Joint Speech Recognition and Speaker Diarization via Sequence Transduction
This addresses the need for accurate speaker attribution in conversational speech applications, such as medical dialogues, by integrating tasks that are typically handled separately, though it is incremental in combining existing sequence transduction methods.
The paper tackles the problem of jointly performing speech recognition and speaker diarization in conversations, proposing a novel approach using a recurrent neural network transducer that improves word-level diarization error rate from 15.8% to 2.2% on a medical conversation corpus.
Speech applications dealing with conversations require not only recognizing the spoken words, but also determining who spoke when. The task of assigning words to speakers is typically addressed by merging the outputs of two separate systems, namely, an automatic speech recognition (ASR) system and a speaker diarization (SD) system. The two systems are trained independently with different objective functions. Often the SD systems operate directly on the acoustics and are not constrained to respect word boundaries and this deficiency is overcome in an ad hoc manner. Motivated by recent advances in sequence to sequence learning, we propose a novel approach to tackle the two tasks by a joint ASR and SD system using a recurrent neural network transducer. Our approach utilizes both linguistic and acoustic cues to infer speaker roles, as opposed to typical SD systems, which only use acoustic cues. We evaluated the performance of our approach on a large corpus of medical conversations between physicians and patients. Compared to a competitive conventional baseline, our approach improves word-level diarization error rate from 15.8% to 2.2%.