A Word-Granular Adversarial Attacks Framework for Causal Event Extraction.

A Word-Granular Adversarial Attacks Framework for Causal Event Extraction.
复制标题

DOI:
10.3390/e24020169
复制
发表时间:
2022-01-24
期刊:
Entropy (Basel, Switzerland)
影响因子:
--
通讯作者:
Zuo X
Zuo X
中科院分区:
其他
文献类型:
--
作者:
Zhao Y;Zuo W;Liang S;Yuan X;Zhang Y;Zuo X

文献摘要

参考文献

相似文献

掩蔽词作为一种数据扩充方法,在许多自然语言处理任务中得到了广泛的应用。但是,大多数掩码方法都是基于规则的,与下游任务无关。本文提出了一种新的掩码字生成器,称为Actor-Critic掩码模型(ACMM),该模型能够根据下游任务的性能自适应地调整掩码策略。为了验证该方法的有效性,我们在两个因果事件抽取数据集上进行了实验。实验结果表明,与各种基于规则的掩蔽方法相比,本文提出的方法生成的掩蔽语句能够显著增强模型的泛化能力,提高模型的性能。
As a data augmentation method, masking word is commonly used in many natural language processing tasks. However, most mask methods are based on rules and are not related to downstream tasks. In this paper, we propose a novel masking word generator, named Actor-Critic Mask Model (ACMM), which can adaptively adjust the mask strategy according to the performance of downstream tasks. In order to demonstrate the effectiveness of the method, we conducted experiments on two causal event extraction datasets. Experiment results show that, compared with various rule-based masking methods, the masked sentences generated by our proposed method can significantly enhance the generalization of the model and improve the model performance.
DOI: 10.1109/taslp.2021.3124365
发表时间: 2021-01-01
影响因子: 5.4
作者:
Cui, Yiming;Che, Wanxiang;Yang, Ziqing
通讯作者: Yang, Ziqing
DOI: 10.3390/e23101311
发表时间: 2021-10-06
期刊: Entropy (Basel, Switzerland)
影响因子: --
作者:
Chen S;Jiang C;Li J;Xiang J;Xiao W
通讯作者: Xiao W
DOI: 10.3390/e23111422
发表时间: 2021-10-28
期刊: Entropy (Basel, Switzerland)
影响因子: --
作者:
Schomacker T;Tropmann-Frick M
通讯作者: Tropmann-Frick M
DOI: 10.1162/tacl_a_00300
发表时间: 2020-01-01
影响因子: 10.9
作者:
Joshi, Mandar;Chen, Danqi;Levy, Omer
通讯作者: Levy, Omer
DOI: 10.1016/j.eswa.2018.08.009
发表时间: 2019-01-01
影响因子: 8.5
作者:
Li, Pengfei;Mao, Kezhi
通讯作者: Mao, Kezhi