Bernd Rosenow

Anyonic exchange phases can be isolated with just a few-percent systematic error using an open quantum dot architecture, sidestepping the delicate gate-tuning instabilities that routinely compromise fractional quantum Hall interferometry.

Softmax classification's slow α^{-1/3} generalization scaling isn't just about spectral properties, but also stems from a boundary-layer effect where only data near decision boundaries drive learning.