Theory and applications of causal inference and causal learning (Sommersemester 2020)

Lecturer: Prof. Dr. Christoph Lippert (Digital Health - Machine Learning) , Matthias Kirchler (Digital Health - Machine Learning) , Dr. rer. nat. Stefan Konigorski (Digital Health - Machine Learning)

General Information

• Weekly Hours: 2
• Credits: 3
• Graded: yes
• Enrolment Deadline: 06.04.2020 - 22.04.2020
• Teaching Form: Seminar
• Enrolment Type: Compulsory Elective Module
• Course Language: English
• Maximum number of participants: 15

Programs, Module Groups & Modules

Description

Please note: Course starts 04/21/20 via Zoom (Link below)
Recommendation: read chapter 1 of Reference 1 (see below) before the first class

Causality underlies everything that happens around and in us. In recent years, causality has also gotten much more attention in research, and different streams of research have been established in the fields of statistics, epidemiology, econometrics, and machine learning. In this class, we will go through part of different standard textbooks as well papers to build up the basic concepts and mathematical foundation of causal inference and causal learning. In addition, we will consider applications with a focus on health applications as well as the potential for applications, since these are still in development.

• Type: Advanced seminar with discussion of book chapters & papers on theory and applications of causal inference and causal learning
• Helpful introductory background reading: J Pearl (2018). The book of why
• Evaluation: participation in weekly discussions, presentation

Structure of the classes:

• Discussion of the chapter(s) that were assigned for the respective class. At the beginning of the class, everybody describes 3 things on the chapter: what he learned, things/concepts that stood out for him/her, and what he would like to go through in the class in more detail..
• It is expected that everybody has read all materials for each class and actively contributes to the discussions.
• There is one main responsible student for each session, who leads the session and discussions. In addition to leading the discussions, the student prepares some additional details for presentation in the class (e.g. go through one example/proof in detail, or provide code or results from an analysis). This can be chosen by the student or a topic can be given.

Recommendation: read chapter 1 of Reference 1 (see below) before the first class

Overview:

#	Date	Topic
1	21.04.2020	Introduction to Causal Inference, Presentation of topics and seminar structure
2	28.04.2020	Randomized Experiments
3	05.05.2020	Observational Studies
4	12.05.2020	Effect Modification & Interaction
5	19.05.2020	Causal Graphs & Inference in Graphical Models
6	26.05.2020	Bias due to Confounding & Selection Bias
7	02.06.2020	Measurement Bias & Random Variability
8	09.06.2020	Details on main concepts in causal inference
9	16.06.2020	Inverse probability weighting & standardization
10	23.06.2020	G-estimation & propensity scores
11	30.06.2020	Instrumental variables & Mendelian Randomization
12	07.07.2020	(Emulated) target trials
13	14.07.2020	Variable selection for Causal inference,  Learning cause effect models I
14	21.07.2020	Connections to Machine Learning

Requirements

open for Master- and PhD students in Digital Health & Data Engineering.

Prerequisite: statistical/machine learning knowledge at the level of the ‘Deep Learning’ class by Christoph Lippert or the ‘Data Analysis using R, 1 – Statistical Epidemiology’ by Stefan Konigorski

Recommendation: read chapter 1 of Reference 1 (see below) before the first class

Literature

 (Main reference for the class is book 1):

1. Hernán MA, Robins JM (2020). Causal Inference: What If. Boca Raton: Chapman & Hall/CRC. Available from:  https://www.hsph.harvard.edu/miguel-hernan/causal-inference-book/
2. Peters J, Janzing D, Schölkopf B (2017). Elements of Causal Inference - Foundations and Learning Algorithms. MIT Press. Available from:  https://mitpress.mit.edu/books/elements-causal-inference
3. Bishop, C. M. (2006). Pattern recognition and machine learning. Springer.

Learning

The course will take place via Zoom

Zoom link: https://zoom.us/j/95749197426 (meeting ID 957 4919 7426)

For password, please contact  stefan.konigorski(at)hpi.de 

Examination

Grade:

• Active contribution to discussions in class: 30%
• Presentation in one class: 70%

Dates

• Time: Tuesdays 11.00-12.30
• Place: G1 E15/16/ via zoom

• First class: 21.4.2020, last class 21.07.2020

Zurück