25–29 May 2026
Chulalongkorn University
Asia/Bangkok timezone

Dr.Sai: A Pioneering LLM-Based Autonomous Agent for Physics Analysis

26 May 2026, 16:15
18m
MHMK 202

MHMK 202

Oral Presentation Track 4 - Distributed computing Track 4 - Distributed computing

Speaker

Zhengde Zhang (中国科学院高能物理研究所)

Description

We present the Dr.Sai series—a suite of multi-agent systems driven by large language models. The system interprets natural language requests from physicists, decomposes them into tasks such as data skimming, fitting, and others, calls relevant scientific tools, and executes the entire workflow end-to-end.
We will demonstrate its latest advances in decay-chain analysis, scattering data refinement, and AI-ready data processing, while focusing on the core technologies that enable these capabilities—including connecting agents with experimental data, integrating specialized computing environments such as BOSS, supporting seamless model switching, implementing human-AI interaction mechanisms, and managing long-running tasks.
By automating complex workflows and ensuring full traceability, Dr.Sai significantly enhances research efficiency and offers a practical blueprint for an "AI Scientist" assistant in data-intensive scientific fields.

Authors

Beijiang Liu Changzheng YUAN Mr Fazhi Qi Ke Li (Chinese Academy of Sciences (CN)) Zhengde Zhang (中国科学院高能物理研究所)

Presentation materials

There are no materials yet.