Machine Learning Forum

Machine Learning Forum

At a glance


The Machine Learning Forum offers a wide variety of advanced training options for students and developers in industry. On one hand, the new university courses offered for Master’s students establish a wide range of options for deep immersion based on various seminars, practicums, and lectures. On the other hand, hands-on labs oriented toward industry work through the theoretical content, combining it with practical industry issues. This work aims to develop solutions for areas relevant to Germany, such as Industry 4.0, production, and the automotive sector. In addition, the Machine Learning Forum helps small and midsize businesses to use data made available in the context of digitalization by applying ML and to make decision-making processes easier.

The cooperation between the department chairs of Friedrich-Alexander University Erlangen-Nuremberg (FAU) and the departments at Fraunhofer is based on the transformation of lecture and seminar contents into practice-oriented, multiday workshops. The direct reuse and ongoing evolution of themes used in the cooperative development of scripts and content for exercises results in a continuous exchange between research and industry via labs and forums. This leads to continual improvement of the respective teaching concepts for students and industry developers.


Click here to see the various events

Seminars referring to Machine Learning

The goal of machine learning, a subfield of artificial intelligence, is not the explicit programming of the computer, but instead enabling it to learn independently from existing data.

The Machine Learning lecture at FAU provides insight into fundamental optimization processes, state-of-the-art machine learning approaches, and Monte Carlo methods. In addition, the associated Machine Learning seminar gives students an overview of various machine learning algorithms.

For an industry perspective, the industry lab Localization and Machine Learning is offered. The lab focuses on both the comprehensive implementation of machine learning projects and on examples in the areas of logistics, automotive applications, virtual reality, and localization that are relevant to practical situations.

The Machine Learning Forum event is a network platform that connects research and industry. In the future, it will take place twice a year. In addition to specialized lectures held by the university and practical lectures from industry experts, the event will stimulate the contribution of ideas and subjects from small and midsize businesses, job placement for graduates, and the initiation of (association) projects.


At a glance

Fraunhofer IIS is offering a two-day seminar on machine learning (ML) in a professional environment to help industry in the use of machine learning. In this seminar you will learn to successfully implement ML projects. From the definition of their business goals to the test and their use in live operation. A broad selection of treated algorithms and illustrative examples sharpens the view of the application areas in their company. Another important point is the identification of pitfalls in the processing chain of learning processes, as well as the use of appropriate countermeasures or "best practices".


Learning objectives an competences

By participating in the seminar, you will:

  • understand the basics of machine learning
  • learn to cellect and structure data effeicently
  • get to know unsupervised and supervised learning methods
  • use proven procedures for quick results and decisions
  • get to know examples of the use of machine learning in context
  • detect and eliminate errors in the processing chain



Seminar: Machine Learning

Day 1: theory and practice Day 2: Project ans Best practice
  • Methods of unsupervised learning
  • Practical examples
  • Algorithms of supervised learning
    • Regression
    • Classification
  • Practical examples
  • Insight into process mining for process transparency
  • Practical examples

Implementation of a ML project

  • From business goal to live operation

Detect and correct errors

  • Over and underfitting
  • Bias / Variance Trade-Off
  • parameter optimization
  • Unbalanced records
  • Concept drift of data sources


Who should attend?

  • Technical decision makers
  • Developers and engineers in industrial companies
  • Inquisitive of all kinds

Seminars referring to Deep Learning

Deep learning is one of the key technologies for future developments in the field of digital signal processing, such as computer vision. Conventional methods reach their limits in countless tasks in the industry, e.g. due to changing image recording conditions or large variations of the test objects. Deep Learning overcomes such difficulties by using large amounts of sample data. Signal processing is about to take a promising leap in development, opening up entirely new fields of application.  


At a glance

Deep Learning is a technology inspired by the way the human brain works. Multilayer artificial neural networks are capable of analyzing large amounts of data for specific tasks and recognizing the relevant solution patterns. For example, for visual quality control, a distinction between good and bad parts can be learned automatically simply by using a sufficient number of example images.

The "Deep Learning Theory & Applications" lecture is already offered at FAU and provides an overview of neural networks and their differentiation from established machine learning approaches.

Based on this lecture, Fraunhofer IIS offers two day-long workshops "Deep Learning and Computer Vision" as part of the ML-Forum, thus building a bridge from university lectures to industrial applications. The focus of both workshops is in the field of computer vision. In addition, other topics will be addressed to give you an insight into the variety of possible application domains.

In a well-balanced combination of theory and practice, module 1 introduces the topic with the basics of deep learning and concrete examples from image analysis. Module 2 expands your knowledge from module 1 with a deeper look into the functionality of deep learning architectures and other practical exercises. You will have the opportunity to test the use of concrete tools and to experience Deep Learning "in action". In this way you get to know possibilities and limits and will develop a feeling for the potential of this technology for your own company.  


Learning objectives and competences

By participating in the seminars, you will:

  • understand the basics of deep learning
  • get to know various types of artificial neural networks
  • learn to implement basic deep learning workflows for image analysis
  • manipulate training parameters to improve performance
  • learn to deal with common challenges



Workshop I: Basics of Deep Learning

Deep Learning and Computer Vision I  
  • Basics of machine learning
  • Introduction to deep learning
  • Introduction to TensorFlow
  • Multilayer perceptrons
  • Convolutional Neural Networks
  • Hands-on exercises


Workshop II: Advanced topics of deep learning

Deep Learning and Computer Vision II
  • Semantic segmentation
  • Object detection
  • Generative Adversarial Networks
  • Sequence models


Who should attend?

  • industrial managers with strategic decision-making responsibility
  • key members of internal R & D teams
  • algorithm engineers and programmers
  • industrial consultant


Goodfellow, I., Bengio, Y., Courville, A., & Bengio, Y. (2016). Deep learning (Vol. 1). Cambridge: MIT press.

Géron, A. (2017). Hands-on machine learning with Scikit-Learn and TensorFlow: concepts, tools, and techniques to build intelligent systems. "O'Reilly Media, Inc.".

Tibshirani, R., & Friedman, J. (2001). The elements of statistical learning: data mining, inference, and prediction. Heidelberg: Springer.

Feel free to contact us, if you’re interested in our workshops – we look forward to hearing from you!

Jens-Uwe Garbas

Contact Press / Media

Dr.-Ing. Jens-Uwe Garbas

Head of Group Intelligent Systems

Fraunhofer IIS
Am Wolfsmantel 33
91058  Erlangen, Germany

Phone +49 9131 776 5160

Fax +49 9131 776 5108

Seminars referring to Deep Learning Hardware Architectures

Because the hardware in deep learning applications is subject to difficult requirements, special hardware architectures are introduced and discussed in this module.

Deep Learning Hardware Architectures, a new course in the form of a practicum, will be held in the Computer Architecture department. In the theoretical section, students become familiar with various hardware architectures and debate their suitability for deep learning algorithms. The practical section includes trials of the various hardware architecture using application examples related closely to industry.

A Deep Learning Hardware Architectures lab based on the practicum content will be designed. Initially, this lab will introduce the architecture of a selected deep learning system and demonstrate its suitability for certain applications. In the practical section, the system will be implemented based on a specific example.

Seminars referring to Reinforcement Learning

Reinforcement Learning (RL) represents another step toward true AI. In RL an agent learns which actions to take under specific situations in order to solve a given problem. Unlike Supervised Learning, here the agent is not provided with labeled examples of the correct actions; instead the agent needs to discover a strategy that maximizes reward through trial-and-error.

The “Reinforcement Learning” practicum offered at the University is an in-depth course on the topic. Through an engaging mix of theory, practical exercises and a final Project, the practicum provides both the necessary theoretical foundations as well as hands-on experience on state-of-the-art algorithms in the field.

The two Industry Labs – described below in detail – are focused on the application of Reinforcement Learning to real-world business use cases from the Industry


At a glance

What is Reinforcement Learning (RL) and how can your company benefit from it?

Fraunhofer IIS offers two seminars on the thematic areas of: (1) Reinforcement Learning Foundations and (2) Advanced Topics in Reinforcement Learning, to enable you to answer these questions. Through a combination of theory and engaging case examples from the industry, these seminars will allow you to understand the value and the implications of this technology in your business. With the help of hands-on exercises, you will learn how to re-formulate several problem types to the RL paradigm and design (or select from a large base) efficient algorithms to solve them. In addition, implementing RL solutions in real-world use cases in the “project” sessions will equip you with a solid understanding on how to apply RL algorithms in practice using available state-of-the-art software frameworks.


Learning Objectives

Through participation in the seminars you will:

  • Understand the foundations of Reinforcement Learning
  • Learn how to formulate a given problem within the Reinforcement Learning paradigm
  • Study the different types of Reinforcement Learning algorithms
  • Implement Reinforcement Learning Algorithms for real-world problems using state-of-the-art software
  • Learn how to apply Reinforcement Learning in real-world autonomous systems



Seminar 1: Reinforcement Learning Foundations

Day 1: The basics Day 2: Project and Best Practice
  • Markov Decision Processes and Dynamic Programming
  • Monte Carlo and Temporal Difference methods
  • Q-Learning
  • Value Function Approximation
  • Practical Exercises
  • DQN algorithm
  • Practical Project:
    • Define state/action spaces and reward functions in real-world problems
    • Evaluate the learning progress of an autonomous agent
    • Follow best practices for stable learning
    • Work on State-Of-The-Art tools for (Deep) Reinforcement Learning


Seminar 2: Advanced Topics in Reinforcement Learning

Day 1: Deep Reinforcement learning in practice
  • Deep Reinforcement Learning:
    • Policy Search and Actor Critic Methods
    • Model-based Reinforcement Learning
    • End to End Reinforcement Learning
    • Reinforcement Learning for real-world Autonomous Systems
  • Practical Project:
    • Solve real use cases from the Industry
    • Compare different Deep Reinforcement Learning approaches


Who should attend this seminars?

  • Industry managers with strategic decision-making responsibilities
  • Key members of in-house R&D teams
  • Algorithm engineers and programmers
  • Industry consultants



RS. Sutton, AG. Barto: Reinforcement learning: An introduction (2nd edition, in progress), Cambridge: MIT press, 2018.

C. Szepesvári: Algorithms for reinforcement learning, Morgan and Claypool, 2009.

DP. Bertsekas: Dynamic programming and optimal control (Vol I and II), Belmont, MA: Athena scientific, 2017.

A. Geramifard, et. al.: A tutorial on linear function approximators for dynamic programming and reinforcement learning, Foundations and Trends® in Machine Learning, 2013.

I. Goodfellow, Y. Bengio, A. Courville: Deep learning, Cambridge: MIT press, 2016.

V. Mnih, et. al., Human-level Control through Deep Reinforcement Learning, Nature, 2015.

TP. Lillicrap, et. al.:  Continuous control with deep reinforcement learning, arXiv preprint arXiv:1509.02971, 2015.

MP. Deisenroth, G. Neumann, J. Peters: A Survey on Policy Search for Robotics, Foundations and Trends® in Robotics, 2014.



Machine Learning for Smart Sensing and Electronics