arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2404.04530cs.CL
收藏

基于形态学的位置编码研究

A Morphology-Based Investigation of Positional Encodings

IIT Bombay(印度理工学院孟买分校);Google Research(谷歌研究院);NICT(日本情报通信研究机构)

Poulami Ghosh, Shikhar Vashishth, Raj Dabre, Pushpak Bhattacharyya

AI总结:

本研究探讨语言形态学复杂性与预训练语言模型中位置编码利用率的相关性,通过对22种语言和5项下游任务的分析,发现形态学复杂性越高,位置编码的重要性越低,为改进位置编码提供了依据。

摘要:

当代深度学习模型能有效处理具有多样形态学的语言,尽管并未直接将形态学整合其中。形态学与词序密切相关,后者通过位置编码融入基于 Transformer 的模型中。这引发了一个根本性问题:语言的形态学复杂性与预训练语言模型中位置编码的利用率之间是否存在相关性?为寻求答案,我们提出了首项针对此问题的研究,涵盖 22 种语言和 5 项下游任务。我们的发现表明,随着语言形态学复杂性的增加,位置编码的重要性逐渐降低。本研究推动了对位置编码更深入理解的需求,以增强其更好地反映所考虑的不同语言。

英文摘要:

Contemporary deep learning models effectively handle languages with diverse morphology despite not being directly integrated into them. Morphology and word order are closely linked, with the latter incorporated into transformer-based models through positional encodings. This prompts a fundamental inquiry: Is there a correlation between the morphological complexity of a language and the utilization of positional encoding in pre-trained language models? In pursuit of an answer, we present the first study addressing this question, encompassing 22 languages and 5 downstream tasks. Our findings reveal that the importance of positional encoding diminishes with increasing morphological complexity in languages. Our study motivates the need for a deeper understanding of positional encoding, augmenting them to better reflect the different languages under consideration.

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。

↑