My research focuses on the intersection between how humans and machines form beliefs. While machine learning models are now commonplace, their exact nature is far from understood. We are enjoying its benefits yet less attention has been paid to its limitations. My research aims to contribute to understanding the interaction better and contribute by leveraging my cross-domain expertise. 

I am a Ph.D. candidate at the Erasmus Institute for Philosophy and Economics (EIPE), Erasmus School of Philosophy (ESPhil) at the Erasmus University Rotterdam.

Navigating Feedback Systems: Bad News for (Objective) Bayesians

(Objective) Bayesian agents are susceptible to overconfidence in their beliefs under specific circumstances. Agents forming beliefs in feedback context may assign overly confident credences to their beliefs. I demonstrate the importance of considering feedback in systems, and that no agent can have definitive knowledge of whether feedback is present. Furthermore, I explore one potential solution for counteracting overconfidence and for forming robust diachronic beliefs in a feedback context. The solution employs the exploration-exploitation concept from machine learning. Applying exploration-exploitation in a feedback context, allows the agent to improve the robustness of their beliefs over time.

Presentations

26th of June: Pre-read Workshop: The Many Faces of Philosophy of Economics