Careers Denmark
Se alle job

Research & Development

Research Assistant in Explainable and Causality Informed Reinforcement Learning (DEIS)

Aalborg University Aalborg, Denmark Publiceret 12. nov. 2026 Frist 26. nov. · 9 dage tilbage Arbejdsgiverens karrieresideFastansættelseFuldtidNyuddannetEngelsk

Resumé

Develop and evaluate explainable, causality-informed reinforcement learning methods for safe control of cyber-physical systems. The role includes work with UPPAAL Stratego, case studies in water and wastewater infrastructure, collaboration, and scientific dissemination.

Felterne er læst af AI fra opslaget. Har arbejdsgiveren selv udfyldt et felt, bruger vi det.

Kort om jobbet

Employment type: Permanent position · Full time. Requisition: 2026/1315. Country: Denmark.

Research Assistant in Explainable and Causality Informed Reinforcement Learning

We invite applications for a full-time, one-year position as Research Assistant in the Section of Distributed, Embedded and Intelligent Systems (DEIS), Department of Computer Science, Aalborg University. The position focuses on developing reinforcement learning methods that produce understandable control strategies and provide insight into the underlying system dynamics and causal relationships.

The research will address safe and effective control of cyber-physical systems, with particular emphasis on UPPAAL Stratego and collaboration with domain experts and external partners.

Expected starting date: 01.12.2026.

Your work tasks

You will develop and evaluate methods for explainable and causality informed reinforcement learning. The aim is to make learned strategies transparent to human domain experts, support well-founded trust in their decisions, and improve understanding of hidden dynamics and causal relationships in the systems being controlled.

Your tasks will include

  • Developing compact, interpretable representations of learned control strategies, including strategy minimisation with guaranteed preservation of safety properties.
  • Investigating how causal information can inform reinforcement learning and help explain the behaviour and decisions of learned strategies.
  • Working with reinforcement learning and recurrent neural networks in UPPAAL Stratego, including implementation and experimental evaluation.
  • Applying and evaluating the methods on safety-critical cyber-physical systems, with relevant case studies in water and wastewater infrastructure.
  • Collaborating with DEIS researchers, the UPPAAL development team and external partners, and contributing to scientific publications and dissemination.

Dette er et uddrag. Læs hele opslaget hos Arbejdsgiverens karriereside