Engineering Resilient Cognitive Systems (CIT 4230005)
Artificial intelligence (AI) and especially machine learning (ML) are future-oriented drivers of the economy and the key to shaping our future world. In this lecture, we will deal not only with ML, but with complex, learning-enabled software-intensive systems: cognitive systems.
Why should you be interested? Because ML is present not only in the IT world but also in cyber-physical cognitive systems such as autonomous driving, intelligent robotics, or medical products, to name just a few examples. This means that these technologies will pervasively shape our everyday lives. Therefore, safety is also of central importance - it is of utmost importance that we prevent such systems from being able to harm people. For example, the failure of the broad use of autonomous vehicles is not due to ML but to the lack of safety assurance.
In this lecture, which I teach in summer terms, you will learn how to contribute to this success-critical question. To do this, we will focus on the question: How can the engineering of resilient cognitive systems succeed? We will learn how to assure the safety of these systems to reduce the risk to humans to an acceptable level. But we will also talk about how these systems can optimize their utility in every conceivable and unimaginable situation and, at the same time, guarantee safety.
The topics we cover include:
- Resilience: Terminology, basic concepts, conceptual framework
- Safety methodology: Introduction to the terminology and the basic concepts of safety and reliability of software-intensive systems (with a focus on the so-called "Safety of the intended functionality" SOTIF)
- Safety Assurance for AI-based Systems: This involves addressing the unique challenges posed by AI in assuring safety, and establishing fundamental concepts for assuring the safety of AI-based systems.
- Adaptive safety architectures: Introduction to self-adaptive systems and their use to enable the resilience of cognitive systems despite unpredictable environmental conditions.
This lecture is a unique opportunity to gain deep insight into a central future technology and acquire the skills to contribute to its development. Be part of a safe AI future!