Summary of Cfever: a Chinese Fact Extraction and Verification Dataset, by Ying-jia Lin et al.
CFEVER: A Chinese Fact Extraction and VERification Dataset
by Ying-Jia Lin, Chun-Yi Lin, Chia-Jen Yeh, Yi-Ting Li, Yun-Yu Hu, Chih-Hao Hsu, Mei-Feng Lee, Hung-Yu Kao
First submitted to arxiv on: 20 Feb 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary A novel Chinese dataset, CFEVER, is introduced for Fact Extraction and Verification. The dataset consists of 30,012 manually crafted claims based on Chinese Wikipedia content, with labels indicating the degree of factualness (“Supports”, “Refutes”, or “Not Enough Info”). Similar to FEVER, each claim in these categories is annotated with corresponding evidence sentences from single or multiple pages in Chinese Wikipedia. CFEVER demonstrates a Fleiss’ kappa value of 0.7934 for five-way inter-annotator agreement. The dataset serves as a new benchmark for factual extraction and verification, enabling the development of automated systems to aid human fact-checking efforts. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary CFEVER is a special dataset designed to help computers learn how to check if statements are true or false. It has 30,000 claims about things on Chinese Wikipedia, with labels that show how much each claim supports or refutes what it’s saying. The dataset even includes quotes from other places in Wikipedia that support or contradict the claims. CFEVER is useful for training machines to do fact-checking tasks and could help humans do their jobs better. |
