AI applications (5 cr)

Objective

You know application areas of artificial intelligence and machine learning.
You are able to choose a suitable method to implement an artificial intelligence application.
You understand the operating principles and limitations of artificial intelligence.

Content

What is artificial intelligence, machine learning, and deep learning?
What type of machine learning methods exist and what kind of problems can they solve?
How do you implement an artificial intelligence application?

Evaluation

Students can

a. use professional vocabulary systematically.
b. look for information in the key information sources of the field.
c. identify interrelated tasks.
e. use the key models, methods, software and techniques of the professional field

Course material

All materials are available on the learning platform. These include lecture slides, book chapters, and additional supporting resources (templates, cheat sheets, etc.).

Study forms and methods

Each lecture day combines theoretical and practical concepts. Lectures provide a theoretical overview of each topic. Students then work independently to deepen their practical understanding through readings, exercises, and assignments.

There are hands-on exercises (assignments) and two projects covering specific topics. All assignments and projects are mandatory to pass the course (see assessment criteria below).

Timing of exams and assignments

There are no formal exams. Instead, students must submit their exercises and assignments by the given deadlines to receive a grade. Every assignment is required to pass the course. Assignments submitted after the deadline can earn a maximum of 50% of the grade.

Student workload

The course is worth 5 ECTS, equivalent to 135 hours of work. One-third of the workload is completed in classrooms, and the remaining two-thirds are dedicated to projects and assignments.

Course part description

This course covers fundamental AI concepts and practical applications relevant to software engineers.
Throughout the course, students learn how to design, develop, and evaluate AI-driven solutions, reinforcing both theoretical and hands-on skills.

Topics include:

- Overview of AI: History, definitions, and key areas (machine learning, deep learning, etc.).
- Machine Learning Foundations: Supervised vs. unsupervised learning, model evaluation, overfitting/underfitting.
- Neural Networks and Deep Learning: Basic neural network architectures, feed-forward networks, introduction to frameworks like TensorFlow or PyTorch.
- Data Preprocessing and Feature Engineering: Handling missing data, feature selection, and transformation techniques.
- Natural Language Processing: Tokenization, text representation, sentiment analysis, and simple chatbots.
- Computer Vision: Image classification, object detection, and popular architectures.
- Ethical and Responsible AI: Fairness, bias, and societal impacts of AI.
- Practical Implementation: Best practices for using Python (Jupyter notebooks, scripts) and relevant AI libraries (e.g., scikit-learn, Keras).

Further information

Python is used as the main programming language.
Students primarily work with Jupyter notebooks and scripts.
Depending on the chosen project, additional coding skills (e.g., APIs, basic front-end, version control with Git) may be required.

Evaluation scale

1-5

Assessment methods and criteria

Assessment criteria:
- Homework Assignments (40%): Weekly assignments and exercises.
- Midterm Project (20%): A project applying AI to a given problem.
- Final Project (30%): A second project applying AI to a different problem.
- Attitude (10%): Individual engagement and professional conduct throughout the course.

Note: Any assignment or project delivered after the deadline is worth a maximum of 50% of the grade. Also, completing the usual course feedback is mandatory to unlock the final grade in Peppi.