Please refer to the respective ILIAS course for the specific dates.

Requirements

Basic programming knowledge is required. Knowledge of the fundamentals of information technology, signal and system theory, and probability theory is also required.

Content

The topics of machine learning and artificial intelligence are playing an increasingly dominant role in our society. This is due to the increasing computational power and availability of new technologies (modern GPUs, multi-cores, FPGAs, etc.) combined with efficient parallelized frameworks and algorithms for training neural networks and generative AIs.

As data volumes are collected and processed in all walks of life, the areas of application and research in machine learning extend far and wide. For example, Convolutional Neural Networks (CNN) are already replacing classical object recognition methods in image processing for autonomous driving. In medical technology, work is underway on Artificial Neural Networks that can already detect malignant skin lesions from images more reliably than doctors. Through models such as the Generative Pre-trained Transformer (GPT), AI applications such as ChatGPT, Bard and Llama have entered the everyday lives of many people, including students. Machine learning is thus becoming increasingly important as a component of information technology for aspiring electrical engineers.

In order to meet these future challenges as an engineer, it is essential to build up the necessary competences during the studies as well as to gain a basic understanding of the methods and tools in the field of machine learning.

In the course, you will learn the practical use of common algorithms and methods of machine learning in a project-based and "hands-on" approach. You will work in groups and with powerful workstations to implement algorithms and structures such as Perceptron, Decision Trees or Evolutionary Algorithms. At the same time, you will also learn how to use methods and tools that are widely used in business and science (Keras, Tensorflow, Pytorch, Cuda,...) to solve real-world problems (rental prices, medical data,...).

In the "Into-the-Wild" part, you have the exciting opportunity to choose from applied problems from research or your own ideas! The range is wide, from detections in particle physics to predicting soccer games - it's all been there!

Organizational

Attendance is compulsory for all dates and the preliminary meeting.

• You will work on the assignments in teams of two or three.
• The assessment consists of the submitted assignment sheets, the "Into the wild" part and a colloquium.
• More specific details about the assessment will be explained in the introductory session.
• Six ETCS points will be awarded.

Contents

• Preliminary discussion and group assignment
• Processing and analysis of data sets
• Evaluation of ML systems
• Basics of supervised learning
• Unsupervised learning
• Neural Networks
• Evolutionary Algorithms
• Convolutional Neural Networks
• Recurrent Neural Networks
• Into the Wild...

Procedure of the lab dates 1-8

• During each lab session the students have to work on given tasks. These include programming tasks as well as tasks that are to be answered in text form.
• The task sheets are handed out in the form of Jupyter notebooks. With this interactive development environment you can test program codes directly in the task sheet, display solutions and complete the documentation or answers.
• After each lab date, the processed task sheet (the corresponding Jupyter notebook) is handed in.
• It is sufficient to hand in one copy of the notebook per group.
• Successful completion and submission of the assignment sheets is a prerequisite for participation in the oral colloquium at the end of the semester. In case of non-submission, the respective participants will not be admitted to the colloquium!

Into the Wild...

• At the beginning of the second part, different data sets will be presented to choose from for processing.
• Each group decides on one of the presented data sets, of course several groups can use the same data set. The problem is defined by the group itself. It is also possible to create your own problems for your own data sets.
• Based on what has been learned before, a concept for solving the problem is developed. This is implemented and tested.
• Besides working on the data, it is also possible to optimize an existing approach for a data set with respect to runtime or latency. Among other things, this can be achieved by means of suitable hardware.
• Finally, each group prepares a presentation that introduces the developed concept and presents the results. In doing so, a critical reflection of the previously made decisions should also point out possible next steps.

Application procedure

This year, 30 places are expected to be offered for the lab.
Registration for the lab is done through the WiWi portal at the following link: https://portal.wiwi.kit.edu/ys/7360

Materials

The necessary materials for the LAMA will be made available through ILIAS. After successful registration for the LAMA you will get access to the assignments, data sets and additional information material.
The material is provided in English, but can be edited in German.

Tutors wanted

For former participants or other students in the Master's program who would like to work on the topics, there are opportunities to be a tutor in the course. If you are interested, please send an email with the appropriate subject to the contact address: lama∂itiv.kit.edu.