Max Robotics — Global Service Robot Directory + US Certification

About

Tom Silver is an emerging force in AI and robotics research, with a focus on task and motion planning (TAMP), neuro-symbolic learning, and reinforcement learning for complex robotic systems. His work addresses one of the central challenges in modern AI: enabling robots to reason and plan efficiently across continuous, high-dimensional environments with many objects and long task horizons. Silver's most influential contribution, "Planning with Learned Object Importance" (62 citations), demonstrates how graph neural networks can dramatically scale planning by identifying only the objects relevant to solving a given task. His pioneering work on Residual Policy Learning (52 citations) introduced an elegant method for refining imperfect controllers using model-free reinforcement learning, gaining rapid traction in the robotics community. Across multiple papers, he has developed Neuro-Symbolic Relational Transition Models (NSRTs) and frameworks for predicate invention and abstraction discovery — innovations that bridge the gap between data-driven learning and structured symbolic reasoning. Silver's research consistently emphasizes generalization: building systems that learn from few examples and transfer across problem instances. With over 200 cumulative citations across his published work, his contributions are shaping the future of intelligent robot planning, making him a researcher to watch closely in the coming years.