ArgRAG: Explainable Retrieval Augmented Generation using Quantitative Bipolar Argumentation

Yuqicheng Zhu, Nico Potyka, Daniel Hernández, Yuan He, Zifeng Ding, Bo Xiong, Dongzhuoran Zhou, Evgeny Kharlamov, Steffen Staab
Proceedings of The 19th International Conference on Neurosymbolic Learning and Reasoning, PMLR 284:697-718, 2025.

Abstract

Retrieval-Augmented Generation (RAG) enhances large language models by incorporating external knowledge, yet suffers from critical limitations in high-stakes domains—namely, sensitivity to noisy or contradictory evidence and opaque, stochastic decision-making. We propose \textsc{ArgRAG}, an explainable, and contestable alternative that replaces black-box reasoning with structured inference using a Quantitative Bipolar Argumentation Framework (QBAF). \textsc{ArgRAG} constructs a QBAF from retrieved documents and performs deterministic reasoning under gradual semantics. This allows faithfully explanaining and contesting decisions. Evaluated on two fact verification benchmarks, PubHealth and RAGuard, \textsc{ArgRAG} achieves strong accuracy while significantly improving transparency.

Cite this Paper

BibTeX
@InProceedings{pmlr-v284-zhu25a, title = {ArgRAG: Explainable Retrieval Augmented Generation using Quantitative Bipolar Argumentation}, author = {Zhu, Yuqicheng and Potyka, Nico and Hern\'{a}ndez, Daniel and He, Yuan and Ding, Zifeng and Xiong, Bo and Zhou, Dongzhuoran and Kharlamov, Evgeny and Staab, Steffen}, booktitle = {Proceedings of The 19th International Conference on Neurosymbolic Learning and Reasoning}, pages = {697--718}, year = {2025}, editor = {H. Gilpin, Leilani and Giunchiglia, Eleonora and Hitzler, Pascal and van Krieken, Emile}, volume = {284}, series = {Proceedings of Machine Learning Research}, month = {08--10 Sep}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v284/main/assets/zhu25a/zhu25a.pdf}, url = {https://proceedings.mlr.press/v284/zhu25a.html}, abstract = {Retrieval-Augmented Generation (RAG) enhances large language models by incorporating external knowledge, yet suffers from critical limitations in high-stakes domains—namely, sensitivity to noisy or contradictory evidence and opaque, stochastic decision-making. We propose \textsc{ArgRAG}, an explainable, and contestable alternative that replaces black-box reasoning with structured inference using a Quantitative Bipolar Argumentation Framework (QBAF). \textsc{ArgRAG} constructs a QBAF from retrieved documents and performs deterministic reasoning under gradual semantics. This allows faithfully explanaining and contesting decisions. Evaluated on two fact verification benchmarks, PubHealth and RAGuard, \textsc{ArgRAG} achieves strong accuracy while significantly improving transparency.} }
Endnote
%0 Conference Paper %T ArgRAG: Explainable Retrieval Augmented Generation using Quantitative Bipolar Argumentation %A Yuqicheng Zhu %A Nico Potyka %A Daniel Hernández %A Yuan He %A Zifeng Ding %A Bo Xiong %A Dongzhuoran Zhou %A Evgeny Kharlamov %A Steffen Staab %B Proceedings of The 19th International Conference on Neurosymbolic Learning and Reasoning %C Proceedings of Machine Learning Research %D 2025 %E Leilani H. Gilpin %E Eleonora Giunchiglia %E Pascal Hitzler %E Emile van Krieken %F pmlr-v284-zhu25a %I PMLR %P 697--718 %U https://proceedings.mlr.press/v284/zhu25a.html %V 284 %X Retrieval-Augmented Generation (RAG) enhances large language models by incorporating external knowledge, yet suffers from critical limitations in high-stakes domains—namely, sensitivity to noisy or contradictory evidence and opaque, stochastic decision-making. We propose \textsc{ArgRAG}, an explainable, and contestable alternative that replaces black-box reasoning with structured inference using a Quantitative Bipolar Argumentation Framework (QBAF). \textsc{ArgRAG} constructs a QBAF from retrieved documents and performs deterministic reasoning under gradual semantics. This allows faithfully explanaining and contesting decisions. Evaluated on two fact verification benchmarks, PubHealth and RAGuard, \textsc{ArgRAG} achieves strong accuracy while significantly improving transparency.
APA
Zhu, Y., Potyka, N., Hernández, D., He, Y., Ding, Z., Xiong, B., Zhou, D., Kharlamov, E. & Staab, S.. (2025). ArgRAG: Explainable Retrieval Augmented Generation using Quantitative Bipolar Argumentation. Proceedings of The 19th International Conference on Neurosymbolic Learning and Reasoning, in Proceedings of Machine Learning Research 284:697-718 Available from https://proceedings.mlr.press/v284/zhu25a.html.

Related Material