Uncertainty in Artificial Intelligence
First Name   Last Name   Password   Forgot Password   Log in!
    Proceedings         Authors   Author's Info   Article details         Search    
Learning Convex Inference of Marginals
Justin Domke
Abstract:
Graphical models trained using maximum likelihood are a common tool for probabilistic inference of marginal distributions. However, this approach suffers difficulties when either the inference process or the model is approximate. In this paper, the inference process is first defined to be the minimization of a convex function, inspired by free energy approximations. Learning is then done directly in terms of the performance of the inference process at univariate marginal prediction. The main novelty is that this is a direct minimization of emperical risk, where the risk measures the accuracy of predicted marginals.
Keywords:
Pages: 137-144
PS Link:
PDF Link: /papers/08/p137-domke.pdf
BibTex:
@INPROCEEDINGS{Domke08,
AUTHOR = "Justin Domke ",
TITLE = "Learning Convex Inference of Marginals",
BOOKTITLE = "Proceedings of the Twenty-Fourth Conference Annual Conference on Uncertainty in Artificial Intelligence (UAI-08)",
PUBLISHER = "AUAI Press",
ADDRESS = "Corvallis, Oregon",
YEAR = "2008",
PAGES = "137--144"
}


hosted by DSL   •   site info   •   help