ArgRAG: Explainable Retrieval Augmented Generation using Quantitative Bipolar Argumentation

2025-08-26

AI Summary

Retrieval-augmented generation (RAG) faces two critical challenges in high-stakes applications: sensitivity to noisy or contradictory evidence and opaque, uncontrollable reasoning. To address these, we propose the Quantitative Bipolar Argumentation Framework (QBAF), which models retrieved documents as structured argumentation networks and replaces black-box generation with deterministic, interpretable inference via progressive semantic fusion. QBAF enables contestable and traceable decision-making, substantially improving reliability. Evaluated on the PubHealth and RAGuard fact-checking benchmarks, our method achieves state-of-the-art accuracy while providing fine-grained attribution and conflict resolution—marking the first RAG approach to jointly deliver high precision, strong robustness against evidence noise, and end-to-end interpretability across the entire reasoning pipeline.

Technology Category

• Knowledge Representation and Reasoning: Argumentation
• Reasoning under Uncertainty: Other Foundations of Reasoning under Uncertainty
• Natural Language Processing: Discourse, Pragmatics & Argument Mining

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

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 ArgRAG, an explainable, and contestable alternative that replaces black-box reasoning with structured inference using a Quantitative Bipolar Argumentation Framework (QBAF). ArgRAG constructs a QBAF from retrieved documents and performs deterministic reasoning under gradual semantics. This allows faithfully explaining and contesting decisions. Evaluated on two fact verification benchmarks, PubHealth and RAGuard, ArgRAG achieves strong accuracy while significantly improving transparency.

Problem

• Addressing opaque decision-making in RAG systems
• Handling noisy and contradictory evidence retrieval
• Providing explainable and contestable reasoning processes

Innovation

• Quantitative Bipolar Argumentation Framework for structured inference
• Deterministic reasoning under gradual semantics for explainability
• Constructing argumentation frameworks from retrieved documents