arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2407.06547cs.CL
收藏

基于Featural InfoWaveGAN解读阿萨姆语元音和谐

Deciphering Assamese Vowel Harmony with Featural InfoWaveGAN

Indian Institute of Technology Guwahati(古瓦哈蒂印度理工学院)

Sneha Ray Barman, Shakuntala Mahanta, Neeraj Kumar Sharma

AI总结:

本研究针对传统音系学习依赖文本数据的局限,用Featural InfoWaveGAN从原始语音学习阿萨姆语元音和谐,验证模型可掌握音位配列规则并实现特征学习,为语音驱动的音系习得研究提供新路径。

摘要:

传统音系学习研究方法主要依赖人工整理的文本数据,这类方法虽具启发性,但会限制口语的文本表征所承载的信息。为突破这一局限,我们探究了Featural InfoWaveGAN模型利用原始语音数据学习迭代式长距离元音和谐的潜力。我们聚焦阿萨姆语——该语言以音系上的逆向、词界元音和谐著称。研究表明,该模型擅长掌握阿萨姆语音位配列的复杂特征,尤其是逆向的迭代式长距离和谐;它还会生成类似人类语言习得中言语错误的非迭代非法形式。统计分析显示,在新测试项中,模型倾向于将特定的[+高,+ATR]元音作为触发项,这表明模型实现了特征学习。增加数据量与设置对照实验可提升模型性能,这与学习的普遍性形成对比。

英文摘要:

Traditional approaches for understanding phonological learning have predominantly relied on curated text data. Although insightful, such approaches limit the knowledge captured in textual representations of the spoken language. To overcome this limitation, we investigate the potential of the Featural InfoWaveGAN model to learn iterative long-distance vowel harmony using raw speech data. We focus on Assamese, a language known for its phonologically regressive and word-bound vowel harmony. We demonstrate that the model is adept at grasping the intricacies of Assamese phonotactics, particularly iterative long-distance harmony with regressive directionality. It also produced non-iterative illicit forms resembling speech errors during human language acquisition. Our statistical analysis reveals a preference for a specific [+high,+ATR] vowel as a trigger across novel items, indicative of feature learning. More data and control could improve model proficiency, contrasting the universality of learning.

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。

↑