Multitask Pointer Network for multi-representational parsing
作者：
Highlights：
• Dependency and constituent trees are used by many artificial intelligence applications.
• These structures are separately produced by either dependency or constituent parsers.
• Our approach can parse any input sentence with both dependency and constituent trees.
• We just need to train a single model to efficiently produce any syntactic structure.
• It achieves state-of-the-art results on Penn treebanks and discontinuous German datasets.
摘要
•Dependency and constituent trees are used by many artificial intelligence applications.•These structures are separately produced by either dependency or constituent parsers.•Our approach can parse any input sentence with both dependency and constituent trees.•We just need to train a single model to efficiently produce any syntactic structure.•It achieves state-of-the-art results on Penn treebanks and discontinuous German datasets.
论文关键词：Natural language processing,Computational linguistics,Parsing,Dependency parsing,Constituent parsing,Neural network,Deep learning
论文评审过程：Received 4 June 2021, Revised 20 September 2021, Accepted 14 November 2021, Available online 26 November 2021, Version of Record 2 December 2021.
论文官网地址：https://doi.org/10.1016/j.knosys.2021.107760
