CDF-RAG: Causal Dynamic Feedback for Adaptive Retrieval-Augmented Generation

2025-04-17

AI Summary

Existing RAG frameworks rely primarily on semantic matching, which struggles to distinguish genuine causal relationships from spurious correlations, resulting in generated responses lacking mechanistic causal explanations. To address this, we propose CausalRAG—the first retrieval-augmented generation framework integrating causal dynamic feedback. CausalRAG innovatively unifies five core components: causal discovery, dynamic query rewriting, structured causal graph retrieval, multi-hop causal reasoning, and causal path verification—ensuring logical coherence and causal consistency throughout inference. Evaluated on four knowledge-intensive benchmarks, CausalRAG achieves significant improvements in response accuracy (+7.2%) and causal correctness (+12.6%). All code and models are publicly released.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Knowledge Representation and Reasoning: Action, Change, and Causality

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Retrieval-Augmented Generation (RAG) has significantly enhanced large language models (LLMs) in knowledge-intensive tasks by incorporating external knowledge retrieval. However, existing RAG frameworks primarily rely on semantic similarity and correlation-driven retrieval, limiting their ability to distinguish true causal relationships from spurious associations. This results in responses that may be factually grounded but fail to establish cause-and-effect mechanisms, leading to incomplete or misleading insights. To address this issue, we introduce Causal Dynamic Feedback for Adaptive Retrieval-Augmented Generation (CDF-RAG), a framework designed to improve causal consistency, factual accuracy, and explainability in generative reasoning. CDF-RAG iteratively refines queries, retrieves structured causal graphs, and enables multi-hop causal reasoning across interconnected knowledge sources. Additionally, it validates responses against causal pathways, ensuring logically coherent and factually grounded outputs. We evaluate CDF-RAG on four diverse datasets, demonstrating its ability to improve response accuracy and causal correctness over existing RAG-based methods. Our code is publicly available at https://github.com/ elakhatibi/CDF-RAG.

Problem

• Enhancing causal reasoning in Retrieval-Augmented Generation (RAG) systems
• Distinguishing true causal relationships from spurious associations
• Improving factual accuracy and explainability in generative reasoning

Innovation

• Iteratively refines queries for causal accuracy
• Retrieves structured causal graphs dynamically
• Validates responses against causal pathways