Page 177 - 《软件学报》2021年第10期

郭军军 等:融入案件辅助句的低频和易混淆罪名预测 3149

4 总结与展望

本文研究旨在提升低频易混淆罪名的预测准确率.针对低频罪名训练数据少和易混淆罪名案情描述不易
区分等原因导致两者预测准确率低这一问题,我们引入了案件辅助句这一概念,提出了一种融入案件辅助句与
案情描述构建双向互注意力建模的方法.此外,我们分别计算案情描述与案件辅助句不同粒度的特征.本文模型
在中国刑事案件数据集上取得当前最显著的效果.当然,我们的工作目前还有很多的不足之处,比如案件辅助句
的定义比较广泛,以及是否可自动构建方案更细化的案件辅助句,这是有待解决的工作之一.
未来的研究工作主要分为 3 个方面.
(1) 考虑基于神经网络等模型自动构建更为细化的案件辅助句,提升案件辅助句与案情描述更深层次的
语义信息交互;
(2) 当前工作只考虑了单项罪名指控,将来的工作设想结合多项指控案例数据,改进该模型为多罪名预
测,这更符合我国刑事案件罪名自动判决的目的;
(3) 法律判决预测任务包含法条推荐、罪名预测、刑期预测等多项子任务,将来的工作还考虑结合多项
子任务,提升低资源数据的判决预测准确率.

致谢 本文工作是在作者导师的悉心指导下完成的,深表感谢.同时,也真诚地感谢团队的其他老师和同学的耐
心解答,以及在法律判决任务方面做了大量工作的前辈们.

References:
[1] Xiao CJ, Zhong HX, Guo ZP, Tu CC, Liu ZY, Sun MS. Cail2018: A large-scale legal dataset for judgment prediction. 2018.
https://arxiv.org/abs/1807.0247
[2] Kort, F. Predicting Supreme Court decisions mathematically: A quantitative analysis of the “right to counsel” cases. The American
Political Science Review, 1957,51(1):112.
[3] Nagel SS. Applying correlation analysis to case prediction. Texas Law Review, 1963,42:1006.
[4] Keown R. Mathematical models for legal prediction. Computer/Law Journal, 1980,2:829.
[5] Liu CL, Hsieh CD. Exploring phrase-based classification of judicial documents for criminal charges in Chinese. In: Proc. of the
ISMIS. 2006. 681690.
[6] Chao LL, Cheng TC, Jim HH. Some case-refinement strategies for case-based criminal summary judgments. Journal of Information
Science and Engineering, 2004,20(4):783800.
[7] Sulea OM, Zampieri M, Vela M, Josef VG. Exploring the use of text classication in the legal domain. In: Proc. of the ASAIL. 2017.
[8] Zhong HX, Guo ZP, Tu CC, Xiao CJ, Liu ZY, Sun MS. Legal judgment prediction via topological learning. In: Proc. of the 2018
Conf. on Empirical Methodsin Natural Language Processing. Brussels: ACL, 2018. 35403549.
[9] Luo BF, Feng YS, Xu JB, Zhang X, Zhao DY. Learning to predict charges for criminal cases with legal basis. In: Proc. of the 2017
Conf. on Empirical Methodsin Natural Language Processing. Copenhagen: ACL, 2017. 27272736.
[10] Jiang X, Ye H, Luo ZC, Chao WH. Interpretable rationale augmented charge prediction system. In: Proc. of the 27th Int’l Conf. on
Computational Linguistics. Santa Fe: ACL, 2018. 146151.
[11] Long SB, Tu CC, Liu ZY, Sun MS. Automatic judgment prediction via legal reading comprehension. In: Proc. of the 18th China
National Conf. on Computational Linguistics. Kunming: CCL, 2019. 558572.
[12] Wang JW, Zhang H, Tan HY, Wang YL, Zhao HY, Li R. Multi-label charge prediction based on semantic differences of words.
Journal of Chinese Information Processing, 2019,33(10):127134 (in Chinese with English abstract).
[13] Lin WC, Kuo TT, Chang TJ, Yen CA, Chao JC, Lin SD. Exploiting machine learning models for chinese legal documents labeling,
case classification, and sentencing prediction. Int’l Journal of Computational Linguistics & Chinese Language Processing, 2012,
17(4):4967 (in Chinese with English abstract).
[14] Liu ZL, Zhang MS, Zhen RR, Gong ZQ, Yu N, Fu GH. Multi-task learning model for legal judgment predictions with charge
keywords. Journal of Tsinghua University (Science and Technology), 2019,59(7):497504 (in Chinese with English abstract).