Researchers from AC at the KI Institute will present a paper tackling key challenges in explainable graph RAG. The work, titled “ArgRAG: Explainable Retrieval-Augmented Generation using Quantitative Bipolar Argumentation”, will appear at NeSy, a leading conference on Neurosymbolic Learning and Reasoning. Here is a summary of the paper:
ArgRAG: Explainable Retrieval-Augmented Generation using Quantitative Bipolar Argumentation
Standard RAG pipelines retrieve unstructured content and generate answers—but often lack transparency. ArgRAG proposes a neurosymbolic alternative that builds a structured argumentation graph from retrieved evidence.
By applying deterministic reasoning over pro and con arguments, ArgRAG produces not only accurate answers but also explainable and contestable ones.
This method bridges symbolic reasoning and neural generation, moving toward more trustworthy AI decision-making.
Zhu, Yuqicheng, Nico Potyka, Daniel Hernández, Yuan He, Zifeng Ding, Bo Xiong, Dongzhuoran Zhou, Evgeny Kharlamov, and Steffen Staab. ‘ArgRAG: Explainable Retrieval Augmented Generation Using Quantitative Bipolar Argumentation’. arXiv [Cs.AI], 2025. arXiv. http://arxiv.org/abs/2508.20131 .