Lectures of the third week
Monday, June 24
Bayesian methods and machine learning (Christian Forssén)
- Bayesian optimization
- Global versus local optimization
- Gaussian process statistical model
- Acquisition functions
- Scanned lecture notes [pdf]
- Bayesian Optimization lecture [ipynb]
- Bayesian optimization in ab initio nuclear physics by A. Ekström, C. Forssén et al.
Model checking (Daniel Phillips)
- Empirical coverage probability, QQ plots, systematic trends in residuals
- Case study: EFT truncation errors
- Model checking, first lecture [pdf]
Tuesday, June 25
Bayesian parameter estimation (Dick Furnstahl)
- Sloppy models
- SVD and PCA
- Lecture material and afternoon exercises [ipynb]
- A Singularly Valuable Decomposition: The SVD of a Matrix (Kalman) [pdf]
Model checking (Daniel Phillips)
- Case study: EFT truncation errors (contd.)
- Gaussian Process model diagnostics
- Model checking, second lecture [pdf]
Wednesday, June 26
Bayesian neutral networks (Christian Forssén)
Why Bayes is Better (4) (Daniel Phillips)
- Experimental design
- Lecture on experimental design [pdf]
- Notebook reprising signal plus background example [ipynb]