A new training strategy with progressive random gate activation for variational quantum algorithms greatly mitigates barren plateaus and avoids the local minima while being quantum resource efficient and noise resilient.
No takes yet. Share an insight, caveat, or question.
Liu et al. (2023) studied this question.