Exponential Stochastic Cellular Automata for Massively Parallel Inference
AISTATS, pp. 966-975, 2016.
We propose an embarrassingly parallel, memory ecient inference algorithm for latent variable models in which the complete data likelihood is in the exponential family. The algorithm is a stochastic cellular automaton and converges to a valid maximum a posteriori fixed point. Applied to latent Dirichlet allocation we find that our algorith...More
PPT (Upload PPT)