Workshop on Statistical Issues arising in data analyses of modern flavour experiments such as Belle, LHCb and NA62. The focus is on statistical and systematical uncertainties of measurements ranging from highest to lowest precision, and the optimal modelling of signals and backgrounds. Together with invited statisticians the most urgent problems and (new) solution approaches are discussed.
Tentative Agenda Outline:
- 1. Day Morning
- Introductory talks on
- statistics - recent developments
- statistical issues in flavour experiments
- Introductory talks on
- 1. Day Afternoon:
- Likelihoods and confidence intervals estimation
- problems with degenerate likelihoods and physical boundaries
- CLs versus Feldman-Cousins vs Bayesian etc.
- weighted events
- evaluation of likelihoods
- publishing likelihoods
- combining measurements
- Likelihoods and confidence intervals estimation
- 2. Day Morning:
- Systematic uncertainty treatment
- profiling vs marginalising
- priors
- template morphing and smoothing
- modelling of the resolution function
- background modelling uncertainties
- theory uncertainties
- fits to flavour data and treatment of data uncertainties in them
- Systematic uncertainty treatment
- 2. Day Afternoon:
- Session:
- contributed talks
- statisticians talks
- Session:
- 3. Day Morning:
- Session: Applications in Machine learning
- model-independent anomaly detection
- Highlight talks
- Session: Applications in Machine learning
- 3. Day Early Afternoon:
- physicist's summary talk
- statistician's summary talk
