Dr. Rina Dechter @ UCI

Dr. Rina Dechter - University of California at Irvine
home \| publications \| book \| courses \| research	Revised on Mar. 09, 2022


CS 295 - Winter 2021-22, Causal Reasoning

\| main \|

Instructor: Rina Dechter
Days,Time: M/W, 11:00 am - 12:20 pm (PT)
Classoom: DBH 1422
Zoom: https://uci.zoom.us/j/94278071276
Office hours: Upon Request

Course Description

This course will cover topics in Causal Inference. The course will run as a seminar where I will give lectures in the first half of the course and then students will be required to read and present papers based on chapters in books to the class for the second half. There will be a course project which can be based on these selected papers and also several homework assignments. This course is intended for PhD students in the area of AI and Machine Learning, with 271 and 273 as prerequisite courses. Additionally, students who took CS 276 will have particularly good preparation for this course. If you are a second-year master student that already took 271 and 273, please talk to me to obtain approval.

Course Topics

• Introduction: Causal Hierarchy
• The Simpson Paradox
• Structural Causal Models
• Identification of Causal Effects
• The Problem of Confounding and the Back-Door Criterion
• Causal Calculus
• Linear Structural Causal Models
• Counterfactuals
• Structural Learning

Textbooks

[P] Judea Pearl, Madelyn Glymour, Nicholas P. Jewell,
     Causal Inference in Statistics: A Primer,
     Cambridge Pess, 2016.
[C] Judea Pearl,
     Causality: Models, Reasoning, and Inference,
     Cambridge Press, 2009.
[W] Judea Pearl, Dana Mackenzie,
     The Book of Why,
     Basic books, 2018.
[PCH] E. Bareinboim, J. Correa, D. Ibeling, T. Icard,
     On Pearl's Hierarchy and the Foundations of Causal Inference,
     Columbia University, 2020. (To appear in: Probabilistic and Causal Inference: The Works of Judea Pearl, ACM Turing Series).
[Darwiche] Adnan Darwiche,
     Modeling and Reasoning with Bayesian Networks,
     Cambridge Press, 2009.

Course Project: TBD

(You will need to be logged into Google with you UCI account to access.)
Project Information
Project Sign-Up Sheet

Syllabus:

Week	Topic	Lectures	Slides	Homework	Reading	Date

Week 1	Introduction: Pearl's Causal Hierarchy Simpson Paradox, Causal Hierarchy, Defining Structural Causal Models	Lec 1 Lec 2 ISAIM Interview of Judea Pearl	Slides 1		[W] Ch. 1 [P] Ch. 1 [PCH] Sec 1.1	M 01/03 virtual (class zoom) W 01/05 11:00-11:30AM (PT) virtual (class zoom) 11:30AM (PT) Zoom: Live ISAIM Interview of Judea Pearl

Week 2	Structural Causal Models, d-Separation Identification of Causal Effects	Lec 3 Lec 4	Slides 2	HW 1	[P] Ch. 2 [PCH] Sec 1.2 [Darwiche]: Ch. 4	M 01/10 virtual (class zoom) W 01/12 virtual (class zoom)

Week 3	Identification of Causal Effects The Back-Door Criterion	Lec 5	Slides 3		[P] Ch. 3 [C] Sec 1.3, 3.1-3.3	M 01/17 [holiday] W 01/19 virtual (class zoom)

Week 4	The Front-Door Criterion The DO Calculus	Lec 6 Lec 7	Slides 4	HW 2	[C] Ch. 3.4-3.5 [P] Ch. 3 Biometrika 1995	M 01/24 virtual (class zoom) W 01/26 virtual (class zoom)

Week 5	Linear Structural Causal Models	Lec 8 Lec 9	Slides 5 Slides 5b		[P] Ch. 3.8 [C] Ch. 5	M 01/31 in-person W 02/02 in-person

Week 6	Counterfactuals	Lec 10 Lec 11	Slides 6	HW 3	[P] Ch. 4 [C] Ch. 7	M 02/07 in-person W 02/09 in-person

Week 7	Algorithmic Approach for Identification P1: Causal Discovery	Lec 12 Lec P1 (Ofek Gila)	Slides 7 Slides P1(a) Slides P1(b)		= Theory of Inferred Causation Causation, Prediction, and Search (Ch. 5)	M 02/14 in-person W 02/16 in-person

Week 8	P2: Learning Causal Effects via Weighted Empirical Risk Minimization	Lec P2	Slides P2		Learning Causal Effects via Weighted Empirical Risk Minimization	M 02/21 [holiday] W 02/23 in-person

Week 9	P3: Probabilistic Evaluation of Counterfactual Queries P4: Estimating Causal Effects Using Weighting-Based Estimators P5: Equivalence and Synthesis of Causal Models P6: Probabilities of Causation - Three Counterfactual Interpretations and Their Identification	Lec P5 Lec P6	Slides P3 Slides P4 Slides P5 Slides P6		Probabilistic Evaluation of Counterfactual Queries Estimating Causal Effects Using Weighting-Based Estimators Equivalence and Synthesis of Causal Models Probabilities of Causation - Three Counterfactual Interpretations and Their Identification	M 02/28 in-person W 03/02 in-person

Week 10	P7: Estimating Identifiable Causal Effects through Double Machine Learning P8: Causal Inference Using Tractable Circuits	Lec P7 Lec P8	Slides P7		Estimating Identifiable Causal Effects through Double Machine Learning Causal Inference Using Tractable Circuits	M 03/07 in-person W 03/09 in-person

Finals Week	TBD					F 03/18

Background Reading

[DF] Elias Bareinboim and Judea Pearl,
     Causal inference and the data-fusion problem,
     PNAS, 2016.
Jin Tian,
     Studies in Causal Reasoning and learning,
     University of California, Los Angeles, 2002.
Jin Tian and Judea Pearl,
     A General Identification Condition for Causal Effects,
     AAAI, 2002.
[PCH] E. Bareinboim, J. Correa, D. Ibeling, T. Icard,
     On the completeness of an identifiability algorithm for semi-Markovian models,
     Annals of Mathematics and Artificial Intelligence, 2008.
Karthika Mohan and Judea Pearl,
     Graphical Models for Processing Missing Data,
     Journal of the American Statistical Association, 2021.
[BMK] Judea Pearl,
     Causal diagrams for empirical research,
     Biometrika, 1995.
Daniel Kumor, Carlos Cinelli, Elias Bareinboim,
     Efficient Identification in Linear Structural Causal Models with Auxiliary Cutsets,
     NeurIPS, 2019.
Judea Pearl,
     Linear Models: A Useful “Microscope” for Causal Analysis,
     Journal of Causal Inference, 2013.