Variational Infinite Hidden Conditional Random Fields

TY - JOUR

T1 - Variational Infinite Hidden Conditional Random Fields

AU - Bousmalis, Konstantinos

AU - Zafeiriou, Stefanos

AU - Morency, Louis-Philippe

AU - Pantic, Maja

AU - Ghahramani, Zoubin

N1 - eemcs-eprint-26757

PY - 2015/9/1

Y1 - 2015/9/1

N2 - Hidden conditional random fields (HCRFs) are discriminative latent variable models which have been shown to successfully learn the hidden structure of a given classification problem. An Infinite hidden conditional random field is a hidden conditional random field with a countably infinite number of hidden states, which rids us not only of the necessity to specify a priori a fixed number of hidden states available but also of the problem of overfitting. Markov chain Monte Carlo (MCMC) sampling algorithms are often employed for inference in such models. However, convergence of such algorithms is rather difficult to verify, and as the complexity of the task at hand increases the computational cost of such algorithms often becomes prohibitive. These limitations can be overcome by variational techniques. In this paper, we present a generalized framework for infinite HCRF models, and a novel variational inference approach on a model based on coupled Dirichlet Process Mixtures, the HCRF-DPM. We show that the variational HCRF-DPM is able to converge to a correct number of represented hidden states, and performs as well as the best parametric HCRFs—chosen via cross-validation—for the difficult tasks of recognizing instances of agreement, disagreement, and pain in audiovisual sequences.

AB - Hidden conditional random fields (HCRFs) are discriminative latent variable models which have been shown to successfully learn the hidden structure of a given classification problem. An Infinite hidden conditional random field is a hidden conditional random field with a countably infinite number of hidden states, which rids us not only of the necessity to specify a priori a fixed number of hidden states available but also of the problem of overfitting. Markov chain Monte Carlo (MCMC) sampling algorithms are often employed for inference in such models. However, convergence of such algorithms is rather difficult to verify, and as the complexity of the task at hand increases the computational cost of such algorithms often becomes prohibitive. These limitations can be overcome by variational techniques. In this paper, we present a generalized framework for infinite HCRF models, and a novel variational inference approach on a model based on coupled Dirichlet Process Mixtures, the HCRF-DPM. We show that the variational HCRF-DPM is able to converge to a correct number of represented hidden states, and performs as well as the best parametric HCRFs—chosen via cross-validation—for the difficult tasks of recognizing instances of agreement, disagreement, and pain in audiovisual sequences.

KW - HMI-HF: Human Factors

KW - variational inference

KW - hidden conditional random fields

KW - dirichlet processes

KW - EC Grant Agreement nr.: FP7/611153

KW - IR-99338

KW - EWI-26757

KW - Discriminative models

KW - Nonparametric models

KW - METIS-315570

KW - EC Grant Agreement nr.: FP7/2007-2013

U2 - 10.1109/TPAMI.2014.2388228

DO - 10.1109/TPAMI.2014.2388228

M3 - Article

VL - 37

SP - 1917

EP - 1929

JO - IEEE transactions on pattern analysis and machine intelligence

JF - IEEE transactions on pattern analysis and machine intelligence

SN - 0162-8828

IS - 9

ER -