Research map / Graph learning / Knowledge and temporal graphs
CausalRAG2: Hierarchical Causal Knowledge Graph Design for RAG
Nengbo Wang, Tuo Liang, Vikash Singh, Chaoda Song, Van Yang, Yu Yin, Jing Ma, JAGDIP SINGH, Vipin Chaudhary
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add hierarchy only (no gates, no causal filtering) (ablation) | CausalRAG2 · HolisQA ablation | F1 / Context Recall: F1 26.8 -> 24.0; CR 54.7 -> 58.0 | Mixed |
| add causal gates on top of hierarchy, without causal filtering (ablation) | CausalRAG2 · HolisQA ablation | F1 / Context Recall: F1 24.0 -> 23.3; CR 58.0 -> 60.2 | Mixed |
| add causal path filtering to structure without it (ablation) | CausalRAG2 · HolisQA ablation | F1: 23.3 -> 36.8 | Improved |
| add causal filtering alone (flat graph, no hierarchy or gates) (ablation) | flat graph RAG · HolisQA ablation | F1: 26.8 -> 30.1 | Improved |
| use spurious-aware causal identification instead of standard (ablation) | CausalRAG2 · HolisQA ablation | F1: 36.8 -> 38.6 | Improved |
| add hierarchical causal organization to a flat causal RAG | CausalRAG · five standard QA datasets | F1: mean +15.5 F1; +24.83 F1 on HotpotQA; +0.62 F1 on HolisQA-CompSci | Improved |
| enlarge seed budget (K0, KL) (ablation) | CausalRAG2 · HolisQA-CompSci sensitivity | Context Recall / F1 | Mixed |
| vary wcausal, gamma and seed budgets (ablation) | CausalRAG2 · HolisQA-CompSci sensitivity | F1, CR, AR | No clear effect |
| use larger-capacity LLM backbone for causal filtering (ablation) | CausalRAG2 · Appendix H stability | subgraph stability | Improved |
| add causal gating and filtering at 1.5M-character corpus scale | CausalRAG2 · HolisQA scalability | mean of F1, CR, AR | Improved |
| add causal gate verification offline | CausalRAG2 · HolisQA cost | offline construction cost per 1K tokens: $0.0015 vs LightRAG $0.0015; GraphRAG $0.0034 | No clear effect |
Closest papers by meaning
- Simple is Effective: The Roles of Graphs and Large Language Models in Knowledge-Graph-Based Retrieval-Augmented Generation (ICLR 2025 · similarity 0.90)
- HyperGraphRAG: Retrieval-Augmented Generation via Hypergraph-Structured Knowledge Representation (NeurIPS 2025 · similarity 0.90)
- Think-on-Graph 2.0: Deep and Faithful Large Language Model Reasoning with Knowledge-guided Retrieval Augmented Generation (ICLR 2025 · similarity 0.90)
- A Systematic Exploration of Knowledge Graph Alignment with Large Language Models in Retrieval Augmented Generation (AAAI 2025 · similarity 0.89)
- Can Knowledge-Graph-based Retrieval Augmented Generation Really Retrieve What You Need? (NeurIPS 2025 · similarity 0.89)
- ReMindRAG: Low-Cost LLM-Guided Knowledge Graph Traversal for Efficient RAG (NeurIPS 2025 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
