CLSP Homepage : Workshop Homepage
Workshop 2007
Workshop 2007 Wednesday, May 16, 2012

Recovery from Model Inconsistency in Multilingual Speech Recognition

Current ASR has difficulties in handling unexpected words that are typically replaced by acoustically acceptable high prior probability words. Identifying parts of the message where such a replacement could have happened may allow for corrective strategies.

We aim to develop data-guided techniques that would yield unconstrained estimates of posterior probabilities of sub-word classes employed in the stochastic model solely from the acoustic evidence, i.e. without use of higher level language constraints.

These posterior probabilities then could be compared with the constrained estimates of posterior probabilities derived with the constraints implied by the underlying stochastic model.
Parts of the message where any significant mismatch between these two probability distributions is found should be re-examined and corrective strategies applied.

This may allow for development of systems that are able to indicate when they "do not know" and eventually may be able to "learn-as-you-go" in applications encountering new situations and new languages.

During the 2007 Summer Workshop we intend to focus on detection and description of out-of-vocabulary and mispronounced words in the 6 language Call-home database. Additionally, in order to describe the suspect parts of the message, we will work on language-independent recognizer of speech sounds that could be applied for phonetic transcription of identified suspect parts of the recognized message.


Click here for technical details.

 
Team Members:
Hynek Hermansky Team Leader IDIAP hynek at idiap dot ch

Lukas Burget Senior Researcher Brno University of Technology burget at fit dot vutbr dot cz
Sanjeev Khudanpur Senior Researcher Johns Hopkins University khudanphur at jhu dot edu
Chin-Hui Lee Senior Researcher Georgia Technical Institute chl at ece dot gatech dot edu
Haizhou Li Senior Researcher Institute for Infocomm Research hli at i2r dot a-star dot edu dot sg
Jon Nedel Senior Researcher Department of Defense jnedel at gmail dot com
Geoffrey Zweig Senior Researcher Microsoft gzweig at microsoft dot com

Pavel Matejka Graduate Student Brno University of Technology matejkap at fit dot vutbr dot cz
Ariya Rastrow Graduate Student Johns Hopkins University ariya at jhu dot edu
Petr Schwartz Graduate Student Brno University of Technology schwarzp at fit dot vutbr dot cz
Rong Tong Graduate Student Nanyang Technological University tongrong at i2r dot a-star dot edu dot sg
Chris White Graduate Student Johns Hopkins University cmileswhite at jhu dot edu

Mirko Hannemann Undergraduate Student Magdeburg University, Germany mirko dot hannemann at idiap dot ch
Sally Isaacoff Undergraduate Student University of Michigan sisaacof at umich dot edu
Puneet Sahani Undergraduate Student NSIT; Delhi University sahani dot puneet at gmail dot com

The Center for Language and Speech Processing
The Johns Hopkins University
3400 North Charles Street, Barton Hall
Baltimore, MD 21218
*Telephone: (410) 516-4237 *Fax: (410) 516-5050 *E-mail: clsp@clsp.jhu.edu