Max Robotics — Global Service Robot Directory + US Certification

About

Frederik Ebert is a prominent robotics and machine learning researcher whose work sits at the intersection of deep reinforcement learning, predictive modeling, and robotic manipulation. His research has fundamentally advanced how robots learn to interact with the physical world using raw sensory inputs, without relying on extensive human supervision. Ebert is perhaps best known for his contributions to visual foresight and model-based deep RL, where his 2018 paper of the same name (264 citations) demonstrated that robots could acquire complex skills directly from visual observations. Complementing this, his work on stochastic adversarial video prediction (226 citations) pushed the boundaries of future-state modeling, enabling robots to reason probabilistically about their environment. His early self-supervised planning work (113 citations) laid important groundwork for autonomous skill acquisition through predictive learning. Beyond vision, Ebert explored tactile sensing for dexterous manipulation (121 citations), broadening the sensory modalities available to learning robots. More recently, his research has tackled generalization at scale — through large cross-domain datasets and zero-shot task transfer (89 and 83 citations, respectively) — reflecting a growing emphasis on building robotic systems that adapt broadly across tasks and environments. Together, his body of work has substantially shaped the trajectory of data-driven robot learning.