Abstract
“诈骗案件分类问题是打击电信网络诈骗犯罪过程中的关键一环,根据不同的诈骗方式、手法等将其分类,通过对不同案件进行有效分类能够便于统计现状,有助于公安部门掌握当前电信网络诈骗案件的分布特点,进而能够对不同类别的诈骗案件作出针对性的预防、监管、制止、侦查等措施。诈骗案件分类属于自然语言处理领域的文本分类任务,传统的基于LSTM和CNN等分类模型能在起到一定的效果,但是由于它们模型结构的参数量的限制,难以达到较为理想的效果。本文基于预训练语言模型Nezha,结合对抗扰动和指数移动平均策略,有助于电信网络诈骗案件分类任务取得更好效果,充分利用电信网络诈骗案件的数据。我们队伍未采用多模型融合的方法,并最终在此次评测任务中排名第三,评测指标分数为0.8625。”- Anthology ID:
- 2023.ccl-3.20
- 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:
- 184–192
- Language:
- Chinese
- URL:
- https://aclanthology.org/2023.ccl-3.20
- DOI:
- Cite (ACL):
- Yongqing Huang, Hailong Yang, and Fu Xuelin. 2023. CCL23-Eval 任务6系统报告:基于预训练语言模型的双策略分类优化算法(System Report for CCL23-Eval Task 6:Double-strategy classification optimization algorithm based on pre-training language model). In Proceedings of the 22nd Chinese National Conference on Computational Linguistics (Volume 3: Evaluations), pages 184–192, Harbin, China. Chinese Information Processing Society of China.
- Cite (Informal):
- CCL23-Eval 任务6系统报告:基于预训练语言模型的双策略分类优化算法(System Report for CCL23-Eval Task 6:Double-strategy classification optimization algorithm based on pre-training language model) (Huang et al., CCL 2023)
- PDF:
- https://preview.aclanthology.org/nschneid-patch-4/2023.ccl-3.20.pdf