||September 10, 2012
||Dr. Ryohei Fujimaki, NEC Laboratories America
||Factorized Asymptotic Bayesian Hidden Markov Models
This talk presents a new model selection method for hidden Markov
models (HMMs), using factorized asymptotic Bayesian inference (FAB).
FAB for HMMs is derived as an iterative lower bound maximization
algorithm of a factorized information criterion (FIC), and has several
desirable properties for learning HMMs, such as asymptotic consistency
of FIC with marginal log-likelihood, a shrinkage effect for hidden state
selection, monotonic increase of the lower FIC bound through the iterative
optimization. Further, it does not have a tunable hyper-parameter,
and thus its model selection process can be fully automated.
Experimental results shows that FAB outperforms states-of-the-art
variational Bayesian HMM and non-parametric Bayesian HMM in
terms of model selection accuracy and computational efficiency.