Simon-Shlomo Poil

I work on one question: how to understand the state of a complex system when its most important properties are not directly visible.

I studied biophysics, and my research in computational neuroscience at VU University Amsterdam asked whether neuronal networks operate near a critical point, why brain rhythms carry correlations over minutes, and whether those correlations can serve as biomarkers of brain disorders.

Two lessons from that work stay with me: weak signals often become informative only when they are interpreted together, and a method becomes valuable at scale only when others can reuse it without breaking it. That is why we built the Neurophysiological Biomarker Toolbox (NBT). It is no longer maintained, but its core methods live on in the tutorials here.

Today the same question returns in AI: how do you know whether a deployed system is still healthy?

Contact: LinkedIn.