HOME    »    PROGRAMS/ACTIVITIES    »    Annual Thematic Program
Fall 2000
IMA Workshop
Mathematical Foundations of Speech Processing and Recognition
September 18-22, 2000


with partial support by The Office of Naval Research


Organizers:

Sanjeev Khudanpur
Center for Language and Speech Processing
Johns Hopk
ins University

khudanpur@jhu.edu

Mari Ostendorf
Department of Electrical Engineering
University of Washington
mo@ee.washington.edu

Roni Rosenfeld
School of Computer Science
Carnegie Mellon University

roni@CMU.EDU


Speech recognition technology has seen a significant breakthrough with the introduction of Hidden Markov Models and related data-driven statistical techniques in the 70's and 80's. But current technology is still very fragile, breaking down with small changes in speaker characteristics, channel characteristics or discourse domain. This points to fundamental weaknesses in signal representation, and existing limitations of the HMM approach. The proposed workshop will bring together researchers actively working on novel solutions to the problems of speech signal representation, acoustic modeling, and model adaptation, together with mathematicians working in the areas of statistical estimation, stochastic processes and clustering. The first two days will consist of overviews of the acoustics, production, perception and automatic recognition of speech and related speech applications, followed by talks about ongoing research in these areas. The goal is to create a synergy between mathematicians and practitioners to pursue the next generation of solutions to speech processing and recognition problems. To this end, we will aim to:

  • Familiarize mathematicians with speech processing and recognition technology, outline underlying fundamental problems and currently popular/successful solutions.

  • Present novel models, ideas or approaches currently being pursued in the speech community.

  • Present recent advances in mathematics which may be relevant to the speech community.

We hope that this will in turn encourage informal discussions, exchanging of ideas, fostering collaborations and formulating specific mathematical problems whose solutions will advance the state of the art.

WORKSHOP SCHEDULE

Monday Tuesday
MONDAY, SEPTEMBER 18
REVIEW Of APPLICATIONS, MATHEMATICAL FOUNDATIONS AND SPEECH SCIENCE All talks are in Lecture Hall EE/CS 3-180 unless otherwise noted.
8:30 am Coffee and Registration

Reception Room EE/CS 3-176

9:10 am Willard Miller, Fred Dulles, and Mari Ostendorff Introduction
9:30 am Sanjeev Khudanpur
Johns Hopkins University

Recognize Speech v/s Wreck a Nice Beach: The Mathematics of Automatic Speech Recognition

Talk   pdf   postscript

10:30 am Break Reception Room EE/CS 3-176
11:00 am-12:00 pm Steve Young
University of Cambridge

The State of the Art in ASR (and Beyond?)

Talk   pdf    powerpoint

2:00 pm Abeer Alwan
UCLA
http://www.icsl.ucla.edu/~spapl
Toward Quantitative Models of Human Speech Perception and Production Mechanisms
3:00 pm Break Reception Room EE/CS 3-176
3:30 pm Discussant: Herve Bourlard
Swiss Federal Institute of Technology at Lausanne (EPFL)

Hard Problems in ASR

Talks:   iima.pdf   iima.ps    keynote.pdf   keynote.ps   likeli.pdf   likeli.ps
4:15 pm IMA Tea IMA East, 400 Lind Hall A variety of appetizers and beverages will be served.
TUESDAY, SEPTEMBER 19 SPEECH MOTIVATED EXTENSIONS TO HMMs All talks are in Lecture Hall EE/CS 3-180 unless otherwise noted.
9:15 am Coffee Reception Room EE/CS 3-176
9:30 am Wendy Holmes
20/20 Speech Ltd.

Segmental HMMs: Modelling Dynamics and Underlying Structure for Automatic Speech Recognition

Talk1  pdf    powerpoint

Talk2   pdf   powerpoint

10:30 am Break Reception Room EE/CS 3-176
11:00 am-12:00 pm Li Deng
Microsoft Research
Switching Dynamic-System Models for Speech Articulation and Acoustics
2:00 pm James Glass
MIT Laboratory for Computer Science
Modelling Graph-based Observation Spaces for Segment-Based Speech Recognition
3:00 pm Break Reception Room EE/CS 3-176
3:30-4:30pm Discussion Panelists: Dave Castanon, Herb Gish and Hollis Fitch
WEDNESDAY, SEPTEMBER 20 SIGNAL REPRESENTATION All talks are in Lecture Hall EE/CS 3-180 unless otherwise noted.
9:15 am Coffee Reception Room EE/CS 3-176
9:30 am Michael W. Macon
Oregon Graduate Institute of Science & Technology
Speech Signal Representations in Speech Synthesis: Current trends and parallels in ASR
10:30 am Break Reception Room EE/CS 3-176
11:00 am-12:00 pm Richard M. Stern
Carnegie Mellon University

Robust Signal Representations for Automatic Speech Recognition

Talk    powerpoint

2:00 pm Benjamin Kedem
University of Maryland

Semiparametric Filtering in Speech Processing

Talk    pdf    postscript

3:30-4:30pm Discussion Panelists: Jordan Cohen, Yoshinori Sagisaka, Kemal Sonmez
THURSDAY, SEPTEMBER 21 LEARNING FROM MACHINE LEARNING All talks are in Lecture Hall EE/CS 3-180 unless otherwise noted.
9:15 am Coffee Reception Room EE/CS 3-176
9:30 am Jeff A. Bilmes
University of Washington

Graphical Models and Applicability to the Speech Recognition Problem

Talk  pdf

10:30 am Break Reception Room EE/CS 3-176
11:00 am-12:00 pm Lawrence K. Saul
AT&T
Dimensionality Reduction in Machine Learning and Multimedia Processing
2:00 pm Discussion Panelist: Partha Niyogi
3:00 pm Break Reception Room EE/CS 3-176
3:30 pm Contributed Talks TBA
6:00 pm Workshop Dinner Jewel of India
FRIDAY, SEPTEMBER 22 ADDRESSING MODEL MISMATCH All talks are in Lecture Hall EE/CS 3-180 unless otherwise noted.
9:15 am Coffee Reception Room EE/CS 3-176
9:30 am Chin-Hui Lee
Bell Laboratories, Lucent Technologies

Parameter Adaptation and Compensation in Designing Maximum A Posteriori Decision Rules for Automatic Speech Recognition

Talk   pdf    postscript

10:30 am Break Reception Room EE/CS 3-176
11:00 am-12:00 pm Ramesh Gopinath
IBM T.J. Watson Research

Gaussianization

Talk   pdf    postscript

12:00 - 1:00 pm Discussants: Ananth Sankar, Mari Ostendorf Revisiting Hard Problems
Monday Tuesday

LIST OF CONFIRMED PARTICIPANTS

as of 9/21/2000
Name Department Affiliation
Abeer Alwan Electrical Engineering UCLA
Joan Bachenko   Linguistech Technologies
Jeff Bilmes Electrical Engineering University of Washington
Herve Bourlard IDIAP Swiss Federal Institute of Technology at Lausanne (EPFL)
Mireille Boutin Mathematics University of Minnesota
Jamylle Carter   Institute for Mathematics and its Applications
David Castanon ENG Electrical & Computer Engineering Boston University
Li-Tien Cheng   Institute for Mathematics and its Applications
Jordan Cohen   Voice Signal Technologies, Inc.
Zoran Cvetkovic Information Sciences Research AT&T Shannon Laboratory
Li Deng   Microsoft Research
Fred Dulles   Institute for Mathematics & its Applications
Selim Esedoglu   Institute for Mathematics and its Applications
Hollis Fitch   Institute for Defense Analysis
Tara Gibbs Linguistics University of Minnesota
Herb Gish   BBN Systems and Technologies
James Glass Lab for Computer Science MIT
Fikri Goksu ECE University of Minnesota
Ramesh A. Gopinath Speech Transcription Algoithms, HLT IBM T.J. Watson Research
Surendra Gupte Computer Science & Engineering University of Minnesota
Mitch Helle-Morrissey Research Lernout & Hauspie
Wendy Holmes   20/20 Speech Ltd.
Kirk Jackson DIRNSA National Security Agency
Benjamin Kedem Mathematics University of Maryland
Irving Kessler   IDA-CCR
Sanjeev Khudanpur Center for Language and Speech Processing Johns Hopkins University
Seongjai Kim Mathematics University of Kentucky
Christopher Lang   Indiana University Southeast
Chin-Hui Lee Dialogue Systems Research Lucent Technologies
Zhijun Liu Mathematics & Statistics Mississippi State University
Michael Macon Center for Spoken Language Understanding Oregon Graduate Institute of Science & Tech.
David McKoskey Research/Development Linguistic Technologies, Inc.
Willard Miller   Institute for Mathematics & its Applications
Partha Niyogi   Bell Laboratories
Alexei Novikov   Institute for Mathematics & its Applications
Peter Olver Mathematics University of Minnesota
Mari Ostendorf Electrical Engineering University of Washington
Sergey Pakhomov ILLASL (Linguistics Program) University of Minnesota
Yingyong Qi   Qualcomm Inc.
Jianliang Qian   Institute for Mathematics and its Applications
Yoshinori Sagisaka   ATR Interpreting Telecommunications Res. Lab
Ananth Sankar   Nuance Communications
Lawrence Saul   AT&T Labs -- Research
Guergana Savova LTI/L&H Lernout and Hauspie
Michael Schonwetter Research Lernout & Hauspie
Kemal Sonmez   SRI International
Richard Stern Engineering Carnegie Mellon University
Jack Xin Mathematics University of Texas at Austin
Xiangsheng Xu Mathematics and Statistics Mississippi State University
Steve Young Engineering Cambridge University


2000-2001 Program: Mathematics in Multimedia

Back to top of page

Connect With Us:
Go