Sequence-to-sequence models for punctuated transcription

In this paper we present an extension of our previously described neural machine translation based system for punctuated transcription. This extension allows the system to map from per frame acoustic features to word level representations by replacing the traditional encoder in the encoder-decoder architecture with a hierarchical encoder. Furthermore, we show that a system combining lexical and acoustic features significantly outperforms systems using only a single source of features on all measured punctuation marks. The combination of lexical and acoustic features achieves a significant improvement in F-Measure of 1.5 absolute over the purely lexical neural machine translation based system.

Hierarchical encoder

Further details can be found in the paper:

Written by Ondrej Klejch on March 9, 2017