chATLAS: A Retrieval-Augmented Conversational Assistant for the ATLAS Experiment
Abstract
Large scientific collaborations produce extensive internal documentation spread across heterogeneous systems, making it difficult for collaborators to locate and reuse existing knowledge. We present chATLAS, a deployed Retrieval-Augmented Generation (RAG) assistant developed for the ATLAS experiment at CERN. The system provides conversational access to more than 30,000 documents from several collaboration resources, material largely unavailable to general-purpose language models, using hybrid dense--sparse retrieval deployed on CERN infrastructure. A domain-specific benchmark built from automatically generated question--answer pairs is used to evaluate retrieval and generation performance and to guide design choices. We study the selection and domain-specific fine-tuning of an embedding model, contextual retrieval, ranking strategies, and generator choice. Fine-tuning combined with contextual retrieval more than doubles how often the correct document is returned as the top result on the most technically challenging documentation source. The deployed system also includes intent routing, safety handling, Mattermost integration, and an MCP server for future agentic workflows. chATLAS demonstrates that domain-adapted RAG can provide source-attributed, domain-specific access to collaboration knowledge, and offers a practical foundation for scientific collaborations and other organisations that rely on accurate, up-to-date retrieval from proprietary documentation.
Related
- Qlippy: A Retrieval-Augmented GenAI Assistant for Reproducible Quantum Workflows and Experiment Tracking
- Integration of Retrieval-Augmented Generation for Knowledge Access in the ELBE Accelerator Control System
- Towards Retrieval Augmented Generation in High-Energy and Astroparticle Physics
- SALSA: Semi-Autonomous Literature Summarization Assistant
- BanglaShop-CRS: A User-Centric Bangla Dataset for Conversational Recommendation
BibTeX
@misc{egan2026chatlas,
title = {{chATLAS: A Retrieval-Augmented Conversational Assistant for the ATLAS Experiment}},
author = {Joe Egan and Ben Elliot and Samuel Van Stroud and Alex Sopio and Ki Ryeong Park and Daniele Dal Santo and Zhaohui Li and Runze Li and Jeremy Couthures and Joe George and Cary Randazzo and Juerg Beringer and Daniel Murnane and Gabriel Facini},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.37697},
eprint = {2610.37697},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.37697},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - chATLAS: A Retrieval-Augmented Conversational Assistant for the ATLAS Experiment
AU - Joe Egan
AU - Ben Elliot
AU - Samuel Van Stroud
AU - Alex Sopio
AU - Ki Ryeong Park
AU - Daniele Dal Santo
AU - Zhaohui Li
AU - Runze Li
AU - Jeremy Couthures
AU - Joe George
AU - Cary Randazzo
AU - Juerg Beringer
AU - Daniel Murnane
AU - Gabriel Facini
PY - 2026
DA - 2026/10/24
PB - arXiv
DO - 10.48550/arxiv.2610.37697
UR - https://arxiv.org/abs/2610.37697
AB - Large scientific collaborations produce extensive internal documentation spread across heterogeneous systems, making it difficult for collaborators to locate and reuse existing knowledge. We present chATLAS, a deployed Retrieval-Augmented Generation (RAG) assistant developed for the ATLAS experiment at CERN. The system provides conversational access to more than 30,000 documents from several collaboration resources, material largely unavailable to general-purpose language models, using hybrid dense--sparse retrieval deployed on CERN infrastructure. A domain-specific benchmark built from automatically generated question--answer pairs is used to evaluate retrieval and generation performance and to guide design choices. We study the selection and domain-specific fine-tuning of an embedding model, contextual retrieval, ranking strategies, and generator choice. Fine-tuning combined with contextual retrieval more than doubles how often the correct document is returned as the top result on the most technically challenging documentation source. The deployed system also includes intent routing, safety handling, Mattermost integration, and an MCP server for future agentic workflows. chATLAS demonstrates that domain-adapted RAG can provide source-attributed, domain-specific access to collaboration knowledge, and offers a practical foundation for scientific collaborations and other organisations that rely on accurate, up-to-date retrieval from proprietary documentation.
ER -