master international Physics of Complex Systems

Computational Sciences, Florent Krzakala

Where: Tour 12-13 salle 523, 40 Rue Jussieu, 75005 Paris, France When: Friday Morning, 8H39-10H30 You can find on overleaf the lecture notes written by the previous student for my course in the master Physics of Complex Systems. Feel free to edit them and to correct mistakes!!

The first lecture is an introduction to probability theory, classical bounds and large deviation theory.

The first Homework is posted: see the file on overleaf. It is due for october 3 (no delay accepted).
Very basic examples of python notebook relevant to Homework 1 are given on my github page, see for instance: for a pooling problem or for a sampling one

A primer on inference, statistics, and maximum likelihood

The second Homework is posted: see the file on overleaf. It is due for october 25 (no delay accepted).

The third lecture starts with an introduction to sampling simple distributions, and pursue towards MCMC, Detailed and Global balance algorithms, and cluster algortihms.

The third Homework is posted: see the file on overleaf. It is due for November 15 (no delay accepted).

. For those who want to work on Ising spins instead of Hardsphere, here is an alternative exercice.

The fourth lecture deels with unsupervised learning, and discusses PCA, Young-Eckart theorem, Johnson Lindenstrauss lemme and the k-means algorithm.

The final Homework is posted: see the file on overleaf. It is due for the end of December.

.

Python notebooks on your computer are great. But another possibility, and q good way to use powerful computer without buying one, is to use google colab, the Colaboratory platform from Google: It requires no specific hardware or software, and even allows you to use GPU computing for free, all by writting a jupyter notebook that you can then share.

A good book for probability and statistics, accessible to students, is Larry A. Wasserman 's All of Statistics

Monte-Carlo methods are well covered in Werner Krauth's Statistical Mechanics: Algorithms and Computations. His MOOC on coursera is also recommended.

A good introduction to statistical learning is given in Elements of Statistical Learning by Jerome H. Friedman, Robert Tibshirani, and Trevor Hastie.

Another great reference is Machine Learning:A Probabilistic Perspective, by Kevin P. Murphy.

Exam from 2018: exam

Exam from 2017: exam