Max Robotics — Global Service Robot Directory + US Certification

About

Kyriacos Shiarlis is a leading researcher at the intersection of robot learning, human-robot interaction, and artificial intelligence. His work focuses on enabling robots to acquire complex, socially appropriate behaviors through Learning from Demonstration (LfD) and Inverse Reinforcement Learning (IRL). Shiarlis made a significant contribution to the field with his work on "Rapidly exploring learning trees," which addresses a critical bottleneck in IRL by allowing robots to learn cost functions for path planning without repeated costly planning procedures. He further advanced task decomposition with "TACO," a method that learns to break down complex tasks into reusable sub-policies, improving data efficiency and generalization across tasks. His work on the TERESA project demonstrates a commitment to socially intelligent robotics, developing a semi-autonomous telepresence system for elderly care. Shiarlis also pioneered the use of deep learning from demonstration to acquire social interaction behaviors for telepresence robots, moving beyond hard-coded social norms. With over 90 citations across his key works, Shiarlis's research is shaping the future of robots that can learn, adapt, and interact seamlessly in human environments.