Prof. Imre Rudas

Obuda University, Hungary



Title of the paper: From Virtual Models to Robotic Minds: Digital Twins as Enablers of Embodied AI


Abstract: Embodied artificial intelligence (AI) is driving a paradigm shift in robotics, demanding agents that can perceive, learn, and act autonomously within complex real-world environments. However, training such intelligent agents requires massive, diverse datasets and extensive physical interaction, posing significant challenges in safety, cost, and scalability. Digital twins—dynamic, real-time virtual replicas of physical systems—have emerged as a foundational solution, offering simulated environments for iterative training, policy optimization, and validation before real-world deployment. This plenary talk explores the transformative role of digital twins as enablers for embodied AI, detailing how they bridge the gap between simulation and physical reality, accelerate learning through synthetic data generation, and enable closed-loop, reproducible experimentation. Recent advances in foundation models, self-driving laboratories, and large-scale video learning illustrate digital twins’ capacity to unlock new levels of generalizability, adaptability, and innovation in robotics. Drawing from the latest research and industry trends, this session highlights the critical intersection of digital twin technology and embodied AI, proposing future directions for robust, scalable, and transparent intelligent systems in Industry 5.0 and beyond.

Bio: To be announced soon




Categories