A Small and Fast BERT for Chinese Medical Punctuation Restoration

TL;DR

This paper introduces a compact, efficient BERT-based model tailored for Chinese medical punctuation restoration, significantly reducing model size while maintaining high accuracy for clinical speech transcription tasks.

Contribution

A novel lightweight pre-trained model utilizing contrastive learning and auxiliary tasks for improved Chinese medical punctuation restoration.

Findings

Achieves 95% performance of larger models

Reduces model size by 90% compared to Chinese RoBERTa

Demonstrates effectiveness in clinical dictation scenarios

Abstract

In clinical dictation, utterances after automatic speech recognition (ASR) without explicit punctuation marks may lead to the misunderstanding of dictated reports. To give a precise and understandable clinical report with ASR, automatic punctuation restoration is required. Considering a practical scenario, we propose a fast and light pre-trained model for Chinese medical punctuation restoration based on 'pretraining and fine-tuning' paradigm. In this work, we distill pre-trained models by incorporating supervised contrastive learning and a novel auxiliary pre-training task (Punctuation Mark Prediction) to make it well-suited for punctuation restoration. Our experiments on various distilled models reveal that our model can achieve 95% performance while 10% model size relative to state-of-the-art Chinese RoBERTa.