Pushing the frontiers of Reinforcement Learning and Optimization — from theory to real-world deployment.
SARL-PLUS — Smart Analytics & Reinforcement Learning — is a research group at the IDA Lab, part of the Faculty of Digital & Analytical Sciences (DAS) at Paris Lodron Universität Salzburg (PLUS).
We are dedicated to advancing research, education, and innovation at the intersection of:
| Area | Focus | |
|---|---|---|
| 🤖 | Reinforcement Learning | Safe RL, Model-Based RL, Meta-RL, Multi-Agent Systems |
| ⚡ | Physics-Informed AI | Hamiltonian Neural Networks, Symplectic Integrators, Structure-Preserving Optimization |
| 🏭 | Real-World Control | Particle Accelerators, Industrial Automation, Scientific Facilities |
| 📐 | Optimization | Bayesian Optimization, Koopman Operator Theory, Gaussian Process Methods |
Our mission is to foster collaboration among researchers, practitioners, and students to explore cutting-edge developments and translate them into high-impact applications.
┌─────────────────────────────────────────────────────────────────┐
│ SARL Research Stack │
├──────────────────────────┬──────────────────────────────────────┤
│ 🧬 Theory │ Lie Algebra · Symplectic Geometry │
│ │ Koopman Operators · Regret Bounds │
├──────────────────────────┼──────────────────────────────────────┤
│ 🤖 Algorithms │ Model-Based RL · Meta-RL (MAML) │
│ │ Bayesian Optimization · Safe RL │
├──────────────────────────┼──────────────────────────────────────┤
│ ⚙️ Applications │ Particle Accelerators (CERN/DESY) │
│ │ Industrial Process Control │
└──────────────────────────┴──────────────────────────────────────┘
Simon Hirlaender |
We actively collaborate with leading international research institutions and facilities:
| 🏛️ Institution | 🔗 Focus Area |
|---|---|
| CERN — Geneva, Switzerland | Particle beam control & RL applications |
| DESY — Hamburg, Germany | Accelerator physics & optimization |
| FAIR/GSI — Darmstadt, Germany | Heavy-ion physics & control systems |
| MedAustron — Wiener Neustadt, Austria | Medical accelerator control |
Team Lead: Simon HirlaenderEmail: simon.hirlaender@plus.ac.atsarl.ida.plus@gmail.comInstitution: IDA Lab — Fachbereich Artificial Intelligence & Human Interfaces (AIHI)Fakultät Digital & Analytical Sciences (DAS)Paris Lodron Universität Salzburg (PLUS)Address: Jakob-Haringer-Straße 6Techno 6 | 5020 Salzburg | AustriaWeb: https://www.plus.ac.at/aihi/der-fachbereich/ida-lab/teams/sarl/🔗 See our full publication list on the SARL team page
Recent highlights:
- 📄 Causal GP-MPC — Causal Gaussian Process MPC for the AWAKE experiment at CERN
- 📄 Koopman World Models — Koopman-stabilised offline reinforcement learning
- 📄 Safe RL via Symplectic Shields — Physics-constrained differentiable optimization
- 📄 METRIC — Model-based Epistemic Training for Reinforcement Learning in Control (FWF Project)
We welcome motivated PhD students, Master's students, and collaborators passionate about reinforcement learning, optimization, and AI for science.
📬 Reach out at sarl.ida.plus@gmail.com or visit our team page.
© SARL-PLUS · IDA Lab · Paris Lodron Universität Salzburg