SDST4612 Interpretable Machine Learning (6 credits) Academic Year 2025
Offering Department SCDS (Department of Statistics and Actuarial Science) Quota
Course Co-ordinator Prof Y Cao, SCDS (Department of Statistics and Actuarial Science) < yuancao@hku.hk >
Teachers Involved (Prof Y Cao,Statistics & Actuarial Science)
Course Objectives This course focuses on the interpretability of machine learning methods. In this course, students will first revisit classic statistical learning models with a focus on interpretability. They will then learn model-agnostic methods that help explain general machine learning models, and explore current progress in interpreting neural networks. Through a combination of lectures, case studies, and hands-on exercises, students will gain the necessary skills to apply these principles to real-world problems.
Course Contents & Topics Interpretability in machine learning and its importance, interpretability in linear regression, logistic regression, generalized linear models, decision trees, model-agnostic interpretation methods (partial dependence plot, individual conditional expectation, surrogate models, etc), example-based explanation methods, neural network interpretation.
Course Learning Outcomes
On successful completion of this course, students should be able to:

CLO 1 Understand the importance of interpretability in machine learning
CLO 2 Apply and interpret classic statistical learning models
CLO 3 Implement cutting-edge methods to interpret and explain blackbox models
CLO 4 Apply the principles of interpretable machine learning to real-world problems
Pre-requisites
(and Co-requisites and
Impermissible combinations)		 Pass in STAT3600 and STAT3612
Offer in 2025 - 2026 N            Examination No Exam     
Offer in 2026 - 2027 N
Course Grade A+ to F
Grade Descriptors
A Demonstrate thorough mastery at an advanced level of extensive knowledge and skills required for attaining all the course learning outcomes. Show strong analytical and critical abilities and logical thinking, with evidence of original thought, and ability to apply knowledge to a wide range of complex, familiar and unfamiliar situations. Apply highly effective organizational and presentational skills.
B Demonstrate substantial command of a broad range of knowledge and skills required for attaining at least most of the course learning outcomes. Show evidence of analytical and critical abilities and logical thinking, and ability to apply knowledge to familiar and some unfamiliar situations. Apply effective organizational and presentational skills.
C Demonstrate general but incomplete command of knowledge and skills required for attaining most of the course learning outcomes. Show evidence of some analytical and critical abilities and logical thinking, and ability to apply knowledge to most familiar situations. Apply moderately effective organizational and presentational skills.
D Demonstrate partial but limited command of knowledge and skills required for attaining some of the course learning outcomes. Show evidence of some coherent and logical thinking, but with limited analytical and critical abilities. Show limited ability to apply knowledge to solve problems. Apply limited or barely effective organizational and presentational skills.
Fail Demonstrate little or no evidence of command of knowledge and skills required for attaining the course learning outcomes. Lack of analytical and critical abilities, logical and coherent thinking. Show very little or no ability to apply knowledge to solve problems. Organization and presentational skills are minimally effective or ineffective.
Communication-intensive Course N
Course Type Lecture-based course
Course Teaching
& Learning Activities
Activities Details No. of Hours
Lectures 36.0
Tutorials 12.0
Reading / Self study 100.0
Assessment Methods
and Weighting
Methods Details Weighting in final
course grade (%)		 Assessment Methods
to CLO Mapping
Assignments Coursework (assignments, tutorials, and class test(s)) 50.0 1,2,3,4
Examination One 2-hour written examination 50.0 1,2,3,4
Required/recommended reading
and online materials		 Christoph Molnar, Interpretable machine learning, 2020. https://christophm.github.io/interpretable-ml-book/
Course Website http://moodle.hku.hk
Additional Course Information