Max Robotics — Global Service Robot Directory + US Certification

About

Thomas L. Griffiths is a leading researcher at the intersection of cognitive science, robotics, and human-robot interaction. His work focuses on developing computational models that enable robots to understand and predict human behavior, with a particular emphasis on goal inference and value alignment. Griffiths' most significant contribution is demonstrating how robots can infer human goals to improve both objective performance and perceived collaboration quality—a finding that has garnered over 47 citations across multiple publications. His research on "Pragmatic-Pedagogic Value Alignment" (2019, 16 citations) explores how robots can learn and adapt to human values through interactive teaching, while "How to Be Helpful to Multiple People at Once" (2020) addresses the critical challenge of multi-user preference aggregation in assistive robotics. More recently, Griffiths has pioneered work on "Preference-Conditioned Language-Guided Abstraction" (2024), using natural language to create more generalizable learning representations for robots. His 2021 paper on "Cognitive Science as a Source of Forward and Inverse Models of Human Decisions" (4 citations) provides a foundational framework for integrating cognitive science principles into robotics and control systems. Through his research, Griffiths is fundamentally reshaping how robots understand and collaborate with humans in real-world applications.