ILLUMINE

ILLUMINE

The ILLUMINE project develops an autonomous AI-driven framework for optimizing scientific workflows, integrating real-time data analysis, adaptive experiment design, and human-AI collaboration to accelerate scientific discovery.  The goal is to close the loop between fast analysis, machine-assisted decision-making and data acquisition.

Key Objectives

  • Develop real-time ML inference and edge-to-HPC analysis pipelines for accelerated processing of high-velocity experimental data (200 GB/s–1 TB/s), integrating edge-based compression algorithms and distributed computing architectures to enable autonomous event selection, noise rejection, and adaptive instrument control.
  • Implement reinforcement learning-driven decision support systems and high-dimensional phase-space search algorithms to autonomously optimize experimental parameters, dynamically balance exploration/exploitation tradeoffs, and maximize data quality under uncertainty constraints.
  • Create a modular Bluesky-based interoperability framework with standardized APIs for cross-facility autonomous workflows, enabling composable components for beamline optimization, AI/ML integration, and real-time experiment steering while maintaining facility-specific customization capabilities.

Explore the ILLUMINE project page

Partners & Collaborators

Jana Thayer (PI, SLAC)

Cong Wang (SLAC)

Frederic P Poitevin (SLAC)

Ryan Herbst (SLAC)

Vivek Thampy (SLAC)

Stuart Campbell (BNL)

Daniel Allan (BNL)

Andi Barbour (BNL)

Thomas Caswell (BNL)

Phillip Maffettone (BNL)

Daniel Olds (BNL)

Mak Rakitin (BNL)

Nathan Urban (BNL)

Nicholas Schwarz (ANL)

Franck Cappello (ANL)

Ian Foster (ANL)

Antonino Miceli (ANL)

Alexander Hexemer (LBNL)

Dylan McReynolds (LBNL)

Jonathan Taylor (ORNL)

Publications & Talks

“Towards Closing the Autonomous Loop at Multiple Facilities: Developing Web-based User Interfaces and Data Infrastructure for Autonomous Experiments and Machine Learning Workflows.” Contributed talk at the XIX International Small Angle Scattering Conference (SAS 2024), Taipei, Taiwan

P. Maffettone, C. Fernando, T. A. Caswell, D. Garilov, P. Shafer, D. Allan, S. I. Campbell, Proceedings of the Workshop on Accelerating Discovery in Natural Science Laboratories with AI and Robotics at the International Conference on Robotics and Automation (2024).