/Chinese-Semantic-Role-Labeling-BERT

Applying BERT on Chinese Semantic Role Labeling

MIT LicenseMIT

Chinese-Semantic-Role-Labeling-BERT

My research internship at Utrecht Univeristy from November 2012 to March 2022. I was supervised by Dr.Tejaswini Deoskar. This github repository contains the code and report.

Key Takeaways:

• Applied BERT on the Chinese Proposition Bank 1.0 dataset and the English Universal Proposition Bank dataset. The model only uses sentence-predicate pairs as the input feature but outperforms the traditional methods which heavily rely on feature engineering.

• Conducted a quantitative analysis on the performance of the model on the aforementioned datasets.

• Conducted a comparative analysis on the model’s differing performance in semantic role label classification on the two aforementioned datasets

  • the code will to be added soon.