Pietro Vischia: Third Intensive Course on Statistics for HEP (academic year 2021/2022)
E/3rd floor-E.349 - Seminar room (E.349)
Marc de Hemptinne (chemin du Cyclotron, 2, Louvain-la-Neuve)
This course aims at providing an introduction to the most commonly used statistical methods for high energy physics.
Morning lectures will be supplemented by afternoon exercise sessions, where the material of the morning will be put in practice. There will be five lecture+exercise events, held on Fridays.
There will be home assignments to help you better digest the content shown in class, but they are not mandatory (i.e. the attendance certificate won't depend on them)
This course provides 3 credits for the UCLouvain doctoral school (CDD Sciences) and also contributes to the activities of the Excellence of Science (EOS) Be.h network
The course will be recorded, and the link to the recordings will be shared with the participants.
The course will take place both in person and online. (the zoom link will be sent to the registered participants by email).
Although the vaccination rate in Belgium is rather high, there's always the possibility that travel restrictions (or restrictions on meetings in person) may be imposed again by the government: in that case, it may be necessary to have the lectures online only. Be careful with your travel arrangements.
Tentative program (subject to changes):
Lesson 1 - Fundaments: Bayesian and frequentist probability, theory of measure, correlation and causality, distributions
Lesson 2 - Point and Interval estimation: maximum likelihood methods, confidence intervals, most probable values, credible intervals
Lesson 3 - Test of hypotheses: frequentist and Bayesian tests, CLs, significance, look-elsewhere effect, reproducibility crysis
Lesson 4 - Commonly-used methods in particle physics: unfolding, ABCD, ABC, MCMC
Lesson 5 - Machine Learning: overview and mathematical foundations, generalities most used algorithms, automatic Differentiation and Deep Learning
Computational requirements:
for the frontal lectures you just need to be able to view the video stream and/or the PDF of the lectures. For the exercise sessions, you are expected to be able to run python2/3 jupyter notebooks on your laptop/desktop, on your computing cluster, or on services like Google COLAB. I'll advertise a full list of required packages (typically as conda yml environment) a few days before the beginning of the course.
I am also trying (no guarantee) to see if I can produce a docker image with all the required software preinstalled.
Organization:
Pietro Vischia (teacher and chair)
Carinne Mertens (secretariat)
