Paper: Detecting Structural Events for Assessing Non-Native Speech

ACL ID W11-1405
Title Detecting Structural Events for Assessing Non-Native Speech
Venue Innovative Use of NLP for Building Educational Applications
Year 2011

Structural events, (i.e., the structure of clauses and disfluencies) in spontaneous speech, are important components of human speaking and have been used to measure language devel- opment. However, they have not been ac- tively used in automated speech assessment research. Given the recent substantial progress on automated structural event detection on spontaneous speech, we investigated the de- tection of clause boundaries and interruption points of edit disfluencies on transcriptions of non-native speech data and extracted fea- tures from the detected events for speech assessment. Compared to features com- puted on human-annotated events, the features computed on machine-generated events show promising correlations to holistic scores that reflect speaking proficiency levels.