Factored Filtering of Continuous-Time Systems

We consider filtering for a continuous-time, or asynchronous, stochastic system where the full distribution over states is too large to be stored or calculated. We assume that the rate matrix of the system can be compactly represented and that the belief distribution is to be approximated as a product of marginals. The essential computation is the matrix exponential. We look at two different methods for its computation: ODE integration and uniformization of the Taylor expansion. For both we consider approximations in which only a factored belief state is maintained. For factored uniformization we demonstrate that the KL-divergence of the filtering is bounded. Our experimental results confirm our factored uniformization performs better than previously suggested uniformization methods and the mean field algorithm.

[1]  Daphne Koller,et al.  Expectation Maximization and Complex Duration Distributions for Continuous Time Bayesian Networks , 2005, UAI.

[2]  Jing Xu,et al.  Importance Sampling for Continuous Time Bayesian Networks , 2010, J. Mach. Learn. Res..

[3]  Christel Baier,et al.  Model-Checking Algorithms for Continuous-Time Markov Chains , 2002, IEEE Trans. Software Eng..

[4]  C. Petri.  Kommunikation mit Automaten , 1962 .

[5]  Thomas M. Cover,et al.  Elements of Information Theory , 2005 .

[6]  Nir Friedman,et al.  Mean Field Variational Approximation for Continuous-Time Bayesian Networks , 2009, J. Mach. Learn. Res..

[7]  Gianfranco Ciardo,et al.  Approximate steady-state analysis of large Markov models based on the structure of their decision diagram encoding , 2011, Perform. Evaluation.

[8]  Xavier Boyen,et al.  Tractable Inference for Complex Stochastic Processes , 1998, UAI.

[9]  Michael I. Jordan,et al.  Probabilistic Inference in Queueing Networks , 2008, SysML.

[10]  Cleve B. Moler,et al.  Nineteen Dubious Ways to Compute the Exponential of a Matrix, Twenty-Five Years Later , 1978, SIAM Rev..

[11]  Daphne Koller,et al.  Continuous Time Bayesian Networks , 2012, UAI.

[12]  Xavier Boyen,et al.  Exploiting the Architecture of Dynamic Systems , 1999, AAAI/IAAI.

[13]  Nir Friedman,et al.  Continuous-Time Belief Propagation , 2010, ICML.

[14]  Roger B. Sidje,et al.  Inexact Uniformization Method for Computing Transient Distributions of Markov Chains , 2007, SIAM J. Sci. Comput..

[15]  Yu Fan,et al.  Learning Continuous-Time Social Network Dynamics , 2009, UAI.

[16]  Jing Xu,et al.  Intrusion Detection using Continuous Time Bayesian Networks , 2010, J. Artif. Intell. Res..

[17]  Avi Pfeffer,et al.  Continuous Time Particle Filtering , 2005, IJCAI.