Conversational Query Rewriting with Self-supervised Learning
This work is significant for developers of multi-turn dialogue systems by reducing the reliance on labor-intensive supervised data annotation for conversational query rewriting, offering a more scalable solution.
This paper addresses the challenge of Conversational Query Rewriting (CQR) by proposing a self-supervised learning approach to automatically construct a large-scale CQR dataset, eliminating the need for human annotation. The authors introduce Teresa, a Transformer-based CQR model enhanced with self-attentive keyword detection and intent consistency constraints, which significantly outperforms existing CQR baselines on two public datasets.
Context modeling plays a critical role in building multi-turn dialogue systems. Conversational Query Rewriting (CQR) aims to simplify the multi-turn dialogue modeling into a single-turn problem by explicitly rewriting the conversational query into a self-contained utterance. However, existing approaches rely on massive supervised training data, which is labor-intensive to annotate. And the detection of the omitted important information from context can be further improved. Besides, intent consistency constraint between contextual query and rewritten query is also ignored. To tackle these issues, we first propose to construct a large-scale CQR dataset automatically via self-supervised learning, which does not need human annotation. Then we introduce a novel CQR model Teresa based on Transformer, which is enhanced by self-attentive keywords detection and intent consistency constraint. Finally, we conduct extensive experiments on two public datasets. Experimental results demonstrate that our proposed model outperforms existing CQR baselines significantly, and also prove the effectiveness of self-supervised learning on improving the CQR performance.