Context-Gloss Augmentation for Improving Arabic Target Sense Verification

Abstract

Arabic language lacks semantic datasets and sense inventories. The most common semantically-labeled dataset for Arabic is the ArabGlossBERT, a relatively small dataset that consists of 167K context-gloss pairs (about 60K positive and 107K negative pairs), collected from Arabic dictionaries. This paper presents an enrichment to the ArabGlossBERT dataset, by augmenting it using (Arabic-English-Arabic) machine back-translation. Augmentation increased the dataset size to 352K pairs (149K positive and 203K negative pairs). We measure the impact of augmentation using different data configurations to fine-tune BERT on target sense verification (TSV) task. Overall, the accuracy ranges between 78% to 84% for different data configurations. Although our approach performed at par with the baseline, we did observe some improvements for some POS tags in some experiments. Furthermore, our fine-tuned models are trained on a larger dataset covering larger vocabulary and contexts. We provide an in-depth analysis of the accuracy for each part-of-speech (POS).

Data Augmentation Using Transformers and Similarity Measures for Improving Arabic Text Classification

“…While few studies have focused on augmenting Arabic data [28]- [31], [49], some have used the current DA noising and paraphrasing-based approaches without employing the transformer's powerful models as augmentation techniques [29]- [31], [50]. Other studies have employed transformers to evaluate the augmented Arabic dataset [34], [35].…”

“…Swapping [4], [22] Deletion [4], [9] Insertion [4], [23] Substitution [4], [24] Mixup [4], [15] Paraphrasing-based techniques Thesauruses [6], [12] Rules [13], [14] Machine translation [21] Transformers [13], [35], [47], [48] English Noising-based techniques Swapping [28], [36] Deletion [28], [49] Insertion [29], [31], [53] Substitution [30], [31], [36], [53] Mixup [30], [31], [39], [53] Arabic Paraphrasing-based techniques Thesauruses [31], [34], [40] Rules [29], [31], [35] Machine translation [50] Transformers…”