Automatic Paper Reviewing with Heterogeneous Graph Reasoning over LLM-Simulated Reviewer-Author Debates
Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (37) · 31717-31725
DOI: 10.1609/aaai.v40i37.40439 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Shuaimin Li und 7 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref).
Abstract (Originalsprache)
Existing paper review methods often rely on superficial manuscript features or directly on large language models (LLMs), which are prone to hallucinations, biased scoring, and limited reasoning capabilities. Moreover, these methods often fail to capture the complex argumentative reasoning and negotiation dynamics inherent in reviewer-author interactions. To address these limitations, we propose ReViewGraph (Reviewer-Author Debates Graph Reasoner), a novel framework that performs heterogeneous graph reasoning over LLM-simulated multi-round reviewer-author debates. In our approach, reviewer-author exchanges are simulated through LLM-based multi-agent collaboration. Diverse opinion relations (e.g., acceptance, rejection, clarification, and compromise) are then explicitly extracted and encoded as typed edges within a heterogeneous interaction graph. By applying graph neural networks to reason over these structured debate graphs, ReViewGraph captures fine-grained argumentative dynamics and enables more informed review decisions. Extensive experiments on three datasets demonstrate that ReViewGraph outperforms strong baselines with an average relative improvement of 15.73%, underscoring the value of modeling detailed reviewer–author debate structures.
Zitieren
APA
Li, S., Fan, L., Lin, Y., Li, Z., Wei, X., Ni, S., Alinejad-Rokny, H., & Yang, M. (2026). Automatic Paper Reviewing with Heterogeneous Graph Reasoning over LLM-Simulated Reviewer-Author Debates. Proceedings of the AAAI Conference on Artificial Intelligence, 40(37), 31717-31725. https://doi.org/10.1609/aaai.v40i37.40439
Vancouver
Li S, Fan L, Lin Y, Li Z, Wei X, Ni S, et al. Automatic Paper Reviewing with Heterogeneous Graph Reasoning over LLM-Simulated Reviewer-Author Debates. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(37):31717-31725. doi:10.1609/aaai.v40i37.40439
BibTeX
@article{Li2026,
author = {Li, Shuaimin and Fan, Liyang and Lin, Yufang and Li, Zeyang and Wei, Xian and Ni, Shiwen and Alinejad-Rokny, Hamid and Yang, Min},
title = {Automatic Paper Reviewing with Heterogeneous Graph Reasoning over LLM-Simulated Reviewer-Author Debates},
journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2026},
volume = {40},
number = {37},
pages = {31717--31725},
doi = {10.1609/aaai.v40i37.40439}
}
EUVIMED