Automatic Quality Assessment for Speech Translation Using Joint ASR and MT Features
This work addresses the problem of evaluating speech translation quality for applications like re-scoring or user feedback, but it is incremental as it builds on existing ASR and MT methods.
The paper tackles automatic quality assessment for speech translation by defining it as a sequence labeling problem and proposing word confidence estimators based on ASR, MT, or combined features, using a corpus of 6.7k utterances, with results showing MT features are most influential while ASR features provide complementary information.
This paper addresses automatic quality assessment of spoken language translation (SLT). This relatively new task is defined and formalized as a sequence labeling problem where each word in the SLT hypothesis is tagged as good or bad according to a large feature set. We propose several word confidence estimators (WCE) based on our automatic evaluation of transcription (ASR) quality, translation (MT) quality, or both (combined ASR+MT). This research work is possible because we built a specific corpus which contains 6.7k utterances for which a quintuplet containing: ASR output, verbatim transcript, text translation, speech translation and post-edition of translation is built. The conclusion of our multiple experiments using joint ASR and MT features for WCE is that MT features remain the most influent while ASR feature can bring interesting complementary information. Our robust quality estimators for SLT can be used for re-scoring speech translation graphs or for providing feedback to the user in interactive speech translation or computer-assisted speech-to-text scenarios.