uai2026 in Amsterdam, Netherlands

Research

August 17, 2026 - August 21, 2026

From August 17–21, 2026, I had the opportunity to attend the 42nd Conference on Uncertainty in Artificial Intelligence (uai2026), held in Amsterdam, Netherlands. I presented my publication, “Achieving Alignment Through Adaptive Play: Helping Optimize Objectives Without Observing Them,” during a poster session on Tuesday, August 18, 2026, as part of the session covering two topics; Reinforcement Learning and Dynamics and Voting, Choice, and Games.

It was great to speak with other researchers in the field, share my work, and gain valuable feedback and new perspectives on my research. I was especially happy to receive such positive and encouraging comments throughout my poster session.

I would like to thank my advisors, Samuel Burden and Lillian Ratliff, for their continued support and for giving me the opportunity to travel to Amsterdam and attend UAI 2026.

One notable topic from the conference was the Workshop on Safe AI. This workshop focused on safety challenges in agentic AI, particularly systems that can autonomously act, adapt, and interact over extended periods of time. The talks and discussions covered topics including robustness, interpretability, foundations of AI safety, and the challenges associated with deploying increasingly autonomous AI systems. I found these discussions especially interesting because they also highlighted some of the important government and policy considerations surrounding this field.

I particularly enjoyed Peter Mattson’s talk from ML Commons and his comparison between the development of AI systems and the aerospace industry. He described a progression through three stages separated by two important barriers. First, "ideas that don't quit work" must cross a capability barrier to become great research. From there, great research must cross a reliability barrier before it can successfully become products and services used in the real world. I found this framework especially interesting because it emphasizes that demonstrating capability is only one part of developing a technology; ensuring that it is sufficiently reliable for widespread deployment is an equally important challenge. He used the airline industry as an example of a field that successfully developed strong reliability practices, while contrasting it with nuclear energy as an example of how public trust and perceptions of risk can significantly affect the adoption of a technology.

UAIPoster.pdf

Beyond the conference, I really enjoyed exploring and learning about such a historic and vibrant city. Spending time in another country, even for a short period, is always an eye-opening experience and a great opportunity to experience a different culture, way of life, and perspective.

During my time in Amsterdam, I visited numerous museums, landmarks, and attractions, including the Royal Palace Amsterdam, Van Gogh Museum, Rijksmuseum, Stedelijk Museum, Moco Museum, STRAAT Museum, Grachten Museum, Anne Frank House, Heineken Experience, ARTIS Zoo, Zaanse Schans Windmills, Albert Cuyp Market, and the Amsterdam Flower Market.

When I started my PhD, my advisors encouraged me to take advantage of opportunities to submit my research to conferences in locations that interested me. I especially remember one piece of advice from one of my advisors: “Use your PhD to explore the world.” I am incredibly grateful that my first conference publication gave me the opportunity to travel to Amsterdam, the Netherlands, and experience such a beautiful city. Amsterdam is well known for its extensive network of canals, but I was also struck by its wide variety of museums, rich history, distinctive architecture, and strong cycling culture.

Working in the fields of AI, human-machine interaction, and multi-agent systems, traveling to new places can also provide important context for thinking about how our research may eventually be applied in the real world. Before beginning my current research, I worked on autonomous and safe vehicles, and one thing that stood out to me during my time in Amsterdam was how dramatically different the street environment was compared with many large U.S. cities. The traffic lights, signs, lane layouts, interactions between bicycles, pedestrians, and vehicles, and the overall transportation culture were all noticeably different.

Experiencing these differences made me think about how the deployment of autonomous vehicles, and intelligent systems more broadly, may need to vary significantly across different cities and cultures. These are small examples of environmental and human factors that can be difficult to fully appreciate without experiencing them firsthand. It reinforced for me how valuable it is as a researcher to step outside of familiar environments and see how people interact with technology and infrastructure in different parts of the world.

For that reason, I am especially grateful that attending UAI 2026 gave me not only the opportunity to present and discuss my research, but also the time to explore and experience the city of Amsterdam.

Page updated

Google Sites

Report abuse