Any selective breeding program operates similarly to this, and iterated embryo selection operates even more like this; if you consider the full life cycle from ethnicity to individual to assortative mating to embryo selection, much of the logic carries through (and considering such scenarios where there are multiple ‘stages’ is what made me appreciate the asymptotic advantage). Particle filters/swarm optimization algorithms operate similarly, in a sense, and evolutionary computation methods like CMA-ES operate pretty much exactly like this. This can also be a good intuition pump for pipeline scenarios like Scannell’s drug discovery pipeline model.
Background: https://www.gwern.net/Embryo-selection#multi-stage-selection
Any selective breeding program operates similarly to this, and iterated embryo selection operates even more like this; if you consider the full life cycle from ethnicity to individual to assortative mating to embryo selection, much of the logic carries through (and considering such scenarios where there are multiple ‘stages’ is what made me appreciate the asymptotic advantage). Particle filters/swarm optimization algorithms operate similarly, in a sense, and evolutionary computation methods like CMA-ES operate pretty much exactly like this. This can also be a good intuition pump for pipeline scenarios like Scannell’s drug discovery pipeline model.