Speaker
Fazhi Qi
(Institute of High Energy Physics)
Description
IHEP operates a series of large-scale facilities — BESIII, LHAASO, JUNO, HEPS, CSNS, with major roles at the LHC — generating petabyte-scale scientific data annually. This talk surveys the institute's integrated response across data infrastructure, computing platforms, scientific software, and AI for Science.
A federated data infrastructure anchored by the National High Energy Physics Scientific Data Center integrates HTC/HPC resources, distributed storage, and a FAIR-guided data management system covering the full data lifecycle. A next-generation computing architecture is being developed to unify heterogeneous resources across facilities, while scientific software efforts emphasize sustainability and international alignment.
On AI for Science, IHEP is deploying GPU-accelerated platforms, foundation models, and AI agent frameworks for real-time event reconstruction, fast simulation, autonomous beamline optimization, and AI-assisted discovery — supported by an institute-wide collaborative ecosystem bridging domain science and AI research. Current status and future roadmap will be presented.
Author
Fazhi Qi
(Institute of High Energy Physics)
Co-author
Hao Hu
(Institute of High Energy Physics, CAS)