EnLLVM – Fast Approximate Bayesian Inference

Approximate Bayesian inference that only requires a small number of samples and can be applied to high-dimensional non-linear systems with non-Gaussian uncertainties. It leverages the machinery behind the linear latent variable models which provide significant computational speed-up. Works with intractable likelihood functions and the model can be absorbed implicitly and reflected in the uncertainty of the quantities of interest.

  • X. Lin and G. Terejanu, “Fast Approximate Data Assimilation for High-Dimensional Problems,” https://arxiv.org/abs/1708.02340, 2017.
    [BibTeX] [Download PDF]
    @article{LinJ_TAC_2017,
    author = {Xiao Lin and Gabriel Terejanu},
    title = {{Fast Approximate Data Assimilation for High-Dimensional Problems}},
    journal = {{https://arxiv.org/abs/1708.02340}},
    url = {https://arxiv.org/abs/1708.02340},
    year = {2017},
    }