funnily enough, that's the name of their recent recommendation algorithm paper

https://arxiv.org/pdf/2402.17152