Laboratory for Applied Machine Learning Algorithms (LAMA)

Prerequisites

Basic programming skills are required. A solid understanding of the fundamentals of information technology as well as signal and system theory is also required. The lab provides the necessary expertise to understand and leverage the opportunities and challenges of AI in future professional settings.

Note: Unfortunately, master’s students cannot be admitted.

Content: Hands-on experience with real-world AI challenges

The relevance of machine learning and artificial intelligence (AI) in modern society is undeniable. This course offers students a thorough, practice-oriented introduction to the methods and tools of machine learning.

We cover a broad spectrum—from image processing and natural language processing to reinforcement learning.

The course offers:

• Implementation of fundamental algorithms (e.g., perceptrons, decision trees)

• Application of industry-relevant tools (e.g., PyTorch)

• Insights into the architecture and functioning of CNNs and RNNs, and an introduction to transformer models

• Use of high-performance workstations for hands-on exercises

Course Structure and Project Work

The course is divided into two parts. During the first lab sessions, you will work on assigned tasks in teams of two. These include programming assignments and theoretical problems that are solved in interactive Jupyter Notebooks. Successful completion and submission of these assignment sheets is a prerequisite for participating in the oral colloquium at the end of the semester.

The second part is the “Into the Wild” section. Here, you have the creative freedom to pursue your own project ideas or to engage with current research questions; for example, we’ve had projects ranging from GeoGuesser to medical imaging.

The grade is based on the completed worksheets, the “Into the Wild” section, and the colloquium. Attendance is mandatory for all sessions, including the preliminary meeting.

Schedule for Lab Sessions 1–8

• Preliminary Meeting and Group Assignment

• Processing and Analysis of Datasets

• Fundamentals of Supervised Learning

• Unsupervised learning

• Neural Networks

• Fundamentals of Reinforcement Learning and Evolutionary Algorithms

• Convolutional Neural Networks

• Transformers and Generative AI

Organizational Information and Course Content

Tutors Wanted!

Former participants or master’s students who are familiar with the topics can participate in the course as tutors. If you are interested, please send an email with the appropriate subject line to: lama∂itiv.kit.edu.

FAQ:

How do the lab sessions work?

During each lab session, students will work on assigned tasks. These include both programming exercises and tasks that require written answers.

• The assignment sheets are provided in the form of Jupyter Notebooks. Using this interactive development environment, you can test program code directly in the assignment sheet, display solutions, and add documentation or answers.

• After each lab session, the completed assignment sheet (the corresponding Jupyter Notebook) must be submitted.

• Submitting one copy of the notebook per group is sufficient.

• Successful completion and submission of the assignment sheets is a prerequisite for participating in the oral colloquium at the end of the semester. Participants who do not submit their work will not be admitted to the colloquium!

Is attendance mandatory?

• Attendance is mandatory for all lab sessions, including the introductory session.

• Attendance is required both to carry out the work in teams on-site and to provide hands-on instruction in techniques and skills that cannot be learned through private study alone.

What are the details regarding ItW?

• At the beginning of the second part, various datasets will be presented from which students can choose to work on.

• Each group selects one of the presented datasets; of course, multiple groups may use the same dataset. The problem statement is defined by the group itself. Groups may also develop their own problem statements based on their chosen datasets.

• Based on what has been learned previously, a concept for solving the problem is developed. This concept is then implemented and tested.

• In addition to working with the data, it is also possible to optimize an existing approach for a dataset in terms of runtime or latency. Among other things, this can be achieved by using appropriate hardware.

• Finally, each group will prepare a presentation introducing the developed concept and presenting the results. Through a critical reflection on the decisions made earlier, the presentation should also highlight possible next steps.

I’m still on an Erasmus exchange and can’t attend the final sessions—what should I do?

• The general rule is: Send us an email and let us know.

• As long as there are enough spots available, you can participate.

• If there aren’t enough spots, priority will be given.

• If you don’t notify us when you register, you may not be able to complete the examination requirements.