IEEE SPS Speech and Language Technical Committee E-Newsletter

Welcome!

Welcome to the July 2006 issue of the IEEE Signal Processing Society Speech and Language Technical Committee (SLTC) e-Newsletter. It has been an exciting spring, as we have just completed another successful ICASSP conference in Toulouse, France. This issue includes articles about the latest news and upcoming events in the spoken language research community, as well as a letter to all IEEE members regarding the upcoming elections for the Signal Processing Society's Board of Governors.

We would like to thank Isabel Trancoso and Dilek Hakkani-Tur for their contributions to this issue of the newsletter.

As always, we welcome your contributions of news, events, publications, workshops, and career information to the newsletter.  Please send all articles, ideas, and feedback to the SLTC e-Newsletter Editorial Board [speechnewseds <at> ieee <dot> org]. Archives of recent SLTC e-Newsletters can be found on the SLTC website. Instructions to subscribe or unsubscribe to the eNewsletter distribution can be found here.

Have a great summer!

The SLTC e-Newsletter Editorial Board
Mike Seltzer, Brian Mak, and Gokhan Tur
[speechnewseds <at> ieee <dot> org]
 

HeadLINES
Highlights from Speech and Language Technical Committee Meeting at ICASSP 2006
A Letter From Isabel Trancoso: Please Vote in the Upcoming Elections for Member-at-Large of IEEE Signal Processing Society's Board of Governors
IEEE SPS Awards Announced for Speech and Language Researchers
AT&T, ICSI, and Edinburgh University Announce DiSCoH (Spoken Dialogue System for Conference Help)
NIST Announces New Evaluation Initiative in Spoken Term Detection
How to Subscribe or Unsubscribe to SLTC E-Newsletter

Conference and Workshop Announcements

Call for Papers:
IEEE/ACL 2006 Workshop on Spoken Language Technology
Special Issue of the IEEE Transactions on Computers: Emergent Systems, Algorithms and Architectures for Speech-based Human-Machine Interaction

Call for Participation:
NATO Advanced Study Institute XI Course on The Fundamental of Verbal and Non-verbal Communication and the Biometrical Issue
INTERSPEECH 2006: International Conference on Spoken Language Processing
IEEE International Workshop on Multimedia Signal Processing
 

CAREER CENTER

Positions Available:

Machine Translation at SRI International Speech Technology and Research Laboratory
Speech Processing at the IDIAP Research Institute, Switzerland - Senior Researcher
Adaptive Multimodal Interface Research Lab at University of Trento - Multiple Positions

Transitions:
Researchers take new positions

back to top

Highlights from Speech and Language Technical Committee Meeting at ICASSP 2006
by Brian Mak

The Speech and Language Technical Committee (SLTC) of the IEEE Signal Processing Society met on May 18, 2006 during ICASSP 2006 in Toulouse, France. Below is a summary of some major topics  that were discussed.

• ICASSP Summary: ICASSP this year received a record high of 668 papers in the fields of speech and language technologies, not including the special session papers. This is a historical record high in these areas for this conference, and amounts to over 18% increase from last year. The majority of this increase is attributed to more papers on Robustness, Speech Recognition and Synthesis, and Language Processing. Nearly all papers received three reviews from experts in this community. The review committee involved 10 Area Chairs, 48 Technical Committee members and 180 external reviewers. The acceptance rate for this year is about 46%, which is less than the 50% we had last year.
• Growth of SLTC: With the increased scope of the Speech TC to now cover language as well, the SLTC will expand from the current membership of 30 (excluding AE’s) to approximately 45. New members are needed in the areas of language and multimodal processing.
• New member election: Taking into account the number of retiring SLTC members (7), the number of ICASSP reviewers needed this year, and the projected growth of ICASSP paper submissions, the SLTC will need to recruit approximately 16 new members this year though they may not be recruited all at once. The nomination process will start this summer.
• The Conference Board:  The conference board met at ICASSP and discussed the future venues of ICASSP and ICIP. The venues for ICASSP have been decided through 2009. They are:
  • ICASSP 2007 in Honolulu, Hawaii
  • ICASSP 2008 in Las Vegas, Nevada
  • ICASSP 2009 in Taiwan
• IEEE/ACL Workshop on Spoken Language Technology:  (http://www.slt2006.org) will be held in Aruba from December 10-13 this year. It aims to bring the speech processing and natural language processing communities together to share and present recent research and advances in the area of spoken language technology.
• ASRU Proposals: There were two proposals submitted for ASRU 2007: one from Giuseppe Di Fabbrizio (AT&T), Juan M. Huerta (IBM), and Ingrid Kirschning (UDLA-P) to hold it in Puebla, Mexico, and the other from Sadaoki Furui (Tokyo Inst. Tech.), Satoshi Nakamura (ATR), and Tatsuya Kawahara (Kyoto Univ.) to hold it in Kyoto, Japan. A decision will be made in the Fall.
• IEEE Transactions on Audio, Speech, and Language Processing: Mari Ostendorf, the new Editor-in-Chief of IEEE Transactions of Audio, Speech, and Language Processing (TASLP) talked about her vision of TASLP. The journal is doing well in terms of submissions and will be moving to 8 issues per year next year. She is actively soliciting papers in language processing with the expanded scope of the journal and is hoping for a new special issue in that area to be announced soon. She is working on improving the impact factor of the journal, in part by fixing the problem that ICASSP and Interspeech conference proceedings are not indexed by SCI and many citations of TASLP papers are from these conferences. Turn-around time is also a factor in impact and of concern to authors, and her goal for 2007 targets time from submission to publication at less than 12 months. She encouraged all colleagues in the field to submit papers to the journal.
• SLTC Chair: Mazin Gilbert's term as the Chair of SLTC will end this year. New Chairperson election will be held in September and the new chair will start his/her role in January 2007 for a term of 2 years.

back to top

Please Vote in the Upcoming Elections for Member-at-Large of IEEE Signal Processing Society's Board of Governors
by Isabel Trancoso

Dear All,

Every Summer all members of the IEEE Signal Processing Society are asked to elect members-at-large for the Board of Governors (BoG). Unfortunately, most members don't even bother to vote, either because they don't understand the role of the BoG, or they don't know the candidates.

So the purpose of this message is to try to explain how important it will be for all the people working in spoken language processing to vote.

Many of the decisions the IEEE Signal Processing Society makes are passed for a vote in the BoG, so if you vote for somebody working in your field, you increase the chances of your field being better represented in the Society. That is visible for instance in terms of major Awards, changes in the society publications (page budgets of the Transactions, electronic pre-publication on Xplore, etc.), structure of conferences (special sessions and tutorials devoted to spoken language processing), number of fellows, structure of technical committees and councils, etc.

Spoken language processing has a very big weight in the society in terms of the number of papers presented at ICASSP (20 to 25%, typically), and number of members (possibly three to four thousand). But this weight is not at all reflected in the committees where there is one representative per area. Here the Speech and Language Technical Committee is only ONE OUT OF 12. And it is not at all reflected in the current members of the BoG.

So this message is an incentive for the speech and language community to be more active in the society and start by voting for representatives of their field in the coming BoG elections.

This time our candidate is Mazin Gilbert (formerly Rahim). Mazin has done such a terrific work as Chair of the Speech and Language Technical Committee that he does not need any introduction from me. He has been one of the strongest supporters of increasing the weight and visibility of the speech and language area within IEEE, he has done a tremendous work conducting the ICASSP review process in the last couple of years, promoting special sessions, and tutorials, setting up new workshops and many new activities within the STC.

I'm sure he will be the perfect representative of the speech and language community within the BoG. So don't throw your ballot away!

Isabel Trancoso
Member-at-Large
Board of Governors
IEEE Signal Processing Society

back to top

IEEE SPS Awards Announced for Speech and Language Researchers
by Michael L. Seltzer

For a complete list of the SPS award winners in all areas, please see http://www.ieee.org/organizations/society/sp/awards.html.For a complete list of the student winners in all areas, please see http://www.icassp2006.org/SPCWinners.asp

Congratulations to all award winners!!!

back to top

AT&T, ICSI, and Edinburgh University Announce DiSCoH (Spoken Dialogue System for Conference Help)
by Dilek Hakkani-Tur

DiSCoH (Spoken Dialogue System for Conference Help) is a goal-oriented, mixed-initiative, human-machine spoken dialogue system for conference information, developed by researchers at AT&T Labs, ICSI, and Edinburgh University. The main goal of of DiSCoH is to collect naturally spoken human-machine dialogs. All of these dialogues will be logged by DiSCoH using rich representations of context and all data collected with the system will be released to the research community. It is first being deployed for the IEEE SLT Workshop 2006, see http://www.slt06.org

Soon you will be able to call +1 888 687 3887 (toll free for US callers, please check the project web page for the availability date of the system) to access DiSCoH for information about the conference including locations, dates, paper submission deadlines, program, venue, paper status, accommodation options and costs, etc.

The central motivation for this system is that the lack of large, richly annotated dialogue corpora from real spoken dialogue applications is a major barrier to progress in the research community. We believe that such a corpus, for example annotated with speech acts, user utterance transcriptions, user intentions, overall task success, etc., will be an essential resource for research in dialogue management, spoken language understanding, automatic speech recognition, and related tasks.

Please help us to collect the DiSCoH spoken dialogue corpus!

More information about DiSCoH is available at: http://www.icsi.berkeley.edu/~dilek/DisCoH/index.html

back to top

NIST Announces New Evaluation in Spoken Term Detection
by Michael L. Seltzer

NIST has announced a new evaluation initiative in Spoken Term Detection (STD). This evaluation is designed to facilitate research in the area of information retrieval from speech data. Initially, the evaluation will focus on finding short word sequences rapidly and accurately across large heterogeneous collections of audio. Specifically, the STD task will be to create an index from a speech corpus and then find all occurrences, if any, of specified terms in that corpus. The search terms used for evaluation will not be known in advance.  The test corpus will include data from broadcast news, telephone conversations, and meetings, and will be in English, Chinese, and Arabic. System performance will be measured based on accuracy, speed, and robustness. The schedule for the STD evaluation is as follows:

Evaluation Schedule:

• Release of development data:     July 2006

• Dry-run evaluation:                     September 2006

• Evaluation:                                 November 2006

A detailed description of the evaluation plan is located on the STD website http://www.nist.gov/speech/tests/std/. Those interested in participating in the evaluation or interested in this research area can join the STD interest group mailing list by sending email to STD-info@nist.gov.

 back to top

How to Subscribe to or Unsubscribe from the SLTC e-Newsletter
by Michael L. Seltzer

We have created a listserv hosted by IEEE for the distribution of the SLTC e-Newsletter, [ speechnewsdist <at> listserv <dot> ieee <dot> org ]. This list is intended for the purpose of disseminating news and information pertaining to the IEEE SPS Speech and Language Technical Committee (SLTC), and in particular for distributing the electronic newsletter of the SLTC. To receive the newsletter and any other related announcements, you can simply subscribe to the distribution list (tell your friends and colleagues!). To no longer receive the newsletter, you can unsubscribe from the list. Note that if you have received this issue (June 2006) by email, you are currently already subscribed to the list. If you wish to remain subscribed, no action needs to be taken.

To Subscribe:
Send an email with the command "subscribe speechnewsdist" in the message body to [ listserv <at> listserv <dot> ieee <dot> org ].

To Unsubscribe:
Send an email with the command "signoff speechnewsdist" in the message body to [ listserv <at> listserv <dot> ieee <dot> org ].

Note that subscribers cannot post to this distribution list. Please send all contributions, articles, ideas, and feedback to the SLTC e-Newsletter Editorial Board [ speechnewseds <at> ieee <dot> org ].

back to top

The first workshop on Spoken Language Technology (SLT) sponsored by IEEE and ACL will be held on December 10-December 13, 2006. The goal
of this workshop is to bring the speech processing and natural language processing communities together to share and present recent
advances in the area of spoken language technology, and to discuss and foster new research in this area. Spoken language technology is a
vibrant research area, with the potential for significant impact on government and industrial applications.

Workshop Topics

• Spoken language understanding
• Spoken document summarization
• Machine translation for speech
• Spoken dialog systems
• Spoken language generation
• Spoken document retrieval
• Human/Computer Interactions (HCI)
• Speech data mining
• Information extraction from speech
• Question/Answering from speech
• Multimodal processing
• Spoken language systems, applications and standards.

Submissions for the Technical Program

The workshop program will consist of tutorials, oral and poster presentations, and panel discussions. Attendance will be limited with
priority for those who will present technical papers; registration is required of at least one author for each paper. Submissions are
encouraged on any of the topics listed above. The style guide, templates, and submission form will follow the IEEE ICASSP
style. Three members of the Scientific Committee will review each paper. The workshop proceedings will be published on a CD-ROM.

Schedule

• Camera-ready paper submission deadline July 21, 2006
• Hotel Reservation and Workshop registration opens July 30, 2006
• Paper Acceptance / Rejection September 1, 2006
• Hotel Reservation and Workshop Registration closes October 15, 2006
• Workshop December 10-13, 2006

Registration and Information

Registration and paper submission, as well as other workshop information, can be found on the SLT website: http://www.slt2006.org

 Organizing Committee

• General Chair: Mazin Gilbert, AT&T, USA
• Co-Chair: Hermann Ney, RWTH Aachen, Germany
• Finance Chair: Gokhan Tur, SRI, USA
• Publication Chair: Brian Roark, OGI/OHSU, USA
• Publicity Chair: Eric Fosler-Lussier, Ohio State U., USA
• Industrial Chair: Roberto Pieraccini, Tell-Eureka, USA

back to top

Call for Papers:

Special Issue of the IEEE Transactions on Computers:
Emergent Systems, Algorithms and Architectures for Speech-based Human-Machine Interaction

In recent years, there has been a significant progress on software and hardware techniques for humanmachine interaction, specially, for speech-based one. While highly successful in the core and many commercial applications have resulted from the progress, the techniques have a long way to improve, specially in areas of speech signal processing, speech/audio source separation, robustness of speech systems, speech modeling, learning algorithms, real-time decoding algorithms, speech quality enhancement, speech synthesis, speaker identification, and so on, for further impacting on speech-centric human-machine interaction. This special issue intends to bring such techniques among the researchers in the field and promote new methods that may prove useful for further research in this area.

Topics: Papers showing mature results of research, advancing the state-of-the-art in terms of systems, algorithms or architectures for speech-based human machine interaction, are expected, particularly involving:

• speech recognition
• speaker identification
• speech enhancement
• speech and audio source separation
• speech synthesis
• voice morphing
• methods to assist users in speech production and in processing the auditory scene,
• other related applications.

Submission procedure: Please, submit your paper to Manuscript Central at http://cs-ieee.manuscriptcentral.com/, selecting this special issue's title. Submitted articles must not have been previously published or currently submitted for journal publication elsewhere. As an author, you are responsible for understanding and adhering to our submission guidelines. You can access them by clicking on http://www.computer.org/mc/tc/author.htm. Please thoroughly read these before submitting your manuscript. Feel free to contact the Peer Review Manager, Suzanne Werner at swerner@computer.org or the guest editors at guido@ifsc.usp.br , deng@microsoft.com or maki@cslab.kecl.ntt.co.jp , if you have any questions.

Important dates:

• Submission Deadline:.......................................... September 12, 2006.
• Completion of First Round of Reviews:............... November 10, 2006.
• Major Revisions Due:......................................... January, 10, 2006.
• Completion of Second Round of Reviews:...........February 12, 2006.
• Minor Revisions Due:......................................... February 26, 2006.
• Final Decisions Due:.......................................... March 5, 2006.
• Acceptance Letters Sent to Authors:.................. March 7, 2007
• Publication Materials Due:.................................. March 23, 2007.
• Tentative Publication Date:................................. August, 2007.

GUEST EDITORS

Rodrigo Capobianco Guido University of São Paulo, Brazil. e-mail: guido@ifsc.usp.br
Li Deng Microsoft Research, USA. e-mail: deng@microsoft.com
Shoji Makino NTT CS Labs, Japan. e-mail: maki@cslab.kecl.ntt.co.jp

back to top

back to top

INTERSPEECH 2006 - ICSLP, the Ninth International Conference on Spoken Language Processing dedicated to the interdisciplinary study of speech science and language technology, will be held in Pittsburgh, Pennsylvania, September 17-21, 2006, under the sponsorship of the International Speech Communication Association (ISCA).

The Interspeech meetings are considered to be the top international conference in speech and language technology, with more than 1000 attendees from universities, industry, and government agencies. They are unique in that they bring together faculty and students from universities with researchers and developers from government and industry to discuss the latest research advances, technological innovations, and products. The conference offers the prospect of meeting the future leaders of our field, exchanging ideas, and exploring opportunities for collaboration, employment, and sales through keynote talks, tutorials, technical sessions, exhibits, and poster sessions. In recent years the Interspeech meetings have taken place in a number of exciting venues including most recently Lisbon, Jeju Island (Korea), Geneva, Denver, Aalborg (Denmark), and Beijing.

In addition to the regular sessions, a series of special sessions has been planned for the meeting. Potential authors are invited to submit papers for special sessions as well as for regular sessions, and all papers in special sessions will undergo the same review process as papers in regular sessions. Confirmed special sessions and their organizers include:

•  The Speech Separation Challenge, Martin Cooke (Sheffield) and Te-Won Lee (UCSD)
•  Speech Summarization, Jean Carletta (Edinburgh) and Julia Hirschberg (Columbia)
•  Articulatory Modeling, Eric Bateson (University of British Columbia)
•  Visual Intonation, Marc Swerts (Tilburg)
•  Spoken Dialog Technology R&D, Roberto Pieraccini (Tell-Eureka)
•  The Prosody of Turn-Taking and Dialog Acts, Nigel Ward (UTEP) and Elizabeth Shriberg (SRI and ICSI)
•  Speech and Language in Education, Patti Price (pprice.com) and Abeer Alwan (UCLA)
•  From Ideas to Companies, Janet Baker (formerly of Dragon Systems)

IMPORTANT DATES

• Early registration deadline: June 23, 2006
• Tutorial Day: September 17, 2006
• Main Conference: September 18-21, 2006

For further information: http://www.interspeech2006.org or send email to info@interspeech2006.org

Organizer:
Professor Richard M. Stern (General Chair)
Carnegie Mellon University
Electrical Engineering and Computer Science
5000 Forbes Avenue
Pittsburgh, PA 15213-3890
Fax: +1 412 268-3890
email: chair@interspeech2006.org

back to top

Position Available:

RESEARCHER IN MACHINE TRANSLATION
SRI INTERNATIONAL SPEECH TECHNOLOGY AND RESEARCH LABORATORY

 The Speech Technology and Research (STAR) Laboratory at SRI International seeks a self-motivated, team-oriented researcher in machine translation. Highly qualified postdoctoral fellows may also apply.

The STAR Laboratory is engaged in leading-edge research in speech recognition, automatic spoken language translation, speaker recognition and verification, human-machine interfaces, and other areas of speech/language technology, and offers opportunities for basic research as well as prototyping and collaborative productization.  For further details about the SRI STAR Lab please see http://www.speech.sri.com. 

The successful candidate will have the opportunity to work on multiple government-funded research projects and to collaborate with other researchers at SRI and partner institutions.  S/he will work on high-performance, deployable machine translation systems for multiple language pairs with varying levels of resources. A PhD with machine translation background is desired. The candidate must have strong engineering capability, with skills in C/C++ and scripting languages in a Unix/Linux environment. Strong oral and written communication skills are expected. Experience in previous NIST MT evaluations and knowledge of speech recognition are highly desirable.

Candidates must be able to work both independently and cooperatively across multiple projects with dynamically forming teams.  Characteristics of STAR staff are enthusiasm, self-motivation, initiative, and passion for learning. Please apply online via https://sri.ats.hrsmart.com/cgi-bin/a/highlightjob.cgi?jobid=3027

back to top

Position Available:

Senior Researcher position in Speech Processing
at the IDIAP Research Institute, Switzerland

The IDIAP Research Institute (www.idiap.ch) is currently seeking one talented senior researcher in speech processing and/or machine learning with a proven record of high level research and project management, and whose interests are aligned with our existing strengths in speech processing (speech and speaker recognition, language understanding), or in related areas such as computer vision, biometric authentication, and multimedia data mining. A candidate with these strengths will be expected to play a leading role in the research, teaching and strategic development of the Institute.

Most of IDIAP's research activities take place in the framework of National  long term research initiatives such as the National Centre of Competence in Research (NCCR) on “Interactive Multimodal Information Management”  (IM2, see www.im2.ch) or large European projects such as “Augmented Multi-party Interaction” (AMI, see www.amiproject.org).

Specific Knowledge/Skills:

• PhD in electrical engineering, computer science, signal processing, mathematics, statistics, or relevant discipline with a minimum of 5 years (post-PhD) experience.

• Strong record of research and innovation in the area of speech processing, pattern recognition and machine learning (incl. HMM, ANN, SVM, kernel approaches, Gaussian processes).

• Significant experience in software development, including C/C++.

• Experience with large real world audio and video data sets is desirable.

• Experience in project management and supervision of researchers, including PhD students.

• Good organizational and communication skills, both written and oral.

• Ability to interact well with international, multi-disciplinary, R&D teams.

Initiated in 1991, and supported by the Swiss Federal Government, the State of Valais, and the City of Martigny, IDIAP (www.idiap.ch) is an independent, nonprofit research institute located in Martigny (at the edge of the Swiss Alps), and is affiliated with EPFL (Ecole Polytechnique Federale de Lausanne) and the University of Geneva. See http://www.idiap.ch/pages/press/faq.pdf for more information and FAQs.

IDIAP is building a culturally diverse research community and strongly encourages applications from female and minority candidates. Prospective candidates should apply with a cover letter, CV, statement of research interests and accomplishments, and names and email addresses of 3 references. Please send these to jobs@idiap.ch, with a clear reference in subject header to “senior position in speech processing”). More information can also be obtained by contacting Prof. Hervé Bourlard (bourlard@idiap.ch), Director of IDIAP.

Start dates are flexible, but applications received by April 30, 2006, will receive full consideration.

back to top

Positions Available:
 Adaptive Multimodal Interface Research Lab at University of Trento (Italy)

The Adaptive Multimodal Interface Research Lab at University of Trento (Italy) has openings in the following areas:

• Automatic Speech Recognition (PhD Research Fellowship)
• Natural Language Processing (PhD Research Fellowship)
• Machine Learning (PhD Research Fellowship/Senior Researcher)
• HCI/User Interface (Junior Researcher)
• Multimodal/Spoken Dialog (Senior Researcher)

The Adaptive Multimodal Interface research lab pursues excellence research in next-generation interfaces for human-machine and human-human communication. The research positions will be funded by the prestigious Marie Curie Excellence grant awarded by the European Commission for cutting edge and interdisciplinary  research. The candidates for PhD research fellowships should have background in speech, natural language processing or machine learning. The successful applicants should have EE or CS degree with strong academic records. The students will be part of an interdisciplinary research team working on speech recognition, language understanding, spoken dialog, machine learning and adaptive user interfaces. Deadline for application submission is July 11, 2006 (see http://ict.unitn.it/). The candidates for the junior/senior researcher positions should have a PhD degree either in computer science, cognitive science or related disciplines. They will have an established international research track record in their field of expertise and leadership skills. Deadline for application submission is November 1, 2006.
 

The applicants should be fluent in English. The Italian language competence is optional and applicants are encouraged to acquire this skill on the job. The applicants should have good programming skills in
most of the following C++/Java/JavaScript/Perl/Python.Salaries are competitive and depending on qualifications. Relocation package might be available depending on eligibility. University of Trento is an equal opportunity employer. Interested applicants should send their CV along with their statement of research interest and three reference letters to:


Prof. Ing. Giuseppe Riccardi
Email: riccardi@dit.unitn.it
http://www.dit.unitn.it/~riccardi

About University of Trento and Information and Communication Technology Department (DIT)
The University of Trento is constantly ranked as premiere Italian graduate university institution (see www.dit.unitn.it).
DIT Department

• DIT has a strong focus on interdisciplinarity with professors from different faculties of the University
  (Physical Science, Electrical Engineering, Economics, Social Science, Cognitive Science, Computer Science)
  with international background.
• DIT aims at exploiting the complementary experiences present in the various research areas in order to
  develop innovative methods and technologies, applications and advanced services.
• English is the official language.
   

back to top

TRANSITIONS

The STC Newsletter would like to provide announcements of  professors, researchers, and developers in the speech and language community taking new positions.  If you have moved lately or are in the process of moving  to a new position in the near future, send us  your new contact information so it can be posted in the next edition.  

back to top