Exponential Stochastic Cellular Automata for Massively Parallel Inference

    AISTATS, pp. 966-975, 2016.

    Cited by: 18|Bibtex|Views56|Links
    EI

    Abstract:

    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

    Code:

    Data:

    Your rating :
    0

     

    Tags
    Comments