Choose timezone
Your profile timezone:
Description
This tutorial presents a detailed and operational introduction to Bayesian inference using Bilby. The focus is on understanding how posterior distributions are constructed from likelihoods and priors, and on developing full control over the inference process. Rather than relying on predefined workflows, participants are guided through the structure of likelihood functions and the practical consequences of modeling assumptions. A central part of the tutorial is the exploration of sampling methods, with particular attention to nested sampling algorithms such as Dynesty and PyMultiNest. The tutorial includes hands-on sessions in which participants build and modify likelihoods, run inference with different samplers, and interpret the resulting outputs. The aim is to provide a clear and practical understanding of Bayesian inference, and to equip participants with the tools needed to design, diagnose, and validate their own analyses in a wide range of applications.
Introductive material can be found in the Previous tutorial
Schedule:
- Recap and orientation (15')
- Dynesty internals and diagnostics (45')
- Break (15')
- Sampler comparison and selection (35')
- Reparametrization and prior design (35')
- PP tests and injection-recovery validation (20')
- Hierarchical inference with bilby.hyper (40')
- Wrap-up and further directions (20')
Material available:
- Introductive material from previous tutorial
- Tutorial guide
- Jupyter noteboocs
Note: times are indicative
