MCMSki occupies a unique niche in the Bayesian conference landscape: a focused meeting on computational methods held in the relaxed and scenic setting of a mountain ski resort. The combination of serious science and informal atmosphere has made MCMSki one of the most popular and productive gatherings in the Bayesian computation community.
Origins and Spirit
MCMSki was conceived as a conference that would bring together researchers working on Monte Carlo methods and Bayesian computation in an environment that encourages open discussion, collaboration, and creative thinking. The name—a portmanteau of "MCMC" and "ski"—captures the meeting's dual character: rigorous computational research by day, skiing and socializing by evening.
The first MCMSki is held in Bormio, Italy, establishing the format of a Bayesian computation conference at a ski resort.
MCMSki II is held in Bormio, with growing attendance and an expanded scientific program.
MCMSki IV is held in Chamonix, France, attracting over 300 participants and featuring sessions on topics from Hamiltonian Monte Carlo to scalable inference.
MCMSki V is held in Lenzerheide, Switzerland, with continued growth and increasing international participation.
The choice of a ski resort is not merely whimsical. The MCMSki organizers recognized that the best scientific discussions often happen informally—over dinner, on a chairlift, or during a hike. By situating the conference in an environment that naturally encourages relaxation and socialization, MCMSki creates the conditions for the kind of open, exploratory conversations that can lead to new research ideas and collaborations.
Scientific Program
Despite its informal setting, MCMSki features a rigorous and high-quality scientific program. The conference includes invited talks by leading researchers in Bayesian computation, contributed talks and poster sessions, and tutorials on emerging computational methods. Topics at recent meetings have included:
Hamiltonian Monte Carlo and its variants, variational inference and amortized inference, sequential Monte Carlo and particle methods, approximate Bayesian computation (ABC), scalable MCMC for big data, probabilistic programming languages, and the interface between Bayesian computation and machine learning.
Community Impact
MCMSki has become an incubator for new ideas in Bayesian computation. Many important algorithmic developments have been first presented or discussed at MCMSki, and the conference has fostered collaborations that have led to significant advances in the field. The meeting's relatively small size (typically 200-400 participants) ensures that attendees have ample opportunity to interact with speakers and fellow researchers.
"MCMSki is where the Bayesian computation community comes to think big, share crazy ideas, and push the boundaries of what is computationally possible. The ski resort setting is just a bonus."— Christian Robert