Reproduce “Japanese Discourse Structure Analysis Based on Segmented Discourse Representation Theory Using LLMs” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Satomu Tomisaki, Yosuke Mikami, Amane Watahiki, Hitomi Yanaka
The authors test whether an incremental discourse parser can parse Japanese dialogue by translating English gaming corpora into Japanese and evaluating Llamipa, which predicts links and relation labels using prior clauses and discourse structure. On MSDC, Llama-3-based Llamipa reaches weighted-average F1 of 0.79 in both languages. The comparison is on translated data, not naturally produced Japanese conversations.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.