Max Robotics — Global Service Robot Directory + US Certification

About

Aditi Talati is a researcher working at the intersection of human-robot interaction, robot learning, and collaborative autonomy. Her work centers on two critical challenges: enabling robots to realistically model human behavior and developing systems that allow robots to learn what humans truly want. Talati's most influential contribution, "When Humans Aren't Optimal" (2020, 53 citations), challenges a pervasive assumption in robotics — that humans always behave rationally. By incorporating risk-awareness into human behavior models, her research enables robots to collaborate more safely and effectively with real, imperfect human partners. This work represents a meaningful shift away from idealized assumptions toward psychologically grounded human modeling. Equally significant is her development of APReL, an open-source library for active preference-based reward learning (2022, 22 citations). APReL consolidates a fragmented landscape of reward learning algorithms into a unified, accessible framework, lowering barriers for researchers working on value alignment in robotics — ensuring robots act in accordance with genuine human preferences rather than proxy objectives. With over 80 cumulative citations, Talati's research addresses foundational questions about trust, alignment, and collaboration between humans and autonomous systems, making her work essential reading for anyone exploring the future of human-centered robotics.