Agentic Hybrid RAG for Evidence-Grounded Muon Collider Analysis

Abstract

Muon collider research spans accelerator physics, detector instrumentation, and high-energy phenomenology, with relevant evidence scattered across a rapidly expanding and heterogeneous body of scientific literature. As high-energy physics (HEP) increasingly explores agent-assisted analysis workflows, efficiently locating, integrating, and verifying scientific evidence becomes an essential capability. While retrieval-augmented generation (RAG) offers a promising framework for scientific question answering, integrating agentic reasoning without compromising retrieval precision remains a key challenge. In this work, we present agentic hybrid RAG, an evidence-grounded RAG framework for muon collider research. The framework combines a hybrid retriever, integrating sparse lexical and dense semantic retrieval, with an agentic reasoning module for query decomposition, evidence expansion, and grounded answer generation. To enable systematic evaluation, we construct the first benchmark for retrieval-augmented scientific question answering in the muon collider domain, comprising a curated literature corpus together with dedicated retrieval and answer-generation benchmarks covering major detector and physics research topics. Extensive evaluation shows that hybrid retrieval provides the strongest retrieval backbone, while agentic reasoning is most effective for controlled evidence expansion and answer synthesis. Built on this principle, agentic hybrid RAG consistently outperforms representative retrieval and RAG baselines in retrieval effectiveness, answer quality, evidence coverage, and factual grounding. Together, the benchmark and framework provide a foundation for evidence-grounded scientific question answering and future HEP analysis agents operating over large-scale scientific literature. Code is available at \href{https://github.com/AItutorialjrb/RAG_muon_JINST}{this URL}.

Input to the SCAC Facilities Subcommittee from the U.S. Muon Collider Collaboration
S. Jindariani, T. Holmes, P. Meade +5
arXiv29 Sep 2026PreprintCC BY 4.0Full textApplied physics & instrumentation High-energy & particle physicsarXiv
FCC-μe: High Energy, High Luminosity Muon-Electron Collider & Neutrino Factory at CERN
K. Piotrzkowski
arXiv29 Oct 2026PreprintCC BY 4.0Full textApplied physics & instrumentation High-energy & particle physicsarXiv
Vanilla Scotogenic Model at the future Muon Collider
Niloy Mondal, Dipankar Pradhan, Abhik Sarkar +2
HAL15 Sep 2026PreprintarXiv licenceFull textCosmology High-energy & particle physicsHALarXiv
Report of the National Lab Accelerator Study Group for a Muon Collider
S. Gourlay, C. Mitchell, E. Nanni +5
arXiv4 Nov 2026PreprintCC BY 4.0Full textApplied physics & instrumentationarXiv
Neutrino Tridents and W-boson production at the Muon Collider
Felix Kling, Toni Mäkelä, Sebastian Trojanowski +1
arXiv12 Oct 2026PreprintCC BY 4.0Full textHigh-energy & particle physicsarXiv

Cite

BibTeX RIS

BibTeX

@misc{jiang2026agentic,
  title = {{Agentic Hybrid RAG for Evidence-Grounded Muon Collider Analysis}},
  author = {Ruobing Jiang and Dawei Fu and Cheng Jiang and Tianyi Yang and Zijian Wang and Youpeng Wu and Yong Ban and Yajun Mao and Qiang Li},
  howpublished = {arXiv},
  year = {2026},
  month = sep,
  doi = {10.48550/arxiv.2606.10381},
  eprint = {2606.10381},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2606.10381},
  note = {CC BY 4.0},
}

RIS

TY  - UNPB
TI  - Agentic Hybrid RAG for Evidence-Grounded Muon Collider Analysis
AU  - Ruobing Jiang
AU  - Dawei Fu
AU  - Cheng Jiang
AU  - Tianyi Yang
AU  - Zijian Wang
AU  - Youpeng Wu
AU  - Yong Ban
AU  - Yajun Mao
AU  - Qiang Li
PY  - 2026
DA  - 2026/09/28
PB  - arXiv
DO  - 10.48550/arxiv.2606.10381
UR  - https://arxiv.org/abs/2606.10381
AB  - Muon collider research spans accelerator physics, detector instrumentation, and high-energy phenomenology, with relevant evidence scattered across a rapidly expanding and heterogeneous body of scientific literature. As high-energy physics (HEP) increasingly explores agent-assisted analysis workflows, efficiently locating, integrating, and verifying scientific evidence becomes an essential capability. While retrieval-augmented generation (RAG) offers a promising framework for scientific question answering, integrating agentic reasoning without compromising retrieval precision remains a key challenge. In this work, we present agentic hybrid RAG, an evidence-grounded RAG framework for muon collider research. The framework combines a hybrid retriever, integrating sparse lexical and dense semantic retrieval, with an agentic reasoning module for query decomposition, evidence expansion, and grounded answer generation. To enable systematic evaluation, we construct the first benchmark for retrieval-augmented scientific question answering in the muon collider domain, comprising a curated literature corpus together with dedicated retrieval and answer-generation benchmarks covering major detector and physics research topics. Extensive evaluation shows that hybrid retrieval provides the strongest retrieval backbone, while agentic reasoning is most effective for controlled evidence expansion and answer synthesis. Built on this principle, agentic hybrid RAG consistently outperforms representative retrieval and RAG baselines in retrieval effectiveness, answer quality, evidence coverage, and factual grounding. Together, the benchmark and framework provide a foundation for evidence-grounded scientific question answering and future HEP analysis agents operating over large-scale scientific literature. Code is available at \href{https://github.com/AItutorialjrb/RAG_muon_JINST}{this URL}.
ER  -

Abstract

Related

Cite