Master in Artificial Intelligence: Modules M05, M06 and M08

Three modules taught between 2019 and 2023 to master students of the Master in AI run by the Idiap Research Institute with Unidistance: open science and ethics (M05), and machine learning fundamentals (M06 and M08).

The Master in Artificial Intelligence is a distance-learning programme run by the Idiap Research Institute together with Unidistance Suisse. I taught three of its modules between 2019 and 2023: Open Science and Ethics (M05), and Fundamentals of Machine Learning, spread over two semesters (M06 and M08).

M05: Open Science and Ethics

An introductory course on ethics and reproducibility in artificial intelligence, given in the spring semesters. It is composed of two parts: the first covers ethical aspects of AI, the second practical aspects of building AI systems so they are continuously reproducible and extensible.

Program

• Credits: 2 ECTS (50 to 60 working hours)
• Grading: written exam on ethics and the law (30%) and an open-science mini-project (70%)
• Days: Mondays, from 9:00 to 10:30
• Syllabus: AI and the law (week 1), AI and data protection (week 2), AI and ethics with the written exam for part 1 (week 3), reproducibility, what is it? (week 4), data and workflow management (week 5), version control with Git (week 6), code sharing with GitHub (week 7), unit testing and continuous integration (week 8), documentation and reporting (week 9), packaging, deployment and licensing (week 10), mini-project presentations (week 11)

Each topic is covered in a single week. There are no graded assignments other than the mini-project.

M06 and M08: Fundamentals of Machine Learning

These two modules present fundamental tools used in machine learning and pattern recognition, ranging from the most basic to more advanced (logistic regression, principal component analysis, linear discriminant analysis, multi-layer perceptrons, Gaussian mixture models, and support vector machines). They serve as a prerequisite for deep learning and other master specialisations. M06 was given in the autumn semesters and M08 in the spring semesters.

Program

• Credits: 4 ECTS per module (100 to 120 working hours each; 8 in total)
• Grading: lab assignments (50%) and final exam (50%)
• Days: Tuesdays, from 9:00 to 10:30
• Prerequisites: linear algebra, probability and statistics, Python programming
• Syllabus: linear regression (weeks 1 and 2), logistic regression (weeks 3 and 4), decision trees (weeks 5 and 6), boosting (weeks 7 and 8), multi-layer perceptron (weeks 9 and 10), final exam (week 11)

Each topic is approached in two activities: the first covers theoretical content, the second discusses practical aspects and the assignments of the corresponding topic. Between the theoretical and the practical activity a mandatory assignment is provided, containing exercises to be implemented in Python. Students have five days in general to deliver it.

Resources

• Moodle hosts the slides and further course material for all modules (enrolment required)
• Assignments and mini-projects are delivered as Jupyter notebooks through Idiap’s JupyterHub, so no software installation is required on the student’s laptop

Cover: Idiap building by the Idiap Research Institute, used with attribution.