- Introduction: What is causal inference?

- Review of Useful Probability concepts
- Random variable, predictors, divergences

- Introduction to Applications
- Computational neuroscience
- Financial markets
- Social networks

- Pearl Causality
- Causal Bayesian Networks (CBNs)
- Learning CBNS: Faithfulness and identifiably
- Algorithms

- Potential Outcome Model
- Counterfactuals and identification problems
- Graphical causal models

- Randomized Experiments
- Identification of causes in randomized experiments
- Effect modification

- Causality in Times Series
- Granger causality
- More general linear predictors

- Beyond linear models and Granger causality
- Directed information graphs
- Efficient algorithms

- Concrete Applications

- Computational neuroscience
- Financial markets
- Social networks

Causality, structure learning, network inference

This course attempts to be as self contained as possible, but it does approach the topic from a quantitative
point of view and, as such, students should be comfortable with the basics of (i.e. have taken at least one
course in) the following topics before enrolling:
- Statistics
- Probability Theory
- Linear Algebra
- Calculus (integral and differential)
- Programing in Pythor and Matlab

As course work will be largely computational, experience with at least one programming language is also
required.

Knowledge of probability and calculus as well as programming is a must.

• Identify situations in which a problem/data can be thought of as a network.
• Analyze data appropriately using a variety of network causal inference techniques
• Interpret the results of applying causal discovery or inference algorithms

• Continue to work through difficulties or initial failure to find optimal solutions.
• Demonstrate the capacity for critical thinking
• Use both general and domain specific IT resources and tools
• Access and evaluate appropriate sources of information.

In class with supporing problem solving sessions.

TA problem solving sessions, homework, exams, projects

Regular individual assignments: 30%
Midterm project: 30%
Final project: 40%

course notes

• https://go.epfl.ch/MGT-416https://go.epfl.ch/MGT-416