biometrics

Multimodal biometric recognition using nonideal data

Biometric image/signal quality assessment:
    o Face
    o Iris
    o Fingerprint
    o Voice
    o Hand
    o Soft biometrics

Biometric security and privacy
    o Liveness detection
    o Encryption
    o Cancelable biometrics

The special issue will focus both on the development and comparison of novel signal/image processing approaches and on their expanding range of applications.

Authors should follow the EURASIP Journal on Advances in Signal Processing manuscript format described at the journal site http://www.hindawi.com/journals/asp/. Prospective authors should submit an electronic copy of their complete manuscript through the journal Manuscript Tracking System at http://mts.hindawi.com/, according to the following timetable:

Manuscript Due                 October 1, 2008
First Round of Reviews         January 1, 2009
Publication Date               April 1, 2009

Guest Editors

o Natalia A. Schmid, Lane Department of Computer Science and Electrical Engineering, West Virginia University, Morgantown, WV 26506, USA; natalia.schmid@mail.wvu.edu
o Stephanie Schuckers, Electrical & Computer Engineering, Clarkson University, Potsdam, NY 13699, USA; sschucke@clarkson.edu
o Jonathon Phillips, National Institute of Standard and Technology, Gaithersburg, MD 20899, USA; jonathon@nist.gov
o Kevin Bowyer, University of Notre Dame, Notre Dame, IN 46556, USA; kwb@cse.nd.edu 

Call for Papers

Special issue of Eurasip journal on advanced signal processing

Analysis and Signal Processing of Oesophagial and Pathological Voices

Speech is the most important means of communication among humans. Speech, however, is not limited only to the process of communication, but is also very important for transferring emotions, expressing our personality, and reflecting situations of stress. Modern lifestyles have increased the risk of experiencing some kind of voice alteration. It is estimated that around 19% of the population suffer or have suffered from dysphonic voicing. This motivates new and objective ways to evaluate speech, its quality, and its connection with other phenomena.

Speech research to date has favored areas such as synthesis, recognition, and speaker verification. The last few years have witnessed the emerging topic of processing and evaluation of disordered speech. Acoustic analysis is a noninvasive technique providing an efficient tool for the objective diagnosis, the screening of voice diseases, the objective determination of vocal function alterations, and the evaluation of surgical treatment and rehabilitation. Its application extends beyond medicine, and now includes forensic analysis as well as voice quality control for voice professionals. Acoustic analysis may also be seen as complementary to other methods of evaluation based on the direct observation of the vocal folds using videoendoscopy.

This special issue aims to foster an interdisciplinary forum for presenting new work in the analysis and modeling of voice signals and videoendoscopic images, with applications in pathological and oesophageal voices.

Topics of interest include (but are not limited to):

o Automatic detection of voice disorders
o Automatic assessment and classification of voice quality
o New strategies for the parameterization and modeling of normal and pathological voices (biomechanical-based parameters, chaos modeling, etc.)
o Databases of vocal disorders
o Inverse filtering
o Signal processing for remote diagnosis
o Speech enhancement for pathological and oesophageal voices
o Objective parameters extraction from vocal fold images using videolaryngoscopy, videokymography, fMRI, and other emerging techniques
o Multimodal analysis of disordered speech
o Robust pitch extraction algorithms for pathological and oesophageal voices

Since speech communication is fundamental to human interaction, we are moving towards a new scenario where speech is gaining greater importance in our daily lives, and many common speech disorders and dysfunctions would be treated using computer-based or physical prosthetics.

Authors should follow the EURASIP Journal on Advances in Signal Processing manuscript format described at http://www.hindawi.com/journals/asp/. Prospective authors should submit an electronic copy of their complete manuscript through the journal Manuscript Tracking System at http://mts.hindawi.com/ according to the following tentative timetable:

Manuscript Due                 November 1, 2008
First Round of Reviews         February 1, 2009
Publication Date               May 1, 2009

Guest Editors

o Juan I. Godino-Llorente, Department of Circuits and Systems Engineering, Polytechnic University of Madrid (UPM), Ctra. de Valencia, 28031 Madrid, Spain; igodino@ics.upm.es
o Pedro Gómez-Vilda, Department of Computer Science and Engineering, Polytechnic University of Madrid (UPM), Boadilla del Monte, 28660 Madrid, Spain; pedro@pino.datsi.fi.upm.es
o Tan Lee, Department of Electronic Engineering, The Chinese University of Hong Kong, Shatin, New Territories, Hong Kong; tanlee@ee.cuhk.edu.hk 

Special Issue on "Emergent Artificial Intelligence Approaches for Pattern Recognition in Speech and Language Processing"

      Computer Speech and Language, Elsevier       Deadline for paper submission: September 26, 2008.  http://speechlab.ifsc.usp.br/call/csl.pdf                        =

Consonant Challenge:
  Human-machine comparisons of consonant recognition in quiet and noise

                   Interspeech, 22-26 September 2008
                         Brisbane, Australia

* Update:
All information concerning the native listener experiments and baseline 
recognisers
including their results can now be found and downloaded from the Consonant 
Challenge website:
http://www.odettes.dds.nl/challenge_IS08/

* Deadline for submissions:
The deadline and paper submission guidelines for full paper submission (4 
pages) is April
7th, 2008. Paper submission is done exclusively via the Interspeech 2008 
conference website.
Participants of this Challenge are asked to indicate the correct Special 
Session during
submission. More information on the Interspeech conference can be found 
here: http://
www.interspeech2008.org/

* Topic of the Consonant Challenge:
Listeners outperform automatic speech recognition systems at every level 
of speech
recognition, including the very basic level of consonant recognition. What 
is not clear is
where the human advantage originates. Does the fault lie in the acoustic 
representations of
speech or in the recogniser architecture, or in a lack of compatibility 
between the two?
There have been relatively few studies comparing human and automatic 
speech recognition on
the same task, and, of these, overall identification performance is the 
dominant metric.
However, there are many insights which might be gained by carrying out a 
far more detailed
comparison.

The purpose of this Special Session is to make focused human-computer 
comparisons on a task
involving consonant identification in noise, with all participants using 
the same training
and test data. Training and test data and native listener and baseline 
recogniser results
will be provided by the organisers, but participants are encouraged to 
also contribute
listener responses.

* Call for papers:
Contributions are sought in (but not limited to) the following areas:

- Psychological models of human consonant recognition
- Comparisons of front-end ASR representations
- Comparisons of back-end recognisers
- Exemplar vs statistical recognition strategies
- Native/Non-native listener/model comparisons

* Organisers:
Odette Scharenborg (Radboud University Nijmegen, The Netherlands -- 
O.Scharenborg@let.ru.nl)
Martin Cooke (University of Sheffield, UK -- M.Cooke@dcs.shef.ac.uk)

CALL FOR WORKSHOP PROPOSALS EACL 2009, NAACL HLT 2009, AND ACL-IJCNLP 2009

Joint site:  http://www.eacl2009.gr/conference/callforworkshops

10-13 September 2008, Zadar, Croatia

http://www.elmar-zadar.org/

TECHNICAL CO-SPONSORS

IEEE Region 8

EURASIP - European Assoc. Signal, Speech and Image Processing

IEEE Croatia Section

IEEE Croatia Section Chapter of the Signal Processing Society

IEEE Croatia Section Joint Chapter of the AP/MTT Societies

TOPICS

--> Image and Video Processing

--> Multimedia Communications

--> Speech and Audio Processing

--> Wireless Commununications

--> Telecommunications

--> Antennas and Propagation

--> e-Learning and m-Learning

--> Navigation Systems

--> Ship Electronic Systems

--> Power Electronics and Automation

--> Naval Architecture

--> Sea Ecology

--> Special Session Proposals - A special session consist

of 5-6 papers which should present a unifying theme

from a diversity of viewpoints; deadline for proposals

is February 04, 2008.

KEYNOTE TALKS

* Professor Sanjit K. Mitra, University of Southern California, Los Angeles, California, USA:

Image Processing using Quadratic Volterra Filters

* Univ.Prof.Dr.techn. Markus Rupp, Vienna University

of Technology, AUSTRIA:

Testbeds and Rapid Prototyping in Wireless Systems

* Professor Paul Cross, University College London, UK:

GNSS Data Modeling: The Key to Increasing Safety and

Legally Critical Applications of GNSS

* Dr.-Ing. Malte Kob, RWTH Aachen University, GERMANY:

The Role of Resonators in the Generation of Voice

Signals

SPECIAL SESSIONS

SS1: "VISNET II - Networked Audiovisual Systems"

Organizer: Dr. Marta Mrak, I-lab, Centre for Communication

Systems Research, University of Surrey, UNITED KINGDOM

Contact: http://www.ee.surrey.ac.uk/CCSR/profiles?s_id=3D3937

SS2: "Computer Vision in Art"

Organizer: Asst.Prof. Peter Peer and Dr. Borut Batagelj,

University of Ljubljana, Faculty of Computer and Information

Science, Computer Vision Laboratory, SLOVENIA

Contact: http://www.lrv.fri.uni-lj.si/~peterp/ or

http://www.fri.uni-lj.si/en/personnel/298/oseba.html

SUBMISSION

Papers accepted by two reviewers will be published in

symposium proceedings available at the symposium and

abstracted/indexed in the INSPEC and IEEExplore database.

More info is available here: http://www.elmar-zadar.org/

IMPORTANT: Web-based (online) paper submission of papers in

PDF format is required for all authors. No e-mail, fax, or

postal submissions will be accepted. Authors should prepare

their papers according to ELMAR-2008 paper sample, convert

them to PDF based on IEEE requirements, and submit them using

web-based submission system by March 03, 2008.

SCHEDULE OF IMPORTANT DATES

Deadline for submission of full papers: March 03, 2008

Notification of acceptance mailed out by: April 21, 2008

Submission of (final) camera-ready papers : May 05, 2008

Preliminary program available online by: May 12, 2008

Registration forms and payment deadline: May 19, 2008

Accommodation deadline: June 02, 2008

GENERAL CO-CHAIRS

Ive Mustac, Tankerska plovidba, Zadar, Croatia

Branka Zovko-Cihlar, University of Zagreb, Croatia

PROGRAM CHAIR

Mislav Grgic, University of Zagreb, Croatia

CONTACT INFORMATION

Assoc.Prof. Mislav Grgic, Ph.D.

FER, Unska 3/XII

HR-10000 Zagreb

CROATIA

Telephone: + 385 1 6129 851=20

Fax: + 385 1 6129 568=20

E-mail: elmar2008 (_) fer.hr

For further information please visit:

http://www.elmar-zadar.org/

EMUS

Expressivity in Music and Speech

http://recherche.ircam.fr/equipes/analyse-synthese/EMUS/

Fourth Conference

September 2008

Thursday 25 and Friday 26

Expression of emotions in Speech and Music

Microgenesis and semiotics of perceptual process

Place : RISC, 28, rue serpente, Paris 6ème, métro Odéon ou Saint-Michel (plan http://www.risc.cnrs.fr./plan.php), salle S35

Scientific Committee : Antoine Auchlin, Greg Beller, Didier Bottineau, Anne Lacheret, Aliyah Morgenstern, Nicolas Obin

Organasing Committee :Didier Bottineau, Greg Beller, Anne Lacheret, Nicolas Larousse, Aliyah Morgenstern, Nicolas Obin

Speakers : Antoine Auchlin (phonetics & linguistics), Mireille Besson (neurosciences), Didier Bottineau (linguistics), Christophe D’Alessandro (computer sciences & music), Michel Imberty (psychology), Anne Lacheret & Dominique Legallois (phonetic and linguistics), Valérie Pasdeloup & David Piotrowski (phonetics & linguistics), Xavier Rodet (signal processing), Victor Rosenthal (psycholinguistics), Daniel Schon (neurosciences), Jean-Luc Schwartz (computer sciences), Barbara Tillman (neurosciences).

Programme

9.30-10                        Anne Lacheret: _{(MODYCO, UMR 7114, Paris X, Nanterre, Institut Universitaire de France)}: Introduction to the workshop

10-11               Victor Rosenthal _{(MODYCO, UMR 7114, Paris X, Nanterre, France)}: Microgenesis and the expressive form of life

11-12               Antoine Auchlin _{(Department of Linguistics, University of Geneva)}: Meu su, voyons! - from meant meaning to meaning meaning. Notes on enaction, microgenesis and experiential blending

  12-13               Jean Luc Schwartz _{(GIPSA, Grenoble, France)}: From auditory patterns to speech “patterned” by perceptuo-motor interactions

13.15-14.30 : Lunch

14.45-15.45                 Mireille Besson¹, Mitsuko Aramaki¹, Daniele Schön¹, Aline Frey² _{(1Institut de Neurosciences Cognitives de la Méditerranée, CNRS-Marseille Universités, Marseille, 2 Laboratoire Cognitions Humaine et Artificielle; Université Paris 8, France)}: An interdisciplinary approach to the semiotics of sounds

15.45-16.45                 Christophe d’Alessandro, Sylvain Le Beux, Albert Rilliard _{(LIMSI, Orsay, France)}: Towards kinematical modelling of expressive speech prosody: experiments in computerized chironomy

16.45-17.45                 Michel Imberty _{(Département de psychologie, université de Paris X, Nanterre)}: Voice, musicality and temporality

9.30-10.30                         Anne Lacheret¹ & Dominique Legallois² _{(1MODYCO, ParisX Nanterre & IUF, Paris, 2CRISCO, université de Caen)} : Expressivity and emotion in spoken language: what does grammar have to tell us?

10.30-11.30                 Barbara Tillmann* & W. Jay Dowling ** _{(* CNRS-UMR 5020, Lyon, France; ** University of Texas at Dallas, USA.)}: Memory of Music and Poetry: Keeping Details over Time

11.30-12.30                 Xavier Rodet _{(IRCAM, Paris, France)} : Voice transformation: methods, means and applications in music, cinema and multi-media

12.45-14 : Lunch

14.15-15.15                 Daniele Schon _{(Institut de Neurosciences Cognitives de la Méditerranée, CNRS-Marseille, France)} : Music and language: cerebral functions or cultural artifacts?

15.15-16.15                 Valérie Pasdeloup¹ & David Piotroswki² _{(1 LPL, Aix-en-Provence, 2CREA, Paris, France)}: On some aspects of sign perception: the case of the temporal structure of speech

16.15-17.15                 Discussion

COLLOQUE « LANGAGE ET COGNITION »

Ier et 2 octobre 2008

Maison de la Chimie

28, rue Saint-Dominique

75007 PARIS

Organisé par le Réseau Thématique Européen « Langage et Cognition »

Issu de l’ACI COGNITIQUE (1999-2002)

Sous l’égide de la Fondation Maison des Sciences de l’Homme

Adresse du Site : http://www.msh-paris.fr

Adresse du site d’OCTOGONE E.A 4156 :

http://w3.octogone.univ-tlse2.fr/articles.php?lng=fr&pg=20

Coordonnateurs :

Jean-Luc Nespoulous

Université de Toulouse Le Mirail

Institut Universitaire de France

Unité de Recherche Interdisciplinaire OCTOGONE E.A 4156

Laboratoire Jacques-Lordat

Institut des Sciences du Cerveau de Toulouse IFR 96

nespoulo@univ-tlse2.fr

&

Michel Fayol

Université Blaise Pascal, Clermont-Ferrand

Laboratoire de Psychologie Sociale et Cognitive UMR CNRS 6024

Mercredi Ier Octobre

8h30 : Accueil

8h45 : Introduction : Jean-Luc Nespoulous et Michel Fayol

Thème 1 : « Langage et Cognition : de la phrase au discours »

9h :       « From sentence to discourse : spatial framing adverbials as discourse markers »

Michel Charolles & Laure Sarda (Paris)

9h45 : « Effect on understanding of preposed and postposed locative prepositional phrases »

            Saveria Colonna (Paris)

« Effect of locative prepositional phrases on anaphor resolution and lexical desambiguation »

Joël Pynte (Paris)

10h30 : Discussion

10h45 : PAUSE

11h15 : « From Situation Models to Mental Simulations »

            Rolf Zwaan (Rotterdam)

12h : Discussion générale

12h-30 : repas

Thème 2 : « Diversité des Langues et Cognition »

14h30 : « La diversité des langues à travers les phénomènes de classification dans les langues orales et signées : aspects méthodologiques »

Barbara Köpke (Toulouse) & Colette Grinevald (Lyon)

15h15 : « Scripturisation des langues des signes et contraintes liées à la modalité : état des lieux »

            Brigitte Garcia (Paris)

16h : Discussion

16h15 : PAUSE

16h45 : « Formes linguistiques et catégories sémantiques : le cas des hiéroglyphes classificateurs de l’ancienne Egypte »

            Orly Goldwasser (Jerusalem), Colette Grinevald (Lyon) & Danièle Dubois (Paris)

17h30 : Discussion générale

Jeudi 2 Octobre

8h30 : Invited address : « The roles of general cognitive capacities and domain-specific cognitive skills in language comprehension »

            David Caplan (Cambridge) & Gloria Waters (Boston)

Thème 3 : « Langage, Communication, Pragmatique »

9h15 : « Quelles contraintes pragmatiques pour le langage et la communication ? »

            Michèle Guidetti (Toulouse) & Jean-Louis Dessalles (Paris)

10h : « Quel lien entre troubles pragmatiques et déficits cognitifs ? Le cas de la schizophrénie et des lésions cérébrales droites »

Maud Champagne (Montréal)

10h45 : Discussion

11h : PAUSE

11h30 :  « Structure(s) et contexte(s) dans une théorie pragmatique»

            Jef Verschueren (Anvers)

12h15 : Discussion générale

12h-30 : repas

Thème 4 : « Perception & Compréhension »

14h30 : « Speaking in the brain »

            Jean-François Démonet (Toulouse) & Mireille Besson (Marseille)

15h15 : « Abstract representations in speech perception : sounds, words and voices »

James McQueen (Nijmegen)

16h :     Discussion

16h15 : PAUSE

16h45 : « Functional and structural neuroanatomy of language »

            Angela Friederici (Leipzig)

17h30 : Discussion générale

18h : Fin des travaux

2008 IEEE International Workshop on MACHINE LEARNING FOR SIGNAL PROCESSING
(Formerly the IEEE Workshop on Neural Networks for Signal Processing)

October 16-19, 2008 Cancun, Mexico
Fiesta Americana Condesa Cancun, www.fiestamericana.com

Deadlines:
Submission of full paper:                     May 5, 2008
Notification of acceptance:                     June 16, 2008
Camera-ready paper and author registration:     June 23, 2008
Advance registration before:                    July 1, 2008

http://mlsp2008.conwiz.dk/

The workshop will feature keynote addresses, technical presentations, special
sessions and tutorials organized in two themes that will be included in the
registration. Tutorials will take place on the afternoon of 16 October, and
the workshop will begin on 17 October. The two themes for MLSP 2008 are
Cognitive Sensing and Kernel Methods for Nonlinear Signal Processing. Papers
are solicited for, but not limited to, the following areas:

Algorithms and Architectures:
Artificial neural networks, kernel methods, committee models, Gaussian
processes, independent component analysis, advanced (adaptive, nonlinear)
signal processing, (hidden) Markov models, Bayesian modeling, parameter
estimation, generalization, optimization, design algorithms.

Applications:
Speech processing, image processing (computer vision, OCR) medical imaging,
multimodal interactions, multi-channel processing, intelligent multimedia and
web processing, robotics, sonar and radar, biomedical engineering, financial
analysis, time series prediction, blind source separation, data fusion, data
mining, adaptive filtering, communications, sensors, system identification,
and other signal processing and pattern recognition applications.

Implementations:
Parallel and distributed implementation, hardware design, and other general
implementation technologies.

For the fourth consecutive year, a Data Analysis and Signal Processing
Competition is being organized in conjunction with the workshop. The goal of
the competition is to advance the current state-of-the-art in theoretical and
practical aspects of signal processing domains. The problems are selected to
reflect current trends, evaluate existing approaches on common benchmarks, and
identify critical new areas of research. Previous competitions produced novel
and effective approaches to challenging problems, advancing the mission of the
MLSP community. A description of the competition, the submissions, and the
results, will be included in a paper which will be published in the
proceedings. Winners will be announced and awards given at the workshop.

Selected papers from MLSP 2008 will be considered for a special issue of The
Journal of Signal Processing Systems for Signal, Image, and Video Technology,
to appear in 2009. The MLSP technical committee may invite one or more winners
of the data analysis and signal processing competition to submit a paper
describing their methodology to the special issue.

Paper Submission Procedure
Prospective authors are invited to submit a double column paper of up to six
pages using the electronic submission procedure at http://mlsp2008.conwiz.dk.
Accepted papers will be published on a CDROM to be distributed at the
workshop.

MLSP'2007 webpage: http://mlsp2008.conwiz.dk/

MLSP 2008 ORGANIZING COMMITTEE:

General Chair
Jose Principe

Program Chair
Deniz Erdogmus

Technical Chair
Tulay Adali

Publicity Chairs
Ignacio Santamaria
Marc Van Hulle

Publication Chair
Jan Larsen

Data Competition
Ken Hild
Vince Calhoun

Local Arrangements
Juan Azuela

The Tenth International Conference on Multimodal Interfaces (ICMI

2008) will take place in Chania, Greece, on October 20-22, 2008. The

main aim of ICMI 2008 is to further scientific research within the

broad field of multimodal interaction and systems. The conference will

focus on major trends and challenges in this area, including help

identify a roadmap for future research and commercial success. ICMI

2008 will feature a main conference with keynote speakers, panel

discussions, technical paper presentations and discussion (single

track), poster sessions, and demonstrations of state-of-the-art

multimodal concepts and systems. Organized on the island of Crete,

ICMI-08 provides excellent conditions for brainstorming and sharing

the latest advances about multimodal interaction and systems in an

inspired setting full of history, mythology and art.

Paper Submission

There are two different submission categories: regular paper and short

paper. The page limit is 8 pages for regular papers and 4 pages for

short papers. The presentation style (oral or poster) will be decided

based on suitable delivery of the content.

Demo Submission

Proposals for demonstrations shall be submitted to demo chairs

electronically. A 1-2 page description of the demonstration is required.

Doctoral Spotlight

Doctoral Student Travel Support and Spotlight Session. Funds are

expected from NSF to support participation of doctoral candidates at

ICMI 2008, and a spotlight session is planned to showcase ongoing

thesis work. Students

interested in travel support can submit a short or long paper as

specified above.

Topics of interest include

* Multimodal and Multimedia processing

* Multimodal input and output interfaces

* Multimodal applications

* User Modeling and Adaptation

* Multimodal Architectures, Tools and Standards

* Evaluation of Multimodal Interfaces

*Important Dates:*

Paper submission: May 23, 2008

Author notification July 14, 2008

Camera ready deadline: August 15, 2008

Conference: October 20-22, 2008

Organizing Committee

General Co-Chairs

Vassilis Digalakis, TU Crete, Greece

Alex Potamianos, TU Crete, Greece

Matthew Turk, UC Santa Barbara, USA

Program Co-Chairs

Roberto Pieraccini, SpeechCycle, USA

Jian Wang, Microsoft Research, China

Yuri Ivanov, MERL 

Invited Speakers:

·         Nestor Becerra (Universidad de Santiago, Chile)

Aplicaciones de las tecnologías del habla en sistemas CALL (Computer Aided Language Training) y CAPT (Computer Aided Pronunciation Training)

·         Giussepe Ricardi (University of Trento, Italy)

Next Generation Spoken Language Interfaces

·         Björn Granstrom (KTH - Royal Institute of Technology, Suecia)

Embodied conversational agents  in verbal and non-verbal communication

·         Yannis Stilianou (University of Crete, Grecia)

Voice Conversion: State of the art and Perspectives

Important dates:

·         Full paper submission: July 20, 2008

• Notification of acceptance: October 1, 2008
• Conference V JTH 2008: November 12-14, 2008

Contact information:

VJTH’2008

Dept. Electronics and Telecommunications

Faculty of Engineering

Alda. Urkijo s/n

48013 Bilbao

Tel.: +34 946 013 969

Fax.: +34 946 014 259

E-mail: 5jth@ehu.es           Web: http://jth2008.ehu.es

• Language recognition

Registration Form

• Speech synthesis

Registration Form

• Automatic text translation

Registration Form

http://jth2008.ehu.es/form_ALBAYZIN08_TA_en.pdf

Call for Papers/Demos
  
  2nd International Symposium on Universal Communication Dec 15 - 16,
  2008 Osaka International Convention Center, Osaka, Japan
  
  
  
  The development of information network systems enables us to
  communicate with people in remote places "anytime and anywhere",
  enriching human knowledge, affection and sensibility. However, there
  are various barriers which stand in our way to using these systems
  freely and flexibly. In order to discuss how to overcome these
  barriers and create a more human-centered communication environment,
  in 2007 the first International Symposium on Universal Communication
  was held in Kyoto, Japan featuring discussions with well-known
  researchers from around the world. Following its success, the second
  International Symposium on Universal Communication will be held in
  Osaka, Japan.
  
  
  
  Topics of interest are as follows, but not limited to
  
  　- Information retrieval and information analysis
  
  　- Information credibility
  
  　- Knowledge processing
  
  　- Language resources
  
  　- Speech recognition and synthesis
  
  　- Machine translation and speech translation
  
  　- Natural language processing
  
  　- Spoken language processing
  
  　- Multilingual information processing
  
  　- Super high-resolution image technology
  
  　- 3D visualization, imaging and display technologies
  
  　- 3D sound processing
  
  　- Virtual reality, mixed reality and augmented reality
  
  　- Multisensory (visual, acoustic, haptic, olfactory, etc) interaction
  
  　- Human factors
  
  　- Human interface and interaction technologies
  
  　- Real-world sensing technologies
  
  
  
  General Chair
  
  Yuichi Matsushima, National Institute of Information and
  Communications Technology (NICT), Japan
  
  
  
  General Vice Chairs
  
  Kazumasa Enami, NICT, Japan
  
  Hiromitsu Wakana, NICT, Japan
  
  
  
  Technical Program Committee
  
  Chair: Satoshi Nakamura, NICT/ATR, Japan Vice Chair: Naomi Inoue,
  NICT/ATR, Japan
  
  - Akio Ando, NHK Science & Technical Research Laboratories, Japan
  
  - Martin S. Banks, UC Berkeley, USA
  
  - Khalid Choukri, ELDA, France
  
  - Marcello Federico, FBK/IRST, Italy
  
  - Sidney Fels, The University of British Columbia, Canada
  
  - Jukka Häkkinen, University of Helsinki, Finland
  
  - Munpyo Hong, Sungkyunkwan University, Korea
  
  - Kentaro Inui, Nara Institute of Science and Technology, Japan
  
  - Hitoshi Isahara, NICT, Japan
  
  - Ken Kaneiwa, NICT, Japan
  
  - Takashi Kawai, Waseda University, Japan
  
  - Yutaka Kidawara, NICT, Japan
  
  - Kyeong Soo Kim, Swansea University, UK
  
  - Hisashi Miyamori, Kyoto Sangyo University, Japan
  
  - Makoto Okui, NICT, Japan
  
  - Tanja Schultz, Carnegie Melon University, USA
  
  - Yasuyuki Sumi, Kyoto University, Japan
  
  - Eiichiro Sumita, NICT/ATR, Japan
  
  - Yoiti Suzuki, Tohoku University, Japan
  
  - Yasuhiro Takaki, Tokyo University of Agriculture and Technology, Japan
  
  - Kazuya Takeda, Nagoya University, Japan
  
  - Kentaro Torisawa, NICT, Japan
  
  - Chiu-yu Tseng, Institute of Linguistics Academia Sinica, Taiwan
  
  - Kiyotaka Uchimoto, NICT, Japan
  
  - Takehito Utsuro, University of Tsukuba, Japan
  
  - Andy Way, Dublin City University, Ireland
  
  - Andrew Woods, Curtin University of Technology, Australia
  
  - Wieslaw Woszczyk, McGill University, Canada
  
  - Xing Xie, Microsoft Research Asia, China
  
  - Tatsuya Yamazaki, NICT, Japan
  
  - Kwon Yongjin, Korea AeroSpace University, Korea
  
  - Daqing Zhang, Institut TELECOM & Management SudParis, France
  
  
  
  Demo Chair
  
  Kazuhiro Kimura, NICT, Japan
  
  
  
  Technical Advisors
  
  Takashi Matsuyama, Kyoto University, Japan Michitaka Hirose, Tokyo
  University, Japan
  
  
  
  Local Arrangement Co-Chairs
  
  Kazuhiro Kimura, NICT,
  
  Yukio Takahashi, NICT
  
  Contact info: isuc2008@khn.nict.go.jp
  
  
  
  Website
  
  http://www.is-uc.org/2008/
  
  
  
  Venue
  
  Osaka International Convention Center
  
  3-51 Nakanoshima 5-chome,Kita-ku,Osaka, Japan
  http://www.gco.co.jp/english/english.html
  
  
  
  Submission Information
  
  All papers must be submitted through the ISUC 2008 homepage at
  http://www.is-uc.org/2008.
  
  
  
  Paper submission:
  
  The extended abstract must be written in English and is limited to 2
  pages in the IEEE 2-column format. This abstract should also indicate
  whether it is an oral paper or a poster, and be submitted in PDF. Once
  accepted, the camera-ready paper shall be 4~8 pages, not exceeding 8
  pages.
  
  More detailed author guidelines are available at
  http://www.is-uc.org/2008/submission.
  
  
  
  All accepted papers will appear in the conference proceedings
  published by the IEEE Computer Society and will be included in the
  IEEE-Xplore and the IEEE Computer Society (CSDL) digital libraries as
  well as indexed through IET INSPEC, EI (Compendex) and Thomson ISI.
  
  
  
  Demo submission:
  
  Please submit a one-page description of your demo in PDF. This
  description must be written in English and should include: An abstract
  of what you will show, Space needed, Facilities needed including power
  supply and Internet access. A specified submission format will be
  available on the ISUC 2008 homepage.
  
  
  
  Important Dates
  
  Papers:
  
  Extended Abstract due          July 25, 2008
  
  Notification of acceptance      August 29, 2008
  
  Camera-ready papers due      September 26, 2008
  
  
  
  Demos:
  
  Submission of description due September 19, 2008
  
  Notification of acceptance      October 10, 2008
  
  
  
  
  

Consonant Challenge for Indian Languages

Detection and Recognition of Consonants in Indian Language Speech Data

 Call for Participation (linked to SLT 2008)

In order to promote speech technology research in Indian Languages and to better understand any specific issues related to speech recognition of these languages and the possible means to address them, we are pleased to announce a Consonant Challenge in Indian Languages. The task involves detection of consonants (in CV, VC, CVC and VCV positions) in a surprise language. Training data is provided in 6 Indian languages, namely, Assamese, Bengali, Hindi, Marathi, Tamil and Telugu to all registered participants.

Based on the recognition results received by the organizers and evaluated by the program committee, the highest two accuracy results will be awarded a cash prize of USD 500 and USD 250 respectively.

The results will be presented in a special session at SLT 08 in Goa, India.

Background

Consonant detection in speech by a machine based on purely spectral features is always problematic due to a number of reasons like the unvoiced (no-energy) portions of stop consonants that can be confused with real silence, the high energy fricative noise that maybe confused with environmental or additive noise, and the vowel like spectrum of the liquids, the nasals and the semi-vowels that make them hard to distinguish from vowels. This problem is further compounded in Indian languages where the number of consonants can go from around 23 (in Tamil) to almost 40 (in Hindi-Urdu).  For example, acoustic phonetic features like voice and aspiration form a four way contrast in many Indian language stop and affricate consonants. Further, stop consonants occur for at least four, that is, labial, dental, retroflex, and velar, place of articulation (this can go to 5 or 6 for some languages like Malayalam and Hindi-Urdu). Though all Indian Languages come from four major language families (Indo-Aryan, Dravidian, Austronesian and Tibeto-Burman, with the majority from the former two), the languages have co-existed for a long time to have borrowed and shared features even at the phonetic level. For example, the borrowing of retroflex sounds from Dravidian to Indo-European and of aspiration as a feature of stops the other way around.

From a Speech Recognition perspective, a deeper understanding of how consonants are detected and recognized can not only help us better understand how to model these sounds (ref. difference between human and computer consonant recognition) but also, in the specific case of Indian languages, open up research issues  into model adaptation from one language to another (related)language. This might allow researchers to explore ways and means to scale from one language to another where resources in terms of training data are limited

Register

* To register for the CCIL, please mail the organizing chairs by 25th August 2008.

* Data will be released to the registered participants ONLY

* All participants for the CCIL will have to register for the main SLT08 workshop.

Important Dates

* Registration for CCIL : 25th August 2008

* Training Data release to registered participants: 29th August 2008

* Test Data in surprise language made available:  8th September 2008

* Recognition results and paper submission: 29th September 2008

* Results Announced: 20th October 2008

* Camera-ready paper submission: 3rd November 2008

Contact

Please mail the organizing chairs to register for the challenge at:

ramkiag@ee.iisc.ernet.in

kalikab@microsoft.com

http://ragashri.ee.iisc.ernet.in/ILCC

 2008 International Symposium on

         Chinese Spoken Language Processing (ISCSLP 2008)

                          December 16 - 19, 2008

                              Kunming, China

                       http://www.iscslp2008.org

ISCSLP’08 is the flagship conference of ISCA SIG-CSLP (Special Interest Group on Chinese Spoken Language Processing).

ISCSLP'08 will be held during December 16-19, 2008 in Kunming hosted by The University of Science and Technology of China and Yunnan University.

ISCSLP (International Symposium on Chinese Spoken Language Processing) is a conference for scientists, researchers, and practitioners to report and discuss the latest progress in all scientific and technological aspects of Chinese spoken language processing (CSLP). The idea of having a series of regular conferences devoted to CSLP was an outcome of a small-group meeting held in December 1997 in Singapore. The meeting was organized and chaired by Professor Chin-Hui Lee, then worked at Bell Laboratories, USA; and attended by Professors Tai-Yi Huang and Ren-Hua Wang from mainland China, Professors Chorkin Chan and Pak-Chung Ching from Hong Kong, Professor Kim-Teng Lua and Dr. Haizhou Li from Singapore, and Professors Lin-Shan Lee and Hsiao-Chuan Wang from Taiwan. A Steering Committee, being chaired by Professor Chin-Hui Lee and consisting of the abovementioned nine members, was established to oversee the ISCSLP conferences. It was decided that a bi-annual symposium will be organized and hosted initially by research groups from Asia Pacific regions. Since its inception, ISCSLP has become the world's largest and most comprehensive technical conference focused on Chinese spoken language processing and its applications. In ISCSLP 2002, a special interest group was formed as SIG-CSLP of International Speech Communication Association (ISCA). ISCSLP is now an ISCA and IEEE supported event.

We invite your participation in this premier conference, where the language from ancient civilizations embraces modern computing technology. The ISCSLP'08 will feature world-renowned plenary speakers, tutorials, exhibits, and a number of lecture and poster sessions. The concrete version is attached to this mail.

In response to popular requests from authors, the paper submission deadline is extended. The new deadline is Jul 29, 2008.

The Keynote speakers of ISCSLP2008 is as following:

Qiang Huo

Microsoft Research Asia, Beijing, China

Research Area

Automatic speech & speaker recognition and related multidisciplinary research topics

Chinese character recognition

Biometric authentication

Document analysis and recognition

Machine learning, etc.

Shigeki Sagayama

Department of Information Physics and Computing Graduate School of Information Science and Technology, The University of Tokyo, Japan

Research Area

Speech and spoken language processing

Signal processing

Music signal/Information processing

Hand-written character recognition

Multimedia Information processing, etc.

Vincent Vanhoucke

Google, USA

Research Area

Software engineering

Text recognition

Speech recognition

Image processing

Face recognition, etc.

Yuqing Gao

IBM T. J. Watson Research Center, USA

Research Area

Speech recognition

understanding and Translation Research

The large vocabulary continuous speech dictation system

Speech-to-speech translation research, etc.

Hideki Kawahara

Design Information Sciences Department, Faculty of Systems Engineering, Wakayama University, Japan

Research Area

Focus on the use of STRAIGHT in research on human speech perception

Signal processing models of hearing and neural networks

Interaction between speech perception and production, etc.

Yu Hu

Research Director of iFLYTEK, Hefei, China

Research Area

Speech pronunciation evaluation

Speech pronunciation defect detection, etc