2019/09/19 by Adam Fisch, Jiang Guo, Fisch, Adam +3
Computer Science · #Natural Language Processing Techniques #Topic Modeling #Text Readability and Simplification
paper · pdf · doi:10.48550/arxiv.1909.09279
This paper explores the task of leveraging typology in the context of\ncross-lingual dependency parsing. While this linguistic information has shown\ngreat promise in pre-neural parsing, results for neural architectures have been\nmixed. The aim of our investigation is to better understand this\nstate-of-the-art. Our main findings are as follows: 1) The benefit of\ntypological information is derived from coarsely grouping languages into\nsyntactically-homogeneous clusters rather than from learning to leverage\nvariations along individual typological dimensions in a compositional manner;\n2) Typology consistent with the actual corpus statistics yields better transfer\nperformance; 3) Typological similarity is only a rough proxy of cross-lingual\ntransferability with respect to parsing.\n