Skip to yearly menu bar Skip to main content


Poster

Scalable Model Selection for Belief Networks

Zhao Song · Yusuke Muraoka · Ryohei Fujimaki · Lawrence Carin

Pacific Ballroom #34

Keywords: [ Variational Inference ] [ Model Selection and Structure Learning ] [ Graphical Models ] [ Generative Models ] [ Latent Variable Models ]


Abstract: We propose a scalable algorithm for model selection in sigmoid belief networks (SBNs), based on the factorized asymptotic Bayesian (FAB) framework. We derive the corresponding generalized factorized information criterion (gFIC) for the SBN, which is proven to be statistically consistent with the marginal log-likelihood. To capture the dependencies within hidden variables in SBNs, a recognition network is employed to model the variational distribution. The resulting algorithm, which we call FABIA, can simultaneously execute both model selection and inference by maximizing the lower bound of gFIC. On both synthetic and real data, our experiments suggest that FABIA, when compared to state-of-the-art algorithms for learning SBNs, $(i)$ produces a more concise model, thus enabling faster testing; $(ii)$ improves predictive performance; $(iii)$ accelerates convergence; and $(iv)$ prevents overfitting.

Live content is unavailable. Log in and register to view live content