TCDA: Thread-Constrained Discourse-Aware Modeling for Conversational Sentiment Quadruple Analysis

Xinran Li, Xinze Che, Yifan Lyu, Zhiqi Huang, Xiujuan Xu · 2026

Read TCDA: Thread-Constrained Discourse-Aware Modeling for Conversational Sentiment Quadruple Analysis by Xinran Li, Xinze Che, Yifan Lyu, Zhiqi Huang, Xiujuan Xu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-01717.txt.

Description

Xinran Li and four co-authors extract sentiment quadruples from multi-round dialogue: who holds an opinion, about what aspect, with what polarity, in conversation that jumps between threads. Graph networks used for this task import structural noise and ignore speaking order, while standard position codes flatten the difference between word order inside a reply and progress across replies.

Their framework answers both faults. A thread-constrained directed acyclic graph filters cross-thread noise, keeps the dialogue connected through a root anchor, and respects time order. A discourse-aware rotary position code then aligns meaning across layers, measures thread distance in tree steps, and adds utterance-level progress so distant tokens stop diluting. On two benchmark data sets the combined design posts state-of-the-art scores for conversational quadruple analysis.

Catalog record

CREATORXinran Li, Xinze Che, Yifan Lyu, Zhiqi Huang, Xiujuan Xu
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE31 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.01717

Browse the full library · About the Tower of Babel