Bounding the expected run-time of nonconvex optimization with early stopping

Re-assign community