Course Information: Explainable Automated Machine Learning
Course Description
This graduate-level course offers a comprehensive introduction to the field of Explainable Automated Machine Learning (AutoML). The curriculum is designed to provide students with a deep understanding of the core concepts, methodologies, and applications that define this cutting-edge area of AI.
We will begin by examining the frameworks and optimization techniques that automate the process of generating high-performing machine learning pipelines. The course will then pivot to address the critical challenge of model opacity, focusing on state-of-the-art techniques for enhancing the transparency and explainability of AutoML systems. The ultimate objective is to equip students with the knowledge to build systems that are not only efficient and powerful but also trustworthy and capable of generating valuable, interpretable insights from complex optimization processes.
Prerequisites
A strong foundational knowledge of machine learning is essential for success in this course. Therefore, prior completion of the following course is strongly recommended:
- MTAT.03.227 Machine Learning
Schedule & Instructional Staff
- Course Commencement: September 1, 2025
- Lectures: Mondays, 12:15 - 14:00, Hybrid Format
- Practice Sessions: Wednesdays, 16:15 - 18:00, Hybrid Format
- Lecturer: Radwa El Shawi (radwa.elshawi@ut.ee)
- Teaching Assistants: Ahmed Soliman (soliman@ut.ee), Mohamed AbdElRahman (abdelrah@ut.ee)
Assessment and Grading
The final grade will be determined based on the following weighted components:
- Course Projects (40%): Two projects that students will work in teams to design, implement, and evaluate an AutoML solution within a specific domain. These project serves as the primary application of course concepts.
- Midterm Examination (20%): An assessment of foundational concepts covered in the first half of the course.
- Final Examination (20%): A comprehensive examination covering all course material.
- Lab Assignments (10%): Practical exercises designed to reinforce technical skills.
- Paper Presentation (10%): Students will analyze and present a significant research paper from the AutoML literature.
The final score (out of 100) will be mapped to a letter grade (A-F) according to the standard university scale. A deferred examination will be available for students with a valid, documented reason for missing the scheduled final exam.
Passing Requirements
To pass the course, a student must satisfy both of the following conditions:
- Achieve a cumulative total of 51 points or more.
- Obtain a minimum score of 50% on each individual assessment component (Labs, Exams, Presentation, and Project).
Course Policies
- Attendance: While attendance in lectures and practice sessions is not mandatory, it is highly encouraged to keep pace with the course material and participate in discussions.
- Academic Integrity: All submitted work must be the product of the individual student's effort. While high-level discussion of assignment concepts is permitted, the sharing of pseudocode, code, or specific solutions is a violation of academic policy.
- Lab Assignments: Labs are graded on a completion basis (2 points for a satisfactory submission, 0 otherwise).
- Extensions: Extensions for assignment deadlines will be granted only in exceptional circumstances (e.g., medical emergencies) and require formal documentation. Requests must be submitted to the instructor at least two days prior to the due date.
Technical Requirements
- Programming Language: Python (Version 3).
- Environment: Practical sessions will utilize Google Colaboratory (Colab). Students are required to have a Google account and should familiarize themselves with the Colab environment prior to the first session.