Reproduce “The Importance of Speech Genre for End-of-Turn Detector Training” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Rini Sharon, Kadri Hacioglu, Jeena Prakash, Andreas Stolcke
The authors compare end-of-turn detectors trained on conversational speech with detectors trained on complete single-turn speech, using frame-level and utterance-level tests. On CASPER, conversational training raises ConvLSTM F1 from 0.694 to 0.959 and lowers false alarms from 44.7% to 6.6%. Utterance-level gains depend on test genre; source and language differences prevent attributing results to genre alone.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.