Professor of Statistics, University of Jyväskylä. Computational statistics, applied probability, Monte Carlo methods, Bayesian inference.
https://iki.fi/mvihola/
Matti Vihola
Loading...
Nordstat will be in Helsinki June 1-4: www.helsinki.fi/en/conferenc.... Abstract submission DL is in one week (March 30).
Some new work on (dynamic path length) Riemannian manifold HMC. RMHMC with a block-diagonal mass matrix admits an explicit (symmetric & reversible) integrator. We find the parameters of the sampler, including the hierarchical mass matrix using on-line adaptation.
Our paper (with Alain Durmus, Samuel Gruffaz, Miika Kailas and Eero Saksman) on the ergodicity of dynamic path length HMC algorithms appeared in the AAP: doi.org/10.1214/25-A....
The results accommodate the 'biased progressive sampling' path selection used in @mc-stan.org.
Thinking of the next big move at the turn of the year? Here are some in Machine Learning and AI research, for PIs and postdocs:
@ellisinstitute.fi is looking for:
- PIs who will also get a tenure track professorship: DL Jan 12 www.ellisinstitute.fi/PI-recruit-2...
A related recent pre-print by Samuel and co-authors has theoretical backing for the empirical observation that BPS can be superior to multinomial selection rule: arxiv.org/abs/2603.18640.
We are looking for an Assistant or Associate Professor to join the Department of Mathematics and Statistics @uniofjyvaskyla.bsky.social: ats.talentadore.com/apply/assist....
Our invitation to adaptive MCMC theory appeared in the EJP: doi.org/10.1214/26-E.... The 'waning adaptation' condition accommodates both stochastic gradient style continuous adaptation and schemes which update parameters increasingly rarely.
Yvann Le Fay, Nicolas Chopin, Matti Vihola: On the complexity of standard and waste-free SMC samplers https://arxiv.org/abs/2604.03352 https://arxiv.org/pdf/2604.03352 https://arxiv.org/html/2604.03352
We're #hiring! We have three new Assistant Professor positions in Statistics @ucddublin.bsky.social See www.ucd.ie/workatucd/jo... and search for job references:
- Permanent (ref. 019430).
- Temporary 5-year (ref. 019431).
- Temporary 5-year in the area of Machine Learning (ref. 019429).
We are looking for Ecological Statistician to work on long-term biodiversity and ecosystem functioning data
jobs.helsinki.fi/job/Helsinki...
Adaptive Markov chain Monte Carlo (MCMC) algorithms, which automatically tune their parameters based on past samples, have proved extremely useful in practice. The self-tuning mechanism makes them ‘non-Markovian’, which means that their validity cannot be ensured by standard Markov chains theory. Several different techniques have been suggested to analyse their theoretical properties, many of which are technically involved. The technical nature of the theory may make the methods unnecessarily unappealing. We discuss one technique—based on a martingale decomposition—with uniformly ergodic Markov transitions. We provide an accessible and self-contained treatment in this setting, and give detailed proofs of the results discussed in the paper, which only require basic understanding of martingale theory and general state space Markov chain concepts. We illustrate how our conditions can accommodate different types of adaptation schemes, and can give useful insight to the requirements which ensure their validity.
The No-U-Turn Sampler (NUTS) is the computational workhorse of modern Bayesian software libraries, yet its qualitative and quantitative convergence guarantees were established only recently. A signifi...
arxiv.org
The Department of Mathematics and Statistics is seeking to recruit an Assistant/Associate Professor (Tenure Track) in Mathematics starting on August 1, 2026, or as soon as possible after that. Accordi...
