The last decade has seen impressive advances in the capability and performance of speech recognizers. However, despite the extent of progress and the emergence of commercially available products, there are many outstanding research issues that need to be addressed before the core speech recognition problem can be considered solved. Speech recognizers remain extremely sensitive to the acoustic and the linguistic properties of the data, and in particular to mismatches between training conditions and conditions of real use. Also porting to a new language or task requires substantial investment of time and money, and requires the knowledge of speech recognition experts.

Some of the challenges for speech recognition research are to address these problems by investigating robust generic technology, low-cost development strategies, portability and adaptability.

The ISCA ASR-2000 workshop will focus on core speech recognition technology and in particular on recent advances and prospective work. Topics of interest include, but are not limited to, the following:

- Robust generic technology
- Acoustic modeling and adaptation
- Language modeling and adaptation
- Lexical and phonological modeling
- Efficient decoding
- Portability and low-cost system development
- Modeling of speaking styles and non-native speech
- Enhancing the recognizer output (confidence measures,
   punctuation, audio segmentation, speaker and channel
   information, ...)

Prospective authors are invited to electronically submit a full paper (4-8 pages) for review (starting April 1st, 2000). (Paper submission closed)

Important dates:

Electronic submission of draft papers	May 1, 2000
Notification of acceptance	June 18
Early registration	August 1
Final papers due	August 1
Workshop	September 18-20

Click here to indicate your interest in the workshop and to be added to the workshop mailing list.