Faster Verified Explanations for Neural Networks