We are a joint research group between the University of Greenwich and the Rutherford Appleton Laboratory (RAL). We build the software infrastructure to hardwire UK Exascale computing directly to Synchrotron beamlines.
Our work targets the stochastic precursors to battery failure. We replace deterministic testing schedules with autonomous, AI-driven control loops that navigate the search space of material degradation in real-time.
- Autonomous Control: Reinforcement learning agents for active beamline steering (Diamond Light Source / ISIS).
- Theory: Benchmarking Autonomy Taxonomy (arXiv:2601.06978)
- Method: The ESME Framework (arXiv:2601.00851)
- Generative Simulation: Physics-informed diffusion models for surrogate modelling on GH200 hardware.
- High-Throughput Physics: Massively parallel transport solvers for terabyte-scale tomography.
Our code is open-source and developed for Tier-1 National Supercomputing facilities (e.g. Isambard-AI).
| Repository | Domain | Description |
|---|---|---|
| OpenImpala | ⚛️ Physics | Massively parallel solver for transport physics (AMReX/C++). (SoftwareX 2021) |
| PorousDiff | 🧠 GenAI | Conditional 3D diffusion for mechanical surrogate modelling. |
| LiionDB | 🔋 Data | The community standard database for Li-ion battery parameters (Zenodo). |
| OpenLSR-X | 👁️ Vision | SRGAN implementation for Synchrotron XCT super-resolution. |
| BatteryExplorer | 📊 Viz | 4D interactive dashboard for operando failure analysis. |
| MultimodalBenchmark | 💾 Data | Official code for the 3D Multimodal Synchrotron Dataset (Sci. Data 2025). |
PhD Studentship: Generative AI for Experimental Control We are recruiting a doctoral candidate to design the predictive engine for the "Self-Driving Microscope."
- Focus: Bayesian Optimization & Active Learning for X-ray Tomography.
- Site: Harwell Campus (RAL).
- Contact: James.LeHoux@greenwich.ac.uk
© 2026 BASE Laboratory | Part of the M34Impact Institute