Nicholas C. Landolfi

Papers

Total Citations

H-Index

About

Nicholas C. Landolfi is a robotics and machine learning researcher whose work centers on reward learning, human-robot interaction, and the challenge of teaching autonomous systems to understand human preferences. His research addresses one of the most fundamental problems in robot learning: how to efficiently and accurately infer what humans want without burdening them with difficult or unintuitive queries. Landolfi's most cited contribution, "Asking Easy Questions" (2019, 54 citations), introduced a human-centered approach to active reward learning that balances information gain with the cognitive ease of questions posed to human teachers — a meaningful departure from purely uncertainty-driven methods. This work reflects his broader commitment to making robot learning practical and accessible for real users. Alongside collaborators, Landolfi has also advanced methods for integrating diverse human feedback signals. His research on combining demonstrations and preferences (2021, 21 citations; 2019, 18 citations) demonstrates how passive and active data collection can be optimally unified, improving the sample efficiency and accuracy of learned reward functions. Together, these contributions have helped shape the growing field of reward learning from human feedback, with implications for safer and more aligned autonomous systems.