Speaker
Description
We present the NRS Agent, a data analysis agent system for Nuclear Resonant Scattering (NRS) experiments conducted at synchrotron light sources. The NRS Agent is a multi-agent assistant designed for end-to-end Nuclear Resonant Scattering workflows aimed at reducing manual effort and accelerating complex analyses. The agent is designed specifically for data recorded by the High Energy Photon Source (HEPS), a high-energy synchrotron light source under construction in China, although it may also be applicable to other light sources. This agent is developed using the Dr.Sai framework, an AI-driven multi-agent system designed to automate end-to-end physics analysis workflows. The major NRS analysis domains covered by this project are the Coherent Nuclear Resonant Scattering and Nuclear Resonant Inelastic X-ray Scattering (NRIXS) experiments. To provide a comprehensive analytical framework, the NRS Agent integrates the standard analysis packages CONUSS and PHOENIX, which are used for coherent scattering modeling/fitting and extracting material phonon properties from NRIXS data, respectively. The agent automates the entire analysis workflow from inspection of raw measured data and data type conversion to create analysis-ready datasets to generation of structured configurations and execution of CONUSS and PHOENIX pipelines. Broad parameter control, reproducible workflow management, high-throughput parallel analyses across multiple datasets and physical configurations, and wise iterative fitting are among the features that contribute to the agent's utility for end users. This agent serves as a powerful assistant for scientists, significantly enhancing both the speed and accuracy of their analyses.
