arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2201.07434cs.CL
收藏

解释阿拉伯语Transformer模型

Interpreting Arabic Transformer Models

Ahmed Abdelali, Nadir Durrani, Fahim Dalvi, Hassan Sajjad

摘要:

阿拉伯语是一种闪米特语言,广泛使用,有许多方言。鉴于预先训练的语言模型的成功,许多在阿拉伯语及其方言上训练的Transformer模型已经浮出水面。虽然这些模型已经与下游NLP任务进行了比较,但没有进行直接比较内部表示的评估。我们探讨了语言信息是如何在阿拉伯语预训练模型中编码的,这些模型是在不同种类的阿拉伯语上训练的。我们使用三个内在任务对模型进行层和神经元分析:基于MSA(现代标准阿拉伯语)和方言POS标记的两个形态标记任务和方言识别任务。我们的分析启发了有趣的发现,例如:i)单词形态在较低层和中间层学习,ii)方言识别需要更多的知识,因此即使在最后一层也会保留,iii)尽管它们的词汇有很大的重叠,但基于MSA的模型无法捕捉阿拉伯方言的细微差别,iv)我们发现嵌入层中的神经元本质上是多义的,而中间层中的神经元专用于特定属性。

英文摘要:

Arabic is a Semitic language which is widely spoken with many dialects. Given the success of pre-trained language models, many transformer models trained on Arabic and its dialects have surfaced. While these models have been compared with respect to downstream NLP tasks, no evaluation has been carried out to directly compare the internal representations. We probe how linguistic information is encoded in Arabic pretrained models, trained on different varieties of Arabic language. We perform a layer and neuron analysis on the models using three intrinsic tasks: two morphological tagging tasks based on MSA (modern standard Arabic) and dialectal POS-tagging and a dialectal identification task. Our analysis enlightens interesting findings such as: i) word morphology is learned at the lower and middle layers ii) dialectal identification necessitate more knowledge and hence preserved even in the final layers, iii) despite a large overlap in their vocabulary, the MSA-based models fail to capture the nuances of Arabic dialects, iv) we found that neurons in embedding layers are polysemous in nature, while the neurons in middle layers are exclusive to specific properties.

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。

↑