CCL23-Eval 任务6系统报告:基于CLS动态加权平均和数据增强的电信网络诈骗案件分类(System Report for CCL23-Eval Task 6:::Classification of Telecom Internet Fraud Cases Based on CLS Dynamic Weighted Average and Data Augement)

Tianjun Liu (天昀刘,), Tianhua Zhang (张兴华), Mengxiao Song (宋梦潇), Tingwen Liu (柳厅文)


Abstract
“电信网络诈骗领域的案件分类作为文本分类的一项落地应用,其目的是为相关案件进行智能化的分析,有助于公安部门掌握诈骗案件的特点,针对性的预防、制止、侦查。本文以此问题为基础,从模型设计、训练过程、数据增强三个方面进行了研究,通过CLS动态加权平均、Multi-Sample Dropout、对抗训练FGM、回译等方法显著提升了模型对诈骗案件描述的分类性能。”
Anthology ID:
2023.ccl-3.19
Volume:
Proceedings of the 22nd Chinese National Conference on Computational Linguistics (Volume 3: Evaluations)
Month:
August
Year:
2023
Address:
Harbin, China
Editors:
Maosong Sun, Bing Qin, Xipeng Qiu, Jing Jiang, Xianpei Han
Venue:
CCL
SIG:
Publisher:
Chinese Information Processing Society of China
Note:
Pages:
179–183
Language:
Chinese
URL:
https://aclanthology.org/2023.ccl-3.19
DOI:
Bibkey:
Cite (ACL):
Tianjun Liu, Tianhua Zhang, Mengxiao Song, and Tingwen Liu. 2023. CCL23-Eval 任务6系统报告:基于CLS动态加权平均和数据增强的电信网络诈骗案件分类(System Report for CCL23-Eval Task 6:::Classification of Telecom Internet Fraud Cases Based on CLS Dynamic Weighted Average and Data Augement). In Proceedings of the 22nd Chinese National Conference on Computational Linguistics (Volume 3: Evaluations), pages 179–183, Harbin, China. Chinese Information Processing Society of China.
Cite (Informal):
CCL23-Eval 任务6系统报告:基于CLS动态加权平均和数据增强的电信网络诈骗案件分类(System Report for CCL23-Eval Task 6:::Classification of Telecom Internet Fraud Cases Based on CLS Dynamic Weighted Average and Data Augement) (Liu et al., CCL 2023)
Copy Citation:
PDF:
https://preview.aclanthology.org/nschneid-patch-4/2023.ccl-3.19.pdf