利用 Transformer 和相似度度量进行数据增强以改进阿拉伯语文本分类
Data Augmentation using Transformers and Similarity Measures for Improving Arabic Text Classification
AI总结：
提出基于 AraGPT-2 的阿拉伯语文本数据增强方法，并结合多种相似度度量评估生成句，用 AraBERT 验证其在四个情感数据集上的分类增益。
摘要：
学习模型的性能在很大程度上依赖于训练数据的可用性和充足性。为了解决数据集充足性问题，研究人员已广泛探索将数据增强（DA）作为一种很有前景的方法。DA 通过对可用数据施加变换来生成新的数据实例，从而增加数据集的规模和多样性。该方法提升了模型性能和准确率，尤其是在处理分类任务中的类别不平衡问题时。然而，很少有研究探索面向阿拉伯语的 DA，已有研究依赖释义或基于加噪技术等传统方法。本文提出一种新的阿拉伯语 DA 方法，该方法采用近期强大的建模技术 AraGPT-2 来执行增强过程。生成的句子使用 Euclidean、cosine、Jaccard 和 BLEU 距离，从上下文、语义、多样性和新颖性方面进行评估。最后，在情感分类任务上使用 AraBERT transformer 来评估增强后的阿拉伯语数据集的分类性能。实验在四个阿拉伯语情感数据集上进行：AraSarcasm、ASTD、ATT 和 MOVIE。所选数据集在规模、标签数量和类别不平衡方面各不相同。结果表明，所提出的方法在所有数据集上都提升了阿拉伯语情感文本分类性能，F1 分数分别在 AraSarcasm 上提高 4%、在 ASTD 上提高 6%、在 ATT 上提高 9%，以及在 MOVIE 上提高 13%。
英文摘要：
The performance of learning models heavily relies on the availability and adequacy of training data. To address the dataset adequacy issue, researchers have extensively explored data augmentation (DA) as a promising approach. DA generates new data instances through transformations applied to the available data, thereby increasing dataset size and variability. This approach has enhanced model performance and accuracy, particularly in addressing class imbalance problems in classification tasks. However, few studies have explored DA for the Arabic language, relying on traditional approaches such as paraphrasing or noising-based techniques. In this paper, we propose a new Arabic DA method that employs the recent powerful modeling technique, namely the AraGPT-2, for the augmentation process. The generated sentences are evaluated in terms of context, semantics, diversity, and novelty using the Euclidean, cosine, Jaccard, and BLEU distances. Finally, the AraBERT transformer is used on sentiment classification tasks to evaluate the classification performance of the augmented Arabic dataset. The experiments were conducted on four sentiment Arabic datasets: AraSarcasm, ASTD, ATT, and MOVIE. The selected datasets vary in size, label number, and unbalanced classes. The results show that the proposed methodology enhanced the Arabic sentiment text classification on all datasets with an increase in F1 score by 4% in AraSarcasm, 6% in ASTD, 9% in ATT, and 13% in MOVIE.
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。
