Facing Uncertainty in AI - From Verification to Synthesis

Explore the challenges of uncertainty in AI systems through this insightful talk by Prof. Joost-Pieter Katoen from RWTH Aachen University. Delve into the critical question of guaranteeing safety and dependability in AI systems amidst various forms of uncertainty, including noisy data, randomized mechanisms, imperfect visibility, and unpredictable environments. Learn about model-based formal verification and synthesis techniques, with a focus on automation, to address these challenges. Discover how these approaches can be applied to verify randomness and complete partial designs in AI systems through practical examples. Gain valuable insights from a distinguished expert in software modeling, verification, and probabilistic computation, and understand the importance of formal methods in ensuring reliable AI systems.