DAAT-CWS
Coupling Distant Annotation and Adversarial Training for Cross-Domain Chinese Word Segmentation
Paper accepted by ACL 2020
- python == 2.7
- tensorflow == 1.8.0
source domain dataset PKU and five distantly-annotated target datasets are put in data/datasets directory
Run python train.py --tgt_train_path <tgt_train_path> --tgt_test_path <tgt_test_path>
Note:
This code is based on the previous work by chqiwang. Many thanks to chqiwang. The raw text of dataset used in our paper can be found at CWS-NAACL2019