View Code? Open in Web Editor
NEW
project for paper "Harvesting Events from Multiple Sources: Towards a Cross-Document Extraction Paradigm"
Python 99.18%
Jupyter Notebook 0.13%
Shell 0.68%
cles's Introduction
- project for paper "Harvesting Events from Multiple Sources: Towards a Cross-Document Extraction Paradigm"
1. The process of collecting the CLES dataset involved several steps:
![gather process](./figures/%E6%B5%81%E7%A8%8B%E5%9B%BE2.png)
- Firstly, we gather document-level event data from Wikipedia. The original dataset is located in "dataset/document_level_dataset".
- Secondly, we extract events from the documents using OmniEvent as our tool to obtain the raw dataset.
- We utilize human validation to obtain the final dataset, as described in the paper.
![CDEEpipline](./figures/%E6%A8%A1%E5%9E%8B%E5%9B%BE3.png)
cles's People
Contributors
Stargazers
Watchers