E2E Segmenter: Joint Segmenting and Decoding for Long-Form ASR
This addresses the problem of suboptimal segmentation in long-form ASR for applications like YouTube transcription, offering a novel method that integrates segmentation and decoding.
The paper tackles the challenge of improving end-to-end ASR performance on long utterances by replacing separate VAD segmentation with a joint model that predicts boundaries using acoustic and semantic features, resulting in an 8.5% relative WER improvement and 250 ms reduction in median latency on real-world audio up to 30 minutes.
Improving the performance of end-to-end ASR models on long utterances ranging from minutes to hours in length is an ongoing challenge in speech recognition. A common solution is to segment the audio in advance using a separate voice activity detector (VAD) that decides segment boundary locations based purely on acoustic speech/non-speech information. VAD segmenters, however, may be sub-optimal for real-world speech where, e.g., a complete sentence that should be taken as a whole may contain hesitations in the middle ("set an alarm for... 5 o'clock"). We propose to replace the VAD with an end-to-end ASR model capable of predicting segment boundaries in a streaming fashion, allowing the segmentation decision to be conditioned not only on better acoustic features but also on semantic features from the decoded text with negligible extra computation. In experiments on real world long-form audio (YouTube) with lengths of up to 30 minutes, we demonstrate 8.5% relative WER improvement and 250 ms reduction in median end-of-segment latency compared to the VAD segmenter baseline on a state-of-the-art Conformer RNN-T model.