Complex Systems and Introduction to Machine Learning

Contact: Alain Barrat or Christophe Eloy


I-Introduction to complex systems and complex networks.

  • What are complex systems (examples in various fields), how can they be represented, use of network representations.
  • Elements of statistical characterization of networks (paths, degrees, centrality, community detections, degree distribution). Small-world, scale-free properties.
  • Models of networks (random graphs, preferential attachment, copy model).
  • Robustness and resilience of networks.
  • Diffusion and spreading processes: from random walks to epidemiology.
  • Models of social phenomena on networks.
  • Introduction to computational social science, online social networks.

II-Learning from data: machine learning and deep learning

  • Introduction to machine learning (learning problem, types of learning, overview...)
  • Linear models (linear regression with one or multiple variables, logistic regression)
  • Non-linear models (Support Vector Machines, naive Bayes, decision tree)
  • Clustering (k-means, Gaussian mixture models)
  • Neural networks (non-linearities, convolutional NN networks, recurrent NN...)
  • Reinforcement learning (Markov Decision Processes, Dynamic Programming, Model-Free Prediction, Model-Free Control)
  • Deep reinforcement learning (Policy gradient, exploration/exploitation, current models…)