Guo Chen

Guo Chen

Undergraduate student of Southwest University, focusing on Retrieval-Augmented Generation and Large Language Models.

Publications

You can also find my articles on my Google Scholar profile.

ForestBench: A Unified Graph Framework for Evaluating Multi-Agent Collaboration

Published in Under Review, 2026

We introduce a unified graph-based evaluation framework and ForestBench, which maps heterogeneous multi-agent execution traces into comparable collaboration graphs and evaluates them against query-specific forests of verified-success references.

Recommended citation: Chen, Guo, Ziwen Li, Reed Li, Yu Lu, Haibo Shi, Bingbing Xu, and Junjie Huang. "ForestBench: A Unified Graph Framework for Evaluating Multi-Agent Collaboration." arXiv preprint arXiv:2608.08605 (2026).
Download Paper | Download Bibtex

NGM-RAG: Neural Graph Matching based Retrieval-Augmented Generation

Published in The 23rd CCF Conference on Web Information Systems and Applications, 2026

NGM-RAG is a novel RAG framework combined with graph structure, which improves the retrieval and generation effects in multi-hop reasoning and complex question-answering tasks by integrating multiple matching methods.

Recommended citation: Chen, Guo, Ziwen Li, Maolin Zheng, Hao Gao, Junjie Huang, and Tao Jia. "NGM-RAG: Neural Graph Matching based Retrieval-Augmented Generation." arXiv preprint arXiv:2607.11159 (2026).
Download Paper | Download Bibtex

RISE: A Retrieval-Augmented Generation Enhanced Immersive System for Education

Published in Companion Proceedings of the ACM Web Conference, 2026

We have developed an immersive education system based on RAG, MCP, and virtual digital humans.

Recommended citation: Chen, Guo, Haowei Tang, Yan Wei, Yuming Wang, Haoyang Zhang, Yikang Hou, Maolin Zheng, and Junjie Huang. "Rise: A Retrieval-Augmented Generation Enhanced Immersive System for Education." In Companion Proceedings of the ACM Web Conference 2026, pp. 85-88. 2026.
Download Paper | Download Bibtex

LiR^3AG: A Lightweight Rerank Reasoning Strategy Framework for Retrieval-Augmented Generation

Published in The Association for the Advancement of Artificial Intelligence, 2026

We explored the reasoning strategies of reasoning models in RAG multi-hop QA tasks, finally proposed a lightweight framework that improves performance while significantly reducing reasoning overhead by integrating three modules.

Recommended citation: Chen, Guo, Junjie Huang, Huaijin Xie, Fei Sun, and Tao Jia. "LiR3AG: A Lightweight Rerank Reasoning Strategy Framework for Retrieval-Augmented Generation." In Proceedings of the AAAI Conference on Artificial Intelligence, vol. 40, no. 36, pp. 30201-30209. 2026.
Download Paper | Download Bibtex