The Statistical Speech Technology Working Papers (SSTWP) is intended as an opportunity for students, alums, post-docs and others to work out the kinks in their ideas before submitting them elsewhere for publication. SSTWP will be printed as a twice-annual book, copies of which will be given to each author. The book will not be republished on line, though the titles of SSTWP papers will be listed on this page. Each paper will be reviewed by four independent reviewers: Everyone who submits a paper will also be asked to review four other papers. Reviews should give the author detailed advice for improving the paper up to the publication standards of relevant professional societies (IEEE, ACM, ASA, ISCA). Participation in this process is considered to be an important and required part of graduate study in the Statistical Speech Technology group; senior students with many published papers are expected to participate, as a way of sharing their insight with younger students who have not yet published.
SSTWP papers are considered to be unpublished manuscripts. An author may post his or her manuscript on his or her private web page (just like any other unpublished manuscript), but the manuscripts will not be posted here. If you are from outside Urbana, and would like a copy of one of these papers, please write directly to the author.
In spring semester 2010, deadlines will be as follows.
- January 31 - Manuscripts due from everybody
- February 28 - First-round reviews completed
- March 31 - Revised manuscripts submitted
- April 30 - Second-round reviews completed
- May 31 - Final manuscripts submitted
MANUSCRIPTS, SPRING 2010
The following manuscripts are in the SSTWP review process for spring 2010.
- 2010s01 - State-Transition Interpolation and MAP Adaptation for HMM-based Dysarthric Speech Recognition
- 2010s02 - Robust automatic speech recognition using HMM trained in cepstral domain with decoder oriented missing feature classifier
- 2010s03 - Multi-Modal Fusion on Audio-Visual Event Detection
- 2010s04 - Tool for finding words and speech sounds in an unknown database
- 2010s05 - Applying Auditory Salience on Meeting Room Dataset
- 2010s06 - Prosody-Dependent Acoustic Modeling Using Variable-Parameter Hidden Markov Models
- 2010s07 - Finite-State ASR Dictionary using Gesture Pattern Vectors as Units
- 2010s08 - Detection of Auditory Saliency in Conference Room Environments
- 2010s09 - Simple Segmental Tandem Model
- 2010s10 - A Nonlinear One-Dimensional Model of Basilar Membrane Displacement
- 2010s11 - Improved Audio-Visual Speech Conversion with Minimum Conversion Error
- 2010s12 - Speech Recognition Experiments with Whispered Speech Corpora
- 2010s13 - A Fast Implementation of Factor Analysis for Speaker Verification