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
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:127971-127995, 2026.

Abstract

Retrieval augmented generation (RAG) has enhanced large language models by enabling access to external knowledge, with graph-based RAG emerging as a powerful paradigm for structured retrieval and reasoning. However, existing graph-based methods often over-rely on entity-centric node matching and lack explicit causal modeling, leading to unfaithful or spurious answers. Prior attempts to incorporate causality are typically limited to local or single-document contexts and also suffer from information isolation that arises from modular graph structures, which hinders scalability and cross-module causal reasoning. To address these challenges, we propose CausalRAG2, a framework that rethinks knowledge organization for graph-based RAG through causal gating across hierarchical modules. CausalRAG2 explicitly models causal relationships to suppress spurious correlations while enabling scalable reasoning over large-scale knowledge graphs. We also introduce HolisQA, a benchmark for holistic comprehension beyond entity-centric matching. Extensive experiments demonstrate that CausalRAG2 consistently outperforms competitive graph-based RAG baselines across multiple datasets and evaluation metrics. Our work establishes a principled foundation for structured, scalable, and causally grounded RAG systems.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-wang26ea, title = {{C}ausal{RAG}2: Hierarchical Causal Knowledge Graph Design for {RAG}}, author = {Wang, Nengbo and Liang, Tuo and Singh, Vikash and Song, Chaoda and Yang, Van and Yin, Yu and Ma, Jing and Singh, Jagdip and Chaudhary, Vipin}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {127971--127995}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/wang26ea/wang26ea.pdf}, url = {https://proceedings.mlr.press/v306/wang26ea.html}, abstract = {Retrieval augmented generation (RAG) has enhanced large language models by enabling access to external knowledge, with graph-based RAG emerging as a powerful paradigm for structured retrieval and reasoning. However, existing graph-based methods often over-rely on entity-centric node matching and lack explicit causal modeling, leading to unfaithful or spurious answers. Prior attempts to incorporate causality are typically limited to local or single-document contexts and also suffer from information isolation that arises from modular graph structures, which hinders scalability and cross-module causal reasoning. To address these challenges, we propose CausalRAG2, a framework that rethinks knowledge organization for graph-based RAG through causal gating across hierarchical modules. CausalRAG2 explicitly models causal relationships to suppress spurious correlations while enabling scalable reasoning over large-scale knowledge graphs. We also introduce HolisQA, a benchmark for holistic comprehension beyond entity-centric matching. Extensive experiments demonstrate that CausalRAG2 consistently outperforms competitive graph-based RAG baselines across multiple datasets and evaluation metrics. Our work establishes a principled foundation for structured, scalable, and causally grounded RAG systems.} }
Endnote
%0 Conference Paper %T CausalRAG2: Hierarchical Causal Knowledge Graph Design for RAG %A Nengbo Wang %A Tuo Liang %A Vikash Singh %A Chaoda Song %A Van Yang %A Yu Yin %A Jing Ma %A Jagdip Singh %A Vipin Chaudhary %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-wang26ea %I PMLR %P 127971--127995 %U https://proceedings.mlr.press/v306/wang26ea.html %V 306 %X Retrieval augmented generation (RAG) has enhanced large language models by enabling access to external knowledge, with graph-based RAG emerging as a powerful paradigm for structured retrieval and reasoning. However, existing graph-based methods often over-rely on entity-centric node matching and lack explicit causal modeling, leading to unfaithful or spurious answers. Prior attempts to incorporate causality are typically limited to local or single-document contexts and also suffer from information isolation that arises from modular graph structures, which hinders scalability and cross-module causal reasoning. To address these challenges, we propose CausalRAG2, a framework that rethinks knowledge organization for graph-based RAG through causal gating across hierarchical modules. CausalRAG2 explicitly models causal relationships to suppress spurious correlations while enabling scalable reasoning over large-scale knowledge graphs. We also introduce HolisQA, a benchmark for holistic comprehension beyond entity-centric matching. Extensive experiments demonstrate that CausalRAG2 consistently outperforms competitive graph-based RAG baselines across multiple datasets and evaluation metrics. Our work establishes a principled foundation for structured, scalable, and causally grounded RAG systems.
APA
Wang, N., Liang, T., Singh, V., Song, C., Yang, V., Yin, Y., Ma, J., Singh, J. & Chaudhary, V.. (2026). CausalRAG2: Hierarchical Causal Knowledge Graph Design for RAG. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:127971-127995 Available from https://proceedings.mlr.press/v306/wang26ea.html.

Related Material