Overview of Workshop

In a multilingual society like ours and with growing globalization and localization needs, language translation is very much needed in our day-to-day activities. Speech forms a natural man-machine interface. It is a much more convenient medium as compared to entry through the keyboards and reading the output generated by the machine. This workshop on "Speech Enabled Systems for Translation" will address all the major issues related to speech enabling of the translation process. A speech enabled system is aimed to provide an improved interactive environment between the user and the machine. The speech enabled applications can bring in a noticeable improvement in the accuracy as well as throughput at places that deal with large data and/or interact with large number of people. Depending upon the application, the speech interface can be deployed at both the input and output ends or at either end. While speech synthesis is a relatively easier task, automated speech recognition (ASR) at the input end could be quite complex depending upon the environment, medium used, speaker dependencies etc. However, the complexity of ASR could be reduced by domain customization, expectation driven NLP models providing disambiguation, learning based on past behavior etc. Though, we will examine speech enabled systems in general, our focus at this workshop will be on machine translation with human-machine interfaces and post editing experiences.

Updated on 07 May 2012