Edge AI: Robustness, Continual Learning and Real-World Deployment

Description

As AI systems increasingly move from centralized infrastructures to real-world edge environments, European industry faces both major opportunities and significant challenges in deploying adaptive and continually learning AI systems at scale. This session at the ENFIELD Final Conference will focus on the challenges of deploying adaptive AI in dynamic environments and the strategic advantages this technology can offer to key European sectors such as automotive, manufacturing, robotics, and industrial automation. Particular attention will be given to robustness, continual learning, trustworthiness, validation, safety, energy efficiency, and regulatory compliance in industrial Edge AI deployments.

The session will open with a 15-minute presentation providing an overview of the main scientific and technological results on Adaptive AI achieved within the project, together with the key challenges identified throughout the research and deployment activities. This introduction will be followed by a panel discussion involving representatives from European AI Networks of Excellence, research and innovation projects, and European industry-oriented associations. The debate will explore current barriers to real-world adoption, industrial requirements, emerging opportunities for European competitiveness and technological sovereignty, and future directions for trustworthy adaptive AI systems operating in continuously evolving environments.