ODSL Forum: Sampling Techniques in High-Dimensional Parameter Spaces with ScannerBit 2.0 (Will Handley)

Friday 23 Feb 2024, 14:00 → 15:00 Europe/Berlin

Basement Seminar Room (Origins Building)

Abstract: 

Numerical methods for exploring high-dimensional parameter spaces are crucial across a wide variety of scientific fields, including:

 

 

• global fitting particle physics models (Frequentist likelihoods) 
• constraining cosmological parameters (Bayesian posteriors) 
• folding proteins (free energy landscapes) 
• phase diagrams in chemistry and lattice field theory (partition functions) 
• exploring the loss landscape of neural networks (machine learning).

 

 

This pedagogical talk will discuss the theoretical challenges of exploring high-dimensional parameter spaces, a practical framework for assessing the performance of different sampling algorithms, and scientific case-studies from the fields above. We will focus the discussion around the ScannerBit module of the GAMBIT global fitting framework, which includes a variety of sampling algorithms for both Frequentist and Bayesian analyses. This talk coincides with the announcement of ScannerBit 2.0 which incorporates a python interface to state-of-the-art sampling algorithms alongside the tried-and-tested Differential evolution (Diver) and Nested sampling methods (MultiNest & PolyChord).

 

 

• https://arxiv.org/abs/1705.07959 
• https://arxiv.org/abs/2101.04525 
• https://arxiv.org/abs/2205.15570