About
Course description
Machine Learning (ML) is the methodology behind the data science successes achieved by so many industries over the past decade. This course teaches the skills to understand and correctly apply ML at scale. Students will gain exposure to fundamental ML concepts, software tools (notably scikit-learn), the machinery of the best-known algorithms, and the often-overlooked but immensely important tasks of algorithm calibration and evaluation. Great emphasis is put on developing intuitions about models and algorithms. The class is a stepping stone to more advanced ML challenges, in a professional context or subsequent academic studies.
Because ML is a cross-disciplinary field, which includes statistical analysis and computer science, this course is for students with a background in both.
Course format. Fundamental concepts are presented during didactic bite-sized lessons, then illustrated with frequent exercises and a weekly real-world application, ranging from natural language processing to biology. Software tools and best practices are introduced along the way.
Topics:
- Intro to ML concepts (typology of tasks, optimization basics, what does “learning” mean?)
- Simple classification and regression
- Building a complex ML pipeline
- Selecting the best model
- Hyperparameter tuning
- Linear models
- Decision tree models
- Ensembles of models
- Evaluating model performance
Meeting times
| Day | Time | Building | Room |
|---|---|---|---|
| Monday | 10:35 AM – 11:55 AM | C | C-302 |
| Thursday | 10:35 AM – 11:55 AM | C | C-302 |
Final exam: 2026-12-14, 9:00 AM, C-302.
Syllabus and learning outcomes
The full syllabus and tentative schedule is at this address
The learning outcomes are:
- Formulate a real-world problem as a machine learning task (classification, regression, clustering).
- Understand the difference between supervised and unsupervised learning.
- Train, evaluate, and compare machine learning models appropriately (train/test splits, cross-validation, hyper-parameter tuning, relevant metrics).
- Recognize overfitting/underfitting and apply strategies to address them.
- Communicate modeling choices and results clearly, in words, numbers, and graphs.
- Have familiarity and fluency in machine learning concepts and tools, especially scikit-learn.
- Reason critically about the limitations of machine learning models.
Grading policy
- Midterm: 30%
- Final exam: 30%
- Weekly tests: 30%
- Attendance: 10%
Attendance policy
Students at The American University of Paris are expected to attend all scheduled classes; in case of absence, students should contact the instructor to explain the situation. Attendance at all exams is mandatory. Responsibility for communicating about missed course meetings and arranging to make up missed work rests solely with the student.