Digital Reality: From Learned Models via Synthetic Data to Trusted-AI - and Back

Explore the concept of Digital Reality in this invited talk at the NHR PerfLab Seminar by Prof. Dr.-Ing. Philipp Slusallek from Saarland University and the German Research Center for Artificial Intelligence (DFKI). Delve into the challenges of creating AI systems that accurately understand complex reality and learn about the modular approach of independently learning partial models. Discover how simulations can generate synthetic data for training, benchmarking, and certifying Trusted-AI systems. Examine the comprehensive approach developed at DFKI that integrates modeling, simulation, AI, and high-performance computing. Gain insights from Prof. Slusallek's extensive experience in computer graphics, artificial intelligence, and simulated reality as he discusses the importance of AI systems understanding their limitations and improving their digital models of reality.