2026
Sharpness-Aware Minimization Can Hallucinate Minimizers
ICML 2026poster
Sharpness-Aware Minimization (SAM) is widely used to seek flatter minima—often linked to better generalization. In its standard implementation, SAM updates the current iterate using the loss gradient evaluated at a point perturbed by distance $\rho$ along the normalized gradient direction. We show t…