arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2302.00894cs.CL
收藏

如何为文本数据增强选择“好”样本

How to choose "Good" Samples for Text Data Augmentation

Guangdong University of Foreign Studies(广东外语外贸大学);Guangdong University of Technology(广东工业大学);Guangzhou Key Laboratory of Multilingual Intelligent Processing(广州市多语言智能处理重点实验室)

Xiaotian Lin, Nankai Lin, Yingwen Fu, Ziyu Yang, Shengyi Jiang

摘要:

基于深度学习的文本分类模型需要大量的标记数据才能获得有竞争力的性能。然而,对大规模语料库进行标注费时费力。为了解决这个问题,许多研究试图使用数据增强来扩大语料库的大小。然而,数据增强可能潜在地产生一些噪声增强样本。在自然语言处理领域,目前还没有研究增强样本的样本选择问题。在本文中,我们提出了一种新的自训练选择框架,具有两个选择器,以选择高质量的样本从数据增强。具体来说,我们首先使用基于熵的策略和模型预测来选择增强样本。考虑到在上述步骤中,一些高质量的样本可能会被错误地过滤,我们提出从词重叠和语义相似度两个角度来召回它们。实验结果表明,我们的框架的有效性和简单性。

英文摘要:

Deep learning-based text classification models need abundant labeled data to obtain competitive performance. Unfortunately, annotating large-size corpus is time-consuming and laborious. To tackle this, multiple researches try to use data augmentation to expand the corpus size. However, data augmentation may potentially produce some noisy augmented samples. There are currently no works exploring sample selection for augmented samples in nature language processing field. In this paper, we propose a novel self-training selection framework with two selectors to select the high-quality samples from data augmentation. Specifically, we firstly use an entropy-based strategy and the model prediction to select augmented samples. Considering some samples with high quality at the above step may be wrongly filtered, we propose to recall them from two perspectives of word overlap and semantic similarity. Experimental results show the effectiveness and simplicity of our framework.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑