Max Robotics — Global Service Robot Directory + US Certification

About

Atsuto Maki is a leading researcher at the forefront of sensorimotor learning and active vision, whose work bridges robotics, computer vision, and reinforcement learning. His core research explores how robots can autonomously learn to interact with their environment through predictive action policies and sensorimotor contingencies—essentially, learning by doing. Maki’s most impactful contribution is his 2017 paper on "Deep Predictive Policy Training using Reinforcement Learning" (113 citations), which introduced a novel framework for training robots to anticipate the sensory consequences of their actions, overcoming the latency challenges inherent in sensorimotor control. This work has been foundational for developing more dexterous and responsive robotic systems. His earlier research on object categorization through action-effect relations (2015, 19 citations) pioneered a functional approach to robot perception, moving beyond static shape recognition. Maki also made notable contributions to binocular tracking and hand-eye coordination, demonstrating how robots can self-calibrate without prior kinematic models. His recent work on domain randomization for synthetic data generation (2025) extends his expertise to manufacturing applications, showcasing the practical impact of his research. With a career spanning from classic active vision (1996) to modern deep learning, Maki’s work consistently emphasizes the importance of learning from interaction, making him a key figure in the development of intelligent, autonomous robotic systems.