This page is under construction. The programme is being finalised with the speakers. Their names, the course details and the detailed timetable will be added here as they are confirmed, and the full timetable will be published in January 2027.
| 4 topics |
16 sessions of 90 minutes |
24 hours of teaching |
2 poster evenings |
The shape of the week
The school runs from Monday 1 March at noon until Friday 5 March in the early afternoon. It combines several kinds of activity, spread across the week so that every day mixes listening, presenting and discussing:
| Four topics Each is made up of four 90-minute sessions. |
Poster sessions Participants present their research to one another and to the speakers. |
| Career round tables and workshops to be confirmed With researchers and professionals from academia and industry. Details will be announced here. |
Hackathon to be confirmed If material conditions permit. |
The working language is English throughout.
The four topics
Uncertainty modelling
Every model is built on imperfect data and simplifying assumptions. This course is about representing what a model does not know: where uncertainty comes from, how to describe it, how to propagate it through a computation, and how to report it in a way that a decision-maker can use.
Simulation-based inference
Many scientific and industrial fields rely on simulators, from physics and climate to energy systems, that can generate data but whose likelihood cannot be written down. Simulation-based inference provides the tools to learn the parameters of such models from observations, with faithful uncertainty on the result.
Bayesian networks
Bayesian networks represent how variables depend on each other, and let us reason under uncertainty with those dependencies made explicit. The course covers how such models are built, learned from data and queried, as well as where they are used in practice.
Machine learning and HPC resource management
Training and running models consume computing time, energy and hardware resources. This course looks at the resources behind machine learning: how high-performance computing is used in practice, the computational and environmental costs of AI, and how they can be reduced.
Poster sessions
All participants are invited to present a poster on their own research. Poster sessions will take place during two evenings of the week, each opening with short pitches (one minute per presenter). Practical details (format, printing) will be sent to accepted participants.
Careers and exchange
The school is part of a Marie Skłodowska-Curie doctoral programme, and career development is an integral part of MSCA programmes. Round tables and workshops, currently under preparation, will give participants the opportunity to discuss career paths in academia and industry with senior researchers and professionals from the programme's partner organisations. Further details will be announced on this page.
Social programme
Some of the best conversations at a school happen away from the lecture hall. We will start the week together with a welcome reception on Monday evening, and gather again on Thursday evening for a social dinner, offered to all participants. They are two relaxed occasions to get to know the other participants and the lecturers, to carry on the day's discussions, and perhaps to start a collaboration or two.
Attendance and certificate
The school is designed as a full week, and participants are expected to attend the entire week. A certificate of attendance stating the number of teaching hours will be issued at the end of the school. You can submit it to your doctoral school for training credits; recognition is decided by each doctoral school.