Simulation-based inference for rapid Bayesian parameter estimation in epidemiological models: a comparison with MCMC
2026
Read Simulation-based inference for rapid Bayesian parameter estimation in epidemiological models: a comparison with MCMC (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2606-27286.txt.
Description
Bazarova and colleagues compare simulation-based inference with the classic sampling workhorse for epidemic model fitting. Mechanistic disease models guide forecasting and public health decisions, but Bayesian fitting by Markov chain Monte Carlo can crawl. Their SECIR model tracks susceptible, exposed, infectious, and recovered groups through German intensive care data from 2020. Neural posterior estimation, that route, learns posteriors from simulated runs instead. During 2026, short windows and a long multi-wave rebuild are both tested.
Across month-long windows, the neural route recovers posteriors in strong agreement with the sampler and reproduces intensive care trajectories well. Runtimes collapse from roughly a thousand seconds to about a minute on one graphics processor in those tests. The long reconstruction preserves the main posterior shape despite wider uncertainty, again far faster. Near-real-time, repeatable fitting becomes realistic for outbreak response.
Catalog record
| YEAR | 2026 |
|---|---|
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 86 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2606.27286