1 problem

0 votes0 replies0 views
A diffusion-powered escape conjecture for SGD without replacement
The setting concerns stochastic gradient descent (SGD) without replacement, in which the data batches used at successive steps are dependent because they are disjoint, and saddle e…
Conjectureopennonconvex-optimization saddle-point-escape stochastic-gradient-descentarXiv posted 2 days ago

A diffusion-powered escape conjecture for SGD without replacement