Introduction to Machine Learning and AI (S25)

Code

IT-MAL1

Version

4.0

Offered by

ICT Engineering

ECTS

Prerequisites

The course requires basic knowledge about algorithmic thinking, and in general a good understanding of mathematical concepts, obtained through SW-DMA1 or SW-MSE1 or equivalent mathematics courses..
Some programming experience is also presupposed but not strictly required.

Main purpose

This course offers a comprehensive introduction to the core methodologies of machine learning and artificial intelligence, providing both theoretical foundations and hands-on experience. Students will work with a variety of data types, spanning both structured and unstructured datasets, to develop practical skills essential for solving real-world problems. The course emphasizes an understanding of how to analyse, prepare, and explore data before applying machine learning algorithms to uncover patterns and make predictions.

Students will become adept at selecting and tuning machine learning models, all while critically evaluating their performance using relevant metrics. The course fosters the ability to address real-world problems with tailored machine learning solutions.

Key topics:
- Classification: Learning to categorize data into predefined classes.
- Regression: Making accurate predictions of continuous outcomes based on input data.
- Clustering: Unveiling hidden groupings in data.
- Dimensionality Reduction: Simplifying high-dimensional data without significant loss of information.

Knowledge

By the end of the course, students will have in-depth knowledge of key machine learning algorithms, methodologies, tools, and applications, including:
- Data Preparation & Preprocessing: Handling missing data, normalization, and feature engineering.
- Classification Algorithms: Naïve Bayes, k-Nearest Neighbor, Decision Trees, Logistic Regression, Support Vector Machines, Neural Networks.
- Regression Techniques: Simple linear regression, multiple linear regression, Ridge and Lasso regression.
- Dimensionality Reduction Algorithms: Principal component analysis and t-SNE
- Clustering Methods: k-Means, Agglomerative Clustering, DBSCAN.
- Model Evaluation Metrics: Accuracy, precision, recall, F1-score, MSE, cross-validation.

Skills

Upon completion, students will have developed:
- The ability to preprocess data and prepare it for machine learning tasks.
- Proficiency in implementing and fine-tuning classification models using real-world datasets.
- Skills to apply and interpret regression models to predict continuous variables.
- The capability to reduce the dimensionality of datasets while preserving important information.
- Competence in clustering unlabelled data and determining optimal cluster numbers.
- Expertise in using leading machine learning tools (e.g., Scikit-Learn, Keras, TensorFlow).
- The ability to critically assess and improve model performance using various validation techniques.

Competences

Upon successful completion of the course, students will:
- Confidently select appropriate machine learning techniques for specific tasks and problem domains.
- Tune machine learning algorithms to optimize performance for unique datasets.
- Design and implement machine learning systems to solve complex real-world problems.
- Communicate and justify machine learning solutions and decisions to both technical and non-technical stakeholders.

Topics

Teaching methods and study activities

The mode of teaching will be classroom based and will involve lectures by the teacher and project work in class.
There are a total of six group projects during the semester.

Resources

Hands-On Machine Learning with Scikit-Learn, Keras & TensorFlow by Aurélien Géron. 3rd edition, 2022.

Additional material will be uploaded to Itslearning.

Evaluation

Examination

Exam prerequisites:
None

Type of exam:
The exam is a 20-minute oral examination that departs from one of the six group assignments that the student has handed in during the semester, in accordance with deadline.
The exam will also include a discussion of one the other assignments.
External assessment.

Tools allowed:
N/A

Re-exam:
Conducted as the ordinary exam.

Grading criteria

Grading according to the Danish 7-point scale.

Additional information

In spring semester 2025 this course is mandatory for 4. semester and also elective for 6. and 7. semester.

However, there is a difference:
4. semester has external assessment, while the 6. and 7. semesters have internal assessment.

Responsible

Richard Brooks (rib)

Valid from

2/1/2025 12:00:00 AM

Course type

Keywords

Classification, data mining, neural networks, regression, decision trees, algorithms, support vector machines, principal component analysis, clustering