gen_embeddings_for_slu.py: The script for the slot filling task.slu_preprocess.py: The script for preprocessing the data in slot filling task.We list the preprocess details for each file as follows: Preprocessing scripts for cross-domain slot filling and NER are in the preprocess folder. Coach FrameworkĬross-domain Slot Filling: Evaluated on SNIPS dataset, which contains 39 slot types across seven domains (intents) and ~2000 training samples per domain.Ĭross-domain named entity recognition (NER): We take CoNLL-2003 English named entity recognition (NER) dataset as the source domain and the CBS SciTech News NER dataset as the target domain. Furthermore, our model can also be applied to the cross-domain named entity recognition task, and it achieves better adaptation performance than other existing baselines. In addition, we propose a template regularization approach to improve the adaptation robustness by regularizing the representation of utterances based on utterance templates.Įxperimental results show that our model significantly outperforms state-of-the-art approaches in slot filling. It then predicts the specific types for the slot entities. Our model first learns the general pattern of slot entities by detecting whether tokens are slot entities or not. In this paper, we propose a Coarse-to-fine approach (Coach) for cross-domain slot filling. Hence, cross-domain slot filling has naturally arisen to cope with this data scarcity problem. However, such data are not always available. Publisher = "Association for Computational Linguistics",Īs an essential task in task-oriented dialog systems, slot filling requires extensive training data in a certain domain. = "oach: A Coarse-to-Fine Approach for Cross-domain Slot Filling",īooktitle = "Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics", If you use any source codes or ideas included in this repository for your work, please cite the following paper.
TASK COACH PYTHON CODE
This code has been written using PyTorch.
It contains the scripts for the coach framework and the baseline models CT and RZT.
This repository is for the ACL-2020 paper: Coach: A Coarse-to-Fine Approach for Cross-domain Slot Filling. Coach: A Coarse-to-Fine Approach for Cross-domain Slot Filling
0 kommentar(er)
