使用 Bi-LSTM 分类文档以疏通巴西最高法院
信息检索
2018-11-29 v1 机器学习
机器学习
摘要
巴西法院系统目前是世界上积压最严重的司法系统。每天有数千起诉讼案件到达最高法院。这些案件需要被分析以便关联到相关标签并分配给正确的团队。大多数案件以光栅扫描文档的形式到达法院，其质量水平差异很大。分析的第一步之一是对这些文档进行分类。在本文中，我们提出一种双向长短期记忆网络 (Bi-LSTM) 来对这些法律文档片段进行分类。
引用
@article{arxiv.1811.11569,
title = {Document classification using a Bi-LSTM to unclog Brazil's supreme court},
author = {Fabricio Ataides Braz and Nilton Correia da Silva and Teofilo Emidio de Campos and Felipe Borges S. Chaves and Marcelo H. S. Ferreira and Pedro Henrique Inazawa and Victor H. D. Coelho and Bernardo Pablo Sukiennik and Ana Paula Goncalves Soares de Almeida and Flavio Barros Vidal and Davi Alves Bezerra and Davi B. Gusmao and Gabriel G. Ziegler and Ricardo V. C. Fernandes and Roberta Zumblick and Fabiano Hartmann Peixoto},
journal= {arXiv preprint arXiv:1811.11569},
year = {2018}
}
备注
This work was presented at NIPS 2018 Workshop on Machine Learning for the Developing World (ML4D)