[mlthoughts2018]

The paper is from 2019. I think the article title should be updated to reflect that, since this link is nothing but rehashing the original algorithm, despite minor embellished claims about “taking it further.”

- https://arxiv.org/abs/1903.03129