Punctuation Prediction Model for Conversational Speech

TL;DR

This paper develops neural network models to predict punctuation in conversational speech transcripts, improving readability and NLP processing by leveraging time-aligned data and pre-trained embeddings.

Contribution

It introduces two neural network variants trained on Fisher corpus with novel sequence alignment, demonstrating improved punctuation prediction accuracy in speech transcripts.

Findings

CNN models achieve higher precision, especially for question marks

BLSTM models have better recall and fewer overall mistakes

Using time-aligned data and pre-trained embeddings enhances prediction accuracy

Abstract

An ASR system usually does not predict any punctuation or capitalization. Lack of punctuation causes problems in result presentation and confuses both the human reader andoff-the-shelf natural language processing algorithms. To overcome these limitations, we train two variants of Deep Neural Network (DNN) sequence labelling models - a Bidirectional Long Short-Term Memory (BLSTM) and a Convolutional Neural Network (CNN), to predict the punctuation. The models are trained on the Fisher corpus which includes punctuation annotation. In our experiments, we combine time-aligned and punctuated Fisher corpus transcripts using a sequence alignment algorithm. The neural networks are trained on Common Web Crawl GloVe embedding of the words in Fisher transcripts aligned with conversation side indicators and word time infomation. The CNNs yield a better precision and BLSTMs tend to have better recall. While BLSTMs make fewer mistakes overall, the punctuation predicted by the CNN is more accurate - especially in the case of question marks. Our results constitute significant evidence that the distribution of words in time, as well as pre-trained embeddings, can be useful in the punctuation prediction task.