PKU Paraphrase Bank: A Sentence-Level Paraphrase Corpus for Chinese.

Bowei Zhang,Weiwei Sun,Xiaojun Wan,Zongming Guo
DOI: https://doi.org/10.1007/978-3-030-32233-5_63
2019-01-01
Abstract:One of the main challenges of conducting research on paraphrase is the lack of large-scale, high-quality corpus, which is particularly serious for non-English investigations. In this paper, we present a simple and effective unsupervised learning model that is able to automatically extract high-quality sentence-level paraphrases from multiple Chinese translations of the same source texts. By applying this new model, we obtain a large-scale paraphrase corpus, which contains 509,832 pairs of paraphrased sentences. The quality of this new corpus is manually examined. Our new model is language-independent, meaning that such paraphrase corpora for other languages can be built in the same way.
What problem does this paper attempt to address?