arXiv stat.ML
· Papers
Avoiding unsafe sets when training with Langevin Dynamics
arXiv:2607.07538v2 Announce Type: replace-cross Abstract: Training a model with noisy gradient descent can be idealized as overdamped Langevin dynamics, and a natural safety question is to bound the probability $nu_t(mathcal{A}_H) = mathbb{P}(Q_t in mathcal{A}_H)$ that the trajectory lies in a designated failure r