I’ve finally stopped procrastination, and my paper for ICALT 2005 is underway. I have most of my data gathered for results processing, so that’s no problem… but the paper itself isn’t separating from my brain very easily. I’ve managed to procrastinate on a fine array of things today.
Semi-related mind-dump: I really like this quote from SRI’s 1998 paper Calibration Of Machine Scores For Pronunciation Grading (pdf):
The ability to accept speech input allows developers oflanguage learning software to move beyond traditional reading and listening comprehension activities toexercises requiring active speech production such as reading aloud, oral multiple-choice exercises, and opendialogs within limited domains. In an ideal system, the computer may provide feedback on lexical, syntactic,and acoustic aspects of language use.
Two desirable features of speech-enabled computer-based language-learning applications are the ability to recognize accented or mispronounced speech producedby language learners, and the ability to provide meaningful feedback on pronunciation quality.