Speaker: Will Handley (Kavli Institute for Cosmology, Uni of Cambridge)
Title: Nested sampling: powering next-generation inference and machine learning tools for cosmology, particle physics and beyond
Abstract: Nested sampling [1] is a radical alternative to traditional MCMC techniques for integrating and exploring probability distributions. With publicly available implementations such as MultiNest, PolyChord, dynesty and ultranest, it has become widely adopted across science as a powerful tool for computing integrals and scanning & sampling challenging a-priori unknown parameter spaces.
In this talk I will give a pedagogical introduction to the theory of nested sampling, and illustrate with recent novel applications in particle physics, cosmology, Bayesian Neural Networks and beyond [2-8]. I will finish with a discussion of recent innovations in the nested sampling toolkit, and prospects for the frontier of the field.
[1] https://arxiv.org/abs/2205.02030
[2] https://arxiv.org/abs/2211.10391
[3] https://arxiv.org/abs/2205.02030
[4] https://arxiv.org/abs/2205.11151
[5] https://arxiv.org/abs/2106.02056
[6] https://arxiv.org/abs/2105.13923
[7] https://arxiv.org/abs/1908.09139
[8] https://arxiv.org/abs/1902.04029