This paper describes our solution for SereTOD Challenge Track 1: Information extraction from dialog transcripts. We propose a token-pair framework to simultaneously identify entity and value mentions and link them into corresponding triples. As entity mentions are usually coreferent, we adopt a baseline model for coreference resolution. We exploit both annotated transcripts and unsupervised dialogs for training. With model ensemble and post-processing strategies, our system significantly outperforms the baseline solution and ranks first in triple f1 and third in entity f1.
CITATION STYLE
Wang, C., Kong, X., Huang, M., Li, F., Xing, J., Zhang, W., & Zou, W. (2022). A Token-pair Framework for Information Extraction from Dialog Transcripts in SereTOD Challenge. In SereTOD 2022 - Towards Semi-Supervised and Reinforced Task-Oriented Dialog Systems, Proceedings of the Workshop (pp. 19–23). Association for Computational Linguistics (ACL). https://doi.org/10.18653/v1/2022.seretod-1.3
Mendeley helps you to discover research relevant for your work.