Particle Sketches, a way to estimate a very large number of Bayesian models in sub-linear space and time
I have been designing an interesting, novel (as far as I know) data structure. Its goal is to estimate a very large number Bayesian models in the smallest possible space [...]
