arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2404.08997cs.CL
收藏

基于半马尔可夫模型的带标注形态切分

Labeled Morphological Segmentation with Semi-Markov Models

Johns Hopkins University(约翰斯·霍普金斯大学);LMU Munich(慕尼黑大学)

Ryan Cotterell, Thomas Müller, Alexander Fraser, Hinrich Schütze

AI总结:

该研究提出带标注形态切分的新形态处理视角,引入形态句法标签集层级,开发显式建模形态句法的判别式系统\textsc{chipmunk},在六种语言的三项形态任务上均提升性能,形态切分F1值较基线绝对提升2-6个点。

摘要:

我们提出带标注形态切分,这是一种统一了多项任务的形态处理新视角。从标注角度,我们还引入了新的形态句法标签集层级体系。最后,我们开发了\textsc{chipmunk}(一种判别式形态切分系统),与以往研究不同,它显式地对形态句法建模。我们表明,\textsc{chipmunk}在六种语言的三项任务上均提升了性能:(i)形态切分、(ii)词干提取、(iii)形态标签分类。在形态切分任务上,我们的方法相比基线的$F_1$值绝对提升了2–6个点。

英文摘要:

We present labeled morphological segmentation, an alternative view of morphological processing that unifies several tasks. From an annotation standpoint, we additionally introduce a new hierarchy of morphotactic tagsets. Finally, we develop \modelname, a discriminative morphological segmentation system that, contrary to previous work, explicitly models morphotactics. We show that \textsc{chipmunk} yields improved performance on three tasks for all six languages: (i) morphological segmentation, (ii) stemming and (iii) morphological tag classification. On morphological segmentation, our method shows absolute improvements of 2--6 points $F_1$ over the baseline.

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。

↑