In-Situ Artificial Intelligence for Self-* Devices: The Elastic AI Ecosystem (Tutorial)

Artificial Intelligence (AI) is an important topic for today's self-* systems. It can e.g. be used to analyze sensor data, to derive a model of a system's runtime situation, and to make dynamic adaptation decisions. To this end, in-situ AI is a powerful tool that enables individual devices to use AI autonomously, leading to truly decentralized self-* behavior. However, developing, deploying and executing in-situ AI is not trivial. Researchers therefore often fall back to classical Cloud-based AI solutions, which restricts what kinds of research studies are possible. In this tutorial we present our solution for this, the so-called elastic AI ecosystem.