The Societal Impact of AI Decisions

Exploring how robot decision-making drives human behavior

Algorithms for Human-Aligned Decision Making

Designing algorithms that adapt AI behavior to human values

In this line of work, I develop algorithms that enable agents and robots to learn human values during collaboration. On the technical side, I draw from reinforcement learning, deep learning, inverse reinforcement learning, and large language models to build systems that can capture and reason about human values. On the applied side, I test these approaches in interactive settings with robots and AI systems, from assistive physical robots to multi-agent games and language-based collaborators.

Publications:

• Claure, H.*, Narcomey, A.*, Vázquez M. 2025. Fairness Theory: Modeling Reduced Welfare in Human-Robot Interactions. ACM/IEEE International Conference on Human- Robot Interaction (HRI). In Preparation.
• Narcomey, A., Claure, H., Vázquez M. 2025. Robot Delivery of Actionable Counterfactual Neural Network Explanations: Results in Group Perception. IEEE International Conference on Robot and Human Interactive Communication (RO-MAN).
• Claure, H., Candon, K., Vázquez, M. 2024. Dynamic Fairness Perceptions in Human- Robot Interaction. ACM Transactions on Human Robot Interactions (THRI). Under Review.
• Claure, H., Chen, Y., Modi, J., Jung, M., Nikolaidis, S. 2019. Reinforcement Learning with Fairness Constraints for Resource Distribution in Human-Robot Teams. ACM/IEEE International Conference on Human-Robot Interaction (HRI) '20.

Interactive Hardware Design

Building and Studying Hardware Designs on Human Experiences

In these projects I developed different hardware and software platforms as well as explored different aspects of human interaction with technology. For instance, the project image below was a collaboration with Dr. Robert Shepherd's Organic Robotics Laboratory (ORL) at Cornell University and NVIDIA. The goal was to develop a novel type of virtual reality controller sleeve that is capable of providing haptic feedback. We developed and tested a 12 DOF fluidically pressurized soft actuator for persistent and kinesthetic haptic sensations. We tested the controller through a user experience test after the development of a software interface with NVIDIA's game VR Funhhouse. Results demonstrated which haptic stimulations from the controller were more recognizable by human participants as well as how the controller sleeve influenced the game experience.

Publications:

• Claure, H.*, Barrieros, J.*, Peele, B.*, Shepherd, R. 2018. Fluid Elastic Actuators for Haptic Interactions in Virtual Reality. IEE Robotics and Automation Letters. (*First Authors).
• Lee, W. Y., Sakashita, M., Ricci, E., Claure, H., Guimbretière, F., & Jung, M. 2021. Interactive vignettes: Enabling large-scale interactive HRI research. 2021 30th IEEE international conference on robot & human interactive communication (RO-MAN).
• Shen, S., Tennent, H., Claure, H., & Jung, M. 2018. My Telepresence, My Culture? An Intercultural Investigation of Telepresence Robot Operators' Interpersonal Distance Behaviors. Proceedings of the 2018 CHI conference on human factors in computing systems (CHI).
• B.Stogin, L. Gockowski, H. Feldstein, H. Claure, J. Wang, and T. Wong. 2018. Free-standing liquid membranes as unusual particle separators. Science Advances.