mawentao277 / CharBERT

CharBERT: Character-aware Pre-trained Language Model (COLING2020)
Apache License 2.0
117 stars 28 forks source link

将字符替换为部首,应用中文,是否可行? #7

Open guojson opened 3 years ago

NoviScl commented 3 years ago

Interesting idea! This paper (https://arxiv.org/pdf/1901.10125.pdf) has shown that glyph information can be useful for Chinese NLP. So I suspect that 部首 can be useful as well.