Machine Learning for Safety Experts C2208
Topics: Advanced Technologies
The course will enable the learner to apply the fundamental principles behind safety of machine learning to a wide range of applications. The course guides learners through an appropriate selection of methods and tools tailored to the learner’s specific projects. With the acquired knowledge the learner will be able to shape the development and assessment of ML-based safety-related functions enabling their teams to leverage the power of advanced ML techniques without undermining safety.
Learning Objectives
By successfully completing this course, you’ll be able to:
- Recognize the impact of machine learning (ML) on functional safety (ISO 26262) and the safety of the intended functionality (ISO 21448).
- Apply key concepts and terminology associated with machine learning relevant to the formulation of a safety assurance argument (e.g., robustness, bias, prediction certainty)
- Develop a project-specific safety-lifecycle that integrates ML-specific safety activities and safety artefacts into an overall system-level development process
- Derive specific safety requirements for an ML-based function from a system level context
- Analyze a given ML-based function with respect to its safety-related properties and identify insufficiencies relevant to the safety of the function
- Know and apply established methods at design time and operation time to address insufficiencies of the most common ML techniques.
- Evaluate the effectiveness of new ML techniques and countermeasures for insufficiencies toward ensuring the safety of an ML-based function
Who Should Attend
The target learner is an automotive safety engineer of any level of experience who is confronted with the need to assess and ensure the safety of functions and systems based on components using some form of machine learning. As a guidance in times of emerging standards and best practices in a rapidly developing field, the learner will be provided with a structured framework as well as a comprehensive toolkit of methods to become an enabler for the safe use of machine learning in increasingly critical and challenging applications.
Prerequisites
Pre-existing safety knowledge is mandatory, experience of applying standards such as ISO 26262 and ISO 21448 in automotive projects Basic understanding of machine learning techniques
You must complete all course contact hours and successfully pass the learning assessment to obtain CEUs.
Karsten Roscher and Philipp Schleiss
Karsten is a computer scientist at heart and an engineer by choice. With over a decade of research experience in connected mobility, advanced driver assistance systems and machine learning for computer vision and communication systems he is a strong advocate for safe intelligence and the ongoing efforts to reconcile the need for safety with the seemingly unpredictable nature of machine learning. In March 2020, he took over as head of the Dependable Perception & Imaging department at Fraunhofer IKS supporting customers and partners in the development of reliable perception systems in the automotive, rail, industrial automation and medical domain. Karsten is an author of numerous scientific publications and active in both German and European standardization bodies.
Philipp has been researching methods for increasing the dependability of embedded and autonomous systems at Fraunhofer since 2013. In 2019 he started building a department in the area of systems safety engineering to successfully advance the state-of-the-art in the field of safety. Currently, his research interests are focused on tool-based design automation, continuous safety assurance and real-time guarantees for automated driving.
