Incomplete Utterance Rewriting as Sequential Greedy Tagging

TL;DR

This paper introduces a sequence tagging-based model with speaker-aware embeddings for incomplete utterance rewriting, achieving superior accuracy and faster inference across multiple datasets.

Contribution

The paper presents a novel sequence tagging approach with speaker-aware embeddings, improving context extraction and inference speed in incomplete utterance rewriting.

Findings

Achieves top results on nine restoration scores

Outperforms previous models in inference speed

Maintains comparable performance on other metrics

Abstract

The task of incomplete utterance rewriting has recently gotten much attention. Previous models struggled to extract information from the dialogue context, as evidenced by the low restoration scores. To address this issue, we propose a novel sequence tagging-based model, which is more adept at extracting information from context. Meanwhile, we introduce speaker-aware embedding to model speaker variation. Experiments on multiple public datasets show that our model achieves optimal results on all nine restoration scores while having other metric scores comparable to previous state-of-the-art models. Furthermore, benefitting from the model's simplicity, our approach outperforms most previous models on inference speed.