BriefGPT.xyz
Oct, 2018
使用轻量级前馈神经网络的基于转移的句法分析
Transition-based Parsing with Lighter Feed-Forward Networks
HTML
PDF
David Vilares, Carlos Gómez-Rodríguez
TL;DR
本文研究在保证依存分析精度不受影响的情况下，如何去除嵌入式特征和减小其规模，以构建适用于多种不同语言的轻量级解析器，并在 Universal Dependencies 数据集上进行了实验。实验证明，对于多数树库而言，可以去除 grand-daughter 特征且不产生显著差异，同时也证明了可以显著降低嵌入向量的大小。
Abstract
We explore whether it is possible to build
lighter parsers
, that are statistically equivalent to their corresponding standard version, for a wide set of languages showing different structures and morphologies. As testbed, we use the
→