Bayesian Sparse Unsupervised Learning for Probit Models of Binary Data
Department of Statistics, Center for Theoretical Neuroscience and Grossman Center for the Statistics of Mind, Columbia University, New York
Columbia University, 2014
@article{pakman2014bayesian,
title={Bayesian Sparse Unsupervised Learning for Probit Models of Binary Data},
author={Pakman, Ari and Shababo, Ben and Paninski, Liam},
year={2014}
}
We present a unified approach to unsupervised Bayesian learning of factor models for binary data with binary and spike-and-slab latent factors. We introduce a non-negative constraint in the spike-and-slab prior that eliminates the usual sign ambiguity present in factor models and lowers the generalization error on the datasets tested here. For the generative models we use probit functions, which can be sampled without tuning parameters, unlike previous works that used logistic functions. The posterior distributions involve mixtures of binary and truncated Gaussian variables, for which we present exact Hamiltonian Monte Carlo samplers and compare their properties to Gibbs samplers.
April 9, 2014 by hgpu
