Inhalt

Abstract:

In the modern industrial landscape, machine learning (ML) has transitioned from a specialized computational niche to a foundational tool for engineers across all disciplines. This course provides a rigorous introduction to the methods and tools of ML, specifically tailored for engineering applications. As systems become increasingly data-intensive, the ability to extract actionable insights from sensor data, optimize manufacturing processes, and predict system failures is essential for the next generation of engineers.

Throughout the course, we explore how ML can solve complex, non-linear problems that traditional analytical models often struggle to address. We discuss inspiring use cases such as predictive maintenance (anticipating machinery failure before it occurs), computer vision for quality control (identifying defects in high-speed production lines), and load management (forecasting demand and optimizing resource allocation). By bridging the gap between mathematical theory and practical implementation, this course empowers students to treat data as a primary engineering material.

Gliederung:

The course is structured to guide students from the fundamentals of statistical learning to the implementation of advanced neural network architectures:

1. Foundations & motivation: We define the ML landscape, machine learning types (supervised, unsupervised, and reinforcement learning), and the standard machine learning pipeline.
2. Supervised learning - regression: We build the mathematical basis for predicting continuous values, including model formulation, optimization via least squares, and the use of basis functions for non-linear data.
3. Supervised learning - classification: We move from regression to categorical decision-making using the logistic regression framework, sigmoid functions, and cross-entropy loss.
4. Model reliability: We explore the balance between underfitting and overfitting, the bias-variance tradeoff, and cross-validation for model selection.
5. Advanced linear models: We introduce support vector machines (SVM), hard and soft margins, optimization strategies, and the application of kernels for high-dimensional mapping.
6. Dimensionality reduction: We explore unsupervised learning through principal component analysis (PCA), focusing on variance maximization, eigenvector decomposition, and feature extraction to simplify complex engineering datasets.
7. Neural networks & deep learning: We investigate the perceptron as a fundamental block, multilayer perceptrons (MLP) architectures, and the learning process through gradient descent and backpropagation.
8. Computer vision fundamentals: We examine the mathematical foundations of convolutional neural networks (CNNs), pooling layers for dimensionality reduction, and applications in image classification and pixel-wise segmentation.
   
   

Detaillierter Inhalt:

This course is delivered through a series of comprehensive video lectures that pair theoretical mathematical foundations with engineering intuition. The curriculum follows a logical progression starting with the basic machine learning pipeline and linear models for regression and classification. It then addresses critical concepts in model reliability and advanced linear techniques like support vector machines. Students will also explore dimensionality reduction using principal component analysis to handle high-dimensional sensor data efficiently. The final phase of the course transitions into deep learning, covering multilayer perceptrons and convolutional neural networks for spatial data analysis.

A central component of the learning experience is the practical application of theory through two major hands-on exercises:

• Regression for energy prediction: A supervised learning pipeline is implemented to forecast energy requirements for a milling process. The exercise covers data preprocessing, outlier management, and a comparative analysis of linear regression, random forests, and support vector regression (SVR) to optimize industrial load management.
• Deep learning for image classification: A quality monitoring system for electrical drive production is developed. Convolutional neural networks (CNNs) are utilized to detect assembly errors through multi-view imagery. The task explores data augmentation and multi-view analysis to identify defects that are difficult to capture from a single perspective.
  
  

By engaging with these video materials and coding assignments, participants gain the technical proficiency required to deploy machine learning solutions in industrial settings.

Good knowledge of linear algebra and calculus is a prerequisite for this course. Additionally, some prior knowledge of mathematical optimization techniques and general programming is recommended.

Lern-/Qualifikationsziele:

Upon successful completion of this course, students will be able to:

• Analyze engineering problems to determine the appropriate machine learning paradigm (regression, classification, or clustering).
• Implement a full machine learning pipeline, including data preparation, model training, and performance validation.
• Formulate and optimize linear and logistic models using gradient-based methods.
• Evaluate model performance critically, identifying and mitigating issues of overfitting or underfitting.
• Explain the intuition and mathematical mechanics behind support vector machines and the kernel trick.
• Understand the role of dimensionality reduction techniques like PCA to simplify and visualize complex datasets.
• Design and train basic neural networks for tabular and structured data.
• Design and train neural networks specifically optimized for processing visual data using convolutional and pooling layers.
  
  

Lehrveranstaltungstyp:

Virtuelle Vorlesung

Interaktionsformen mit Betreuer/in:

E-Mail

Interaktionsformen mit Mitlernenden:

Forum

Kursdemo:

zur Kursdemo

Nutzung

Kurs ist konzipiert für:

• Bachelor’s degree: 2-6 semester
• Master’s degree: all semesters
  
  

Maschinenbau; Mechatronik; Medizintechnik; Wirtschaftsingenieurwesen; International Production Engineering and Management; Elektrotechnik, Elektronik und Informationstechnik; Informations- und Kommunikationstechnik; Energietechnik; Materialwissenschaft und Werkstofftechnik; Elektro- und Informationstechnik; Medical Engineering and Data Science; Erneuerbare Energien und Energiemanagement

Formale Voraussetzungen:

Course enrolment via the Virtuelle Hochschule Bayern (vhb)

Erforderliche Vorkenntnisse:

Good knowledge of linear algebra and calculus is a prerequisite for this course. Additionally, some prior knowledge of mathematical optimization techniques and general programming is recommended.

Hinweise zur Nutzung:

-

Kursumsetzung (verwendete Medien):

Course consists of lecture modules containing videos with self-assessment questions, along with multiple hands-on exercises

Erforderliche Technik:

A computer able to run a web browser and, optionally, a Python development environment (a mobile device should also work but may not provide the optimal experience)

Nutzungsentgelte:

für andere Personen als (reguläre) Studenten der vhb Trägerhochschulen nach Maßgabe der Benutzungs- und Entgeltordnung der vhb

Rechte hinsichtlich des Kursmaterials:

-

Verantwortlich

Anbieterhochschule:

Uni München (LMU)

Anbieter:

Prof. Dr. Björn Eskofier

Autoren:

Jörg Franke

Nico Hanenkamp

Björn Eskofier

Betreuer:

Thomas Altstidl

Prüfung

Online course exam for students

Art der Prüfung:

schriftlicher Leistungsnachweis (Klausur)

Bemerkung:

The examination is online and the date is announced in the course.

Prüfer:

Prof. Dr. Björn Eskofier

Prüfungsanmeldung erforderlich:

ja

Anmeldeverfahren:

Registration for the exam takes place directly in the course.

Prüfungsanmeldefrist:

–

Prüfungsabmeldefrist:

–

Kapazität:

–

Prüfungsdatum:

–

Prüfungszeitraum:

–

Prüfungsdauer:

90 Minuten

Prüfungsort:

–

Zuständiges Prüfungsamt:

Examination office of the students' home university

Zugelassene Hilfsmittel:

None

Formale Voraussetzungen für die Prüfungsteilnahme:

Course enrolment via the Virtuelle Hochschule Bayern (vhb) and timely registration for the exam directly in the course

Inhaltliche Voraussetzungen für die Prüfungsteilnahme:

Course content

Zertifikat:

Ja (Proof of performance (graded))

Anerkennung:

–