Max Robotics — Global Service Robot Directory + US Certification

About

Juergen Schmidhuber is a pioneering figure in artificial intelligence, renowned for his foundational work in artificial curiosity, intrinsic motivation, and developmental robotics. His research focuses on enabling autonomous agents—from humanoid robots to space explorers—to learn complex sensorimotor skills through self-driven exploration, rather than external rewards. A key contribution is the Curiosity Driven Modular Incremental Slow Feature Analysis (CD-MISFA), which allows robots to form stable, invariant sensory representations by maximizing learning progress. This concept is demonstrated in his highly cited works, including "Continual curiosity-driven skill acquisition from high-dimensional video inputs for humanoid robots" (51 citations) and "Learning skills from play: Artificial curiosity on a Katana robot arm" (30 citations), where robots autonomously learn to grasp, stack blocks, and perceive their environment. Schmidhuber also introduced Upside Down Reinforcement Learning (UDRL, 23 citations), transforming RL into supervised learning by using rewards as inputs. His theoretical contributions include Self-Delimiting Neural Networks (22 citations), bridging algorithmic information theory and neural networks. With over 100,000 citations overall, Schmidhuber’s work on curiosity-driven learning and self-improving systems continues to shape AI research, inspiring new generations of intrinsically motivated agents.