Max Robotics — Global Service Robot Directory + US Certification

About

Johannes Kulick is a roboticist whose research sits at the intersection of machine learning, active perception, and interactive robot learning. His work focuses on enabling robots to autonomously explore and understand their physical environment—a challenge he tackles through information-theoretic approaches and human-robot teaching scenarios. Kulick’s most influential contribution is his 2013 paper on active learning for teaching robots grounded relational symbols (68 citations), which investigates how robots can learn abstract, generalizable models through interactive instruction. He has also made significant advances in physical exploration, developing entropy-based strategies that allow robots to autonomously discover and manipulate an environment’s degrees of freedom, such as opening doors or drawers. His 2017 work on opening a lockbox through physical exploration (15 citations) draws inspiration from animal cognition, specifically cockatoos, to bridge the gap between biological and robotic intelligence. Across his publications, Kulick demonstrates a consistent focus on active, information-gathering strategies—including a notable 2014 comparison of cross-entropy versus entropy for iterative information gathering—that push the boundaries of how robots can learn from sparse data and complex, joint-dependent structures.