arXiv cs.CL· Satomu Tomisaki, Yosuke Mikami, Amane Watahiki, Hitomi Yanaka·· 11 小时前AI 评分18
基于 SDRT 用 LLM 分析日语篇章结构
Japanese Discourse Structure Analysis Based on Segmented Discourse Representation Theory Using LLMs
AI 导读
研究用 LLM 翻译现有英文 SDRT 语料库，构建出日语 SDRT 语料库，并对比基于提示词与微调的 LLM 表现，发现模型在英语和日语上的篇章解析性能相当，说明基于 SDRT 的篇章解析具备跨语言适用性。消融分析显示，所考察的多数日语篇章标记影响很小，模型预测更依赖上下文中的篇章关系而非上下文子句内容，这种结构性依赖或可解释英日类型学差异显著却性能相近的现象。
正文 · 原文
Abstract:Discourse parsing based on Segmented Discourse Representation Theory (SDRT) is crucial for capturing the structural dynamics of dialogues. However, most prior research focuses on English, and its effectiveness for Japanese remains unclear owing to the lack of Japanese SDRT corpora. To address this issue, we construct Japanese SDRT corpora by translating existing English SDRT corpora using an LLM. Through a comparative evaluation of prompt-based and fine-tuned LLMs, we show that the models achieve comparable performance between English and Japanese, suggesting the cross-lingual applicability of SDRT-based discourse parsing. In order to investigate why the English and Japanese parsers achieve similar performance, we conduct ablation analyses on discourse markers and contextual inputs. Our results reveal that most of the Japanese discourse markers we examined have little impact, and the models' predictions depend heavily on contextual discourse relations rather than on the content of the context clauses. This structural reliance might explain the comparable performance despite the substantial typological differences between English and Japanese.
| Comments: | To appear in the Proceedings of the Asia-Pacific Chapter of the Association for Computational Linguistics: Student Research Workshop (AACL-SRW 2026) |
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.36667 [cs.CL] |
| (or arXiv:2610.36667v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.36667 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Satomu Tomisaki [view email]
[v1]
Tue, 27 Oct 2026 05:39:29 UTC (185 KB)
来源：arXiv cs.CL · arxiv.org