Simon Holk

I'm a phd student working in the social robotics team at the RPL department under Iolanda Leite.

My phd revolves around Reinforcement learning from human feedback (RLHF) to adapt learning agents to humans. My work involves preference-based learning, LLM, activate learning, unsupervised learning and how it can all be used to improve the efficiency of human feedback.

• Received 3 nominations for POLITE at ICRA'24. Best human-robot interaction paper, best student paper, and best conference paper.
• 2 papers accepted at ICRA'24, related to enhancing preferences with granual information from humans and semi-supervised query synthesis of human preferences.
• 1 paper presented at the technical track of HRI'24, related to learning from preferences with LLMs (large language models).
• 1 paper presented at IROS'23 in Detroit, Michigan.
• 1 paper presented at ICRA'23 in London, United Kingdom.