March 13, 2004

INTRODUCTION:

Welcome to the IEEE Signal Processing Society Speech Technical Committee (STC) newsletter.  As always, contributions of events, publications, workshops, and career information to the newsletter are welcome.  Please send to Rick Rose (rose@ece.mcgill.ca).   Archives of recent STC Newsletters can be found on the STC website. 

LINKS TO WORKSHOPS AND CONFERENCES:
Links to conferences and workshops organized by date  (Rick Rose)

Summary of ICASSP 2005 Submissions from the

Conference Technical Program Chairs

2004 IEEE Signal Processing Awards

2004 was a very successful year for the Speech area. Four out of the seven SPS awards were won by Speech researchers.  In 2004, the Speech Technical Committee (STC) formed an Awards Subcommittee to coordinate the process of nominating candidates for the three Paper Awards and the four Major Awards. The committee members were Ramesh Gopinath, Li Deng, Alan Black, Kazunori Mano, Isabel Trancoso, and Ananth Sankar. The STC Awards Committee received a total of 23 nominations for the 7 categories from 12 individual nominators. This was followed by a vote within the STC to choose the final 7 nominations for the Speech area. The final nominations were then reviewed and revised by the Awards Subcommittee before submission to the Awards Board. Particular care was taken to highlight the contributions of the final nominees.

Speech won two of the three Paper Awards, and two of the four Major Awards. There were multiple winners in some categories. The Speech winners are listed below. The total number of winners in each category is in parenthesis.

 Technical Achievement Award (2):  Prof. Steve Young

Meritorious Service Award (1):  Prof. Andreas Spanias

Young Author Paper Award (2):  G. Tzanetakis and P. Cook, Musical genre classification of audio signals, IEEE Trans. on Speech and Audio Processing, vol. 10, pp. 293-302, July 2002. The young author was George Tzanetakis

Best paper award (4): E. Bocchieri and B. K.W. Mak, Subspace distribution clustering hidden Markov model, IEEE Trans. on Speech and Audio Processing,
vol. 9, pp. 264-275, March 2001.

The STC takes great pride in the achievement of our Speech winners, and extends our heartiest congratulations to them.
top of page

Special Issue of
The IEEE Transactions on Speech and Audio Processing
Progress in Rich Transcription

Over the past several years, Rich Transcription has emerged as an interdisciplinary field combining automatic speech recognition, speaker identification, and natural language processing with the goal of producing richly annotated speech transcriptions that are useful both to human readers and to automated programs for indexing, retrieval and analysis. The key problems include making more accurate speech transcription technology; improving speaker recognition technology; developing fundamentally new techniques for annotating dialog with semantic intent; and enriching ASR output to present it in a maximally informative manner. These various goals interact with each other, and exploiting synergistic uses of the disparate forms of analysis is critical. With its focus on fundamental research in human communication, Rich Transcription is key to governmental applications in data mining, and to commercial applications such as call center automation and monitoring.

The purpose of this special issue is to present recent advances in all areas of Rich Transcription for Speech, Audio, and Spoken Language Dialog. Original, previously unpublished submissions for the following areas are encouraged:

• Speech Recognition algorithms and methods
• Natural Language Processing for Rich Transcription
• Speaker Recognition algorithms and methods
• Algorithms for exploiting large amounts of training data
• Novel Approaches to feature extraction and ASR
• Unsupervised and semi-supervised training
• Paradigms for Data Collection
• Usability and human factors studies
• Performance analysis and evaluation
• Applications in marketing, business & security
• Tools and solutions for rich transcription
• Incorporation of prosodic and supralexical information 

Prospective authors should prepare manuscripts according to the Information for Authors as published in any recent issue of the Transactions and as available on the web at http://www.ieee.org/organizations/society/sp/infotsa.html. Note that all rules will apply with regard to submission lengths, mandatory overlength page charges, and color charges.

Manuscripts should be submitted electronically through the online IEEE manuscript submission system at http://sps-ieee.manuscriptcentral.com/. When selecting a manuscript type, authors must click on "Special Issue of T-SA on Progress in Rich Transcription." Authors should follow the instructions for the IEEE Transactions on Speech and Audio Processing and indicate in the Comments to the Editor-in-Chief that the manuscript is submitted for publication in the Special Issue on Progress in Rich Transcription. We require a completed copyright form to be signed and faxed to 1-732-562-8905 at the time of submission. Please indicate the manuscript number on the top of the page.

Schedule:

Submission deadline:	1 October 2005
Notification of acceptance:	1 April 2006
Final manuscript due:	31 May 2006
Tentative publication date:	September 2006

Special Issue of
The IEEE Transactions on Speech and Audio Processing
Expressive Speech Synthesis

Expressive Speech Synthesis (ESS) is a multidisciplinary research area that addresses one of the most complex problems in speech and language processing. The challenges posed by ESS have been the subject of several collaborative research projects across universities and laboratories around the world. Over the last decade ESS has benefited from advances in speech and language processing as well as from the availability of large conversational-speech databases. These advances have spurred research on the expressiveness of speech and on conveying paralinguistic information including emotion, speaker-state, and speaker-listener relationships. There have also been substantial efforts towards automating database creation and evaluating the quality of speech synthesised for a variety of tasks that require not just the transmission of information, but also the expression of affect.

The purpose of this special issue is to present recent advances in Expressive Speech Synthesis. Original, previously unpublished research is sought in all areas relevant to the field. In particular, submissions on theory and methods for the following areas are encouraged:

• Signal processing techniques for ESS
• Speech corpus design for ESS
• Input text processing and annotation for ESS
• Prosody processing in ESS
• Voice quality issues in ESS
• Unit selection algorithms for ESS
• Unit concatenation algorithms for ESS
• Expression of affect and emotion in ESS
• ESS in systems and applications
• Multilingual and appplication-specific ESS
• Expressive speech annotation and labelling
• Evaluation of expressivity in speech

Prospective authors should prepare manuscripts according to the Information for Authors as published in any recent issue of the Transactions and as available on the web at http://www.ieee.org/organizations/society/sp/infotsa.html. Note that all rules will apply with regard to submission lengths, mandatory overlength page charges, and color charges.

Manuscripts should be submitted electronically through the online IEEE manuscript submission system at http://sps-ieee.manuscriptcentral.com/. When selecting a manuscript type, authors must click on "Special Issue of T-SA on Expressive Speech Synthesis." Authors should follow the instructions for the IEEE Transactions on Speech and Audio Processing and indicate in the Comments to the Editor-in-Chief that the manuscript is submitted for publication in the Special Issue on Statistical and Perceptual Audio Processing. We require a completed copyright form to be signed and faxed to 1-732-562-8905 at the time of submission. Please indicate the manuscript number on the top of the page.

Schedule:

Submission deadline:	1 June 2005
Notification of acceptance:	1 December 2005
Final manuscript due:	28 February 2006
Tentative publication date:	May 2006

HLT/EMNLP 2005 Call for Papers

Human Language Technology Conference
on Empirical Methods in Natural Language Processing

October 6-8, 2005
Vancouver, B.C., Canada

http://www.cs.utexas.edu/~ml/HLT-EMNLP05/

Submission deadline: June 3, 2005

HLT/EMNLP 2005 continues the conference series jointly sponsored by the Human Language Technology Advisory Board (HLT) and the Association for Computational Linguistics (ACL). This year's conference is co-sponsored by SIGDAT, the ACL's special interest group on linguistic data and corpus-based approaches to NLP, which has traditionally sponsored the Empirical Methods in Natural Language Processing (EMNLP) Conferences. The joint conference provides a unified forum for researchers across a spectrum of disciplines to present recent, high-quality, cutting-edge work, to exchange ideas, and to explore emerging new research directions. The conference especially encourages submissions that discuss synergistic combinations of language technologies (e.g., Speech with Information Retrieval, Machine Translation with Speech, Question Answering with Natural Language Processing, etc.). Particular consideration will be given to papers addressing novel learning tasks and evaluation metrics in speech, natural language processing and information retrieval, including e.g.:

• learning tasks insufficiently addressed in the past, e.g. collaborative learning, learning in the presence of background knowledge, or finding anomalies in data;
• limits of standard evaluation methods on new tasks;
• novel performance measures incorporating user preferences, competence, or relevance to a given problem;
• learning and optimization algorithms addressing the above, e.g. novel statistical methods or cognitively inspired solutions.

We are interested in papers from academia, government, and industry on all areas of traditional interest to the HLT and SIGDAT communities, as well as aligned fields, including but not limited to:

• Speech processing, including:
  • Speech recognition
  • Speech generation
  • Speech summarization
  • Rich transcription: annotation of speech signals with metalinguistic information, such as speaker identity, attitude, emotion, etc.
  • Speech-based human-computer interfaces
• Text summarization
• Question answering
• Paraphrasing
• Computational analysis of phonology, morphology, prosody, syntax, semantics, pragmatics, discourse, style
• Statistical techniques for language processing, including:
  • Corpus-based language modeling
  • Lexical and knowledge acquisition
• Language generation and text planning
• Sentence parsing and discourse analysis
• Multilingual processing, including:
  • Machine translation of speech and text
  • Cross-language information retrieval
  • Multi-lingual speech recognition and language identification
• Evaluation, including:
  • Glass-box evaluation of HLT systems and system components
  • Back-box evaluation of HLT systems in application settings
• Development of language resources, including:
  • Lexicons and ontologies
  • Treebanks, proposition banks, and frame banks
• Understanding of human communication, including:
  • Natural language interfaces
  • Dialogue structure and dialogue systems
  • Message and narrative understanding systems
• Information extraction from multiple media
• Information retrieval, including:
  • Formal models, clustering and classification
  • Web mining for IR
  • Natural language processing for IR
  • Spoken IR
  • Metadata annotation and XML IR

Important Dates:

Submissions

Submissions must describe original, completed, unpublished work, and include concrete evaluation results when appropriate. Papers being submitted to other meetings must provide this information (see submission format). In the event of multiple acceptances, authors are requested to immediately notify the HLT/EMNLP program chair and to choose which meeting to present and publish the work at as soon as possible. We cannot accept for publication or presentation work that will be (or has been) published elsewhere.

Papers must be submitted electronically in Postscript (PS) or Portable Document Format (PDF). They should follow the ACL formatting guidelines and should not exceed eight (8) pages in two-column format, including references and illustrations. Papers exceeding the maximum length may be rejected without review. Authors are encouraged to use the style files provided on the HLT/EMNLP 2005 website. We strongly prefer submissions in PS format. Any author who submits in PDF must assume the responsibility for ensuring that fonts are treated properly so that the paper will print (not just view) anywhere. (This may involve reading the manual.) DOC/RTF formats cannot be accepted.

Reviewing will be blind. No information identifying the authors should be in the paper: this includes not only the authors' names and affiliations, but also self-references that reveal authors' identities; for example, "We have previously shown (Smith 1999)" should be changed to "Smith (1999) has previously shown". Names and affiliations should be listed on a separate identification page.

Papers must be submitted electronically by 12 a.m. GMT on June 3, 2005, through the conference website. In addition, information about each paper must be provided, including:

• Paper title
• Authors' names, affiliations, and contact information
• Contact author
• A short list of keywords (selected from a predefined list)
• Abstract (no more than 300 words)
• A statement whether the paper has been or will be submitted to other conferences

Authors who cannot submit a file electronically should contact the program chairs before the due date to arrange alternative forms of submission.

After notifications of acceptance have been issued, authors will have the opportunity to revise their submissions in accordance with reviewers' comments. The due date for the final submission of camera-ready papers is August 12, 2005.
top of page

CALL FOR PAPERS

IEEE ASRU 2005

Automatic Speech Recognition and Understanding Workshop

Fiesta Americana Grand Coral Beach Resort 

Cancun, Mexico

November 27 – December 1, 2005

The Ninth biannual IEEE workshop on Automatic Speech Recognition and  Understanding (ASRU) will be held November 27-December 1, 2005.   The ASRU Workshops have a tradition of bringing together researchers from academia and industry in an intimate and collegial setting to discuss problems of common interest in automatic speech recognition and understanding. Papers in all areas of human language technology are encouraged to be submitted, with emphasis placed on automatic speech recognition and understanding technology, speech to text systems, spoken dialog systems, multilingual language processing, robustness in ASR, spoken document retrieval, and speech-to-speech translation.
 
The workshop program will consist of invited lectures, oral and poster presentations, and panel discussions. Ample time will be allowed for informal discussions and to enjoy the impressive tropical setting.  The workshop website http://www.asru2005.org will be accessible by January, 2005.

PAPER SUBMISSION:
Prospective authors are invited to submit full-length, 4-6 page papers, including figures and references, to www.asru2005.org . All papers will be handled and reviewed electronically. The ASRU 2005 website will provide you with further details. Please note that the submission dates for papers are strict deadlines.

SPECIAL SESSIONS:
Special sessions proposals should be submitted by June 15, 2005, to asru05-tc@lists.csail.mit.edu and must include a topical title, rationale, session outline, contact information, and a description of how the session will be organized.

TENTATIVE  DATES

May 1, 2005            Workshop registration opens
July 1, 2005              Camera-ready paper submission deadline
August 15, 2005        Paper Acceptance / Rejection notices mailed
Sept. 15, 2005          Revised Papers Due and Author Registration Deadline
Oct. 1, 2005             Hotel Reservation and Workshop Registration
Nov. 27 – Dec.1, 2005    Workshop

ORGANIZING COMMITTEE
           
General Chairs
    Jim Glass, MIT, USA
    Richard Rose, McGill University, Canada
Technical Chairs
    Michael Picheny, IBM, USA
    Renato de Mori, Avignon, France
    Richard Stern, CMU, USA
Publicity Chair
    Ruhi Sarikaya, IBM, USA
Publications Chair
    Dilek Hakkani-Tur, AT&T, USA
Local Arrangements Chair:
    Juan Nolazco, Monterrey, Mexico
Demonstrations Chair
    Anand Venkataraman, SRI, USA

TECHNICAL COMMITTEE
Alex Acero,                                          Sadaoki Furui, Univ. of Tokyo
Srini Bangalore,                                    J.L. Gauvain, LIMSI
Jerome Bellegarda, Apple                     Yuqing Gao, IBM
Mary Harper,                                         Hermann Ney, RTWH Aachen
Julia Hirshberg, Columbia University    Joe Picone, Mississippi State University
Helen Meng, CUHK                             Abeer Alwan, UCLA
Roberto Pieraccini, IBM                        Jeff Bilmes, Univ. of Washington
Alex Rudnicky, CMU                           Herve Boulard, IDIAP
Stephanie Seneff, MIT                           Dan Ellis, Columbia University
Liz Shriberg, SRI                                   Mark Hasegawa-Johnson, UIUC
Gokhan Tur, AT&T                               Hynek Hermansky, IDIAP
Wayne Ward, Univ. of Colorado            Chris Wellekens, EURECOM
Steve Young, Cambridge University      Chin-Hui Lee, Georgia Tech
Eric Fosler-Lussier, Ohio State University   Shri Narayanan, USC

back to top

The 2005 National Institute of Standards

Speaker Recognition Evaluation

NIST has been coordinating Speaker Recognition Evaluations since 1996. Each evaluation begins with the announcement of the official evaluation plan which clearly states the rules and tasks involved with the evaluation. The evaluation culminates with a follow-up workshop, where NIST reports the official results and researchers share in their findings.

Brief History

Since 1996, over 40 research sites have participated in our evaluations. Each year, new researchers in industry and universities are encouraged to participate. Collaboration between universities and industries, is also welcomed. The overall goals of the evaluations have always been to drive the technology forward, to measure the state-of-the-art, and to find the most promising algorithmic approaches.

The 2005 NIST Speaker Recognition Evaluation

The 2005 NIST Speaker Recognition evaluation is part of an ongoing series of yearly evaluations conducted by NIST. These evaluations provide an important contribution to the direction of research efforts and the calibration of technical capabilities. They are intended to be of interest to all researchers working on the general problem of text independent speaker recognition. To this end the evaluation was designed to be simple, to focus on core technology issues, to be fully supported, and to be accessible.

Non-LDC members are required to sign the LDC's license agreement before being granted access to SRE-05 data.

More Information

To find out more about previous evaluations and access related publications,  go to the  NIST  speaker idenitfication web page.  To register your desire to participate in future evaluations, to obtain more information about our evaluations, or to be notified of the next evaluation or developments there of, please e-mail Dr. Alvin Martin at NIST.

Postdoctoral Fellowship Position in Speech Recognition

and Speech Modeling at McGill University

The position is in support of a Canadian funded NSERC project that is being conducted in association with the European 6th Frame Project DIVINES.  The overall goal of the project is to overcome deficiencies in existing acoustic feature analysis and phonetic and lexical modeling techiques.  This will be accomplished through a methodology  involving the diagnosis and modeling of instrinsic variabilities in ASR under a variety of conditions.   

McGill University is located in Montreal, an exciting cosmopolitan city in the Province of Quebec. Montreal is the home of a number of speech recognition Research and Development Laboratories including Centre Recherche  Informatique Montreal (CRIM), ScanSoft Canada, Nuance Canada, Nu Echo, and others.  Montreal is composed of a bilingual population with a blend of European and North American culture. Qualified applicants are invited to submit a resume together with the names and addresses of two references by email to:

ASR Researchers Take New Positions

• In July 2004, Roland Kuhn moved from the Panasonic Speech Technology Laboratory in Santa Barbara to Ottawa, Canada. Since then, his research has focused on statistically-oriented machine translation. His new contact information is:
  
 
Roland Kuhn
Interactive Language Technologies Group
Institute of Information Technology
National Research Council of Canada
Roland.Kuhn@cnrc-nrc.gc.ca
Tel: (819) 934-4222 "

Links to Upcoming Conferences and Workshops

ICASSP2005
Philadelphia, Pennsylvania, May, 2005
http://www.icassp2005.org/

SIGdial Workshop on Discourse and Dialog
Lisbon, Portugal , September 2-3, 2005
http://www.sigdial.org/workshops/workshop6

EUROSPEECH 2005 9th European Conference on Speech Communication and Technology
Lisbon, Portugal, September 4-8, 2005
http://www.interspeech2005.org/

Disfluency in Spontaneous Speech
Aix-en-Provence, September 10-12, 2005
http://www.up.univ-mrs.fr/delic/Diss05

SPECOM 2005 - 10th International Conf. on Speech and Computers
Patras, Greece, October 17-19, 2005
http://www.wcl.ee.upatras.gr/specom2005.htm

IEEE ASRU2005 Automatic Speech Recognition and Understanding Workshop
Cancun, Mexico, November 27 - December 1, 2005
http://www.asru2005.org

ICASSP2006
Toulouse, France May 15-19, 2006

INTERSPEECH 2006 - ICSLP  
Pittsburgh, PA, USA September 17-21, 2006
http://www.interspeech2006.org/

ICASSP2007
Honolulu, Hawaii, USA, 2007, April 17-20

INTERSPEECH 2007
Antwerp, Belgium, August 27-31, 2007
http://www.interspeech2007.org/