Gaussian Lifted Marginal Filtering

Recently, Lifted Marginal Filtering [5] has been proposed, an approach for efficient probabilistic inference in systems with multiple, (inter-)acting agents and objects (entities). The algorithm achieves its efficiency by performing inference jointly over groups of similar entities (i.e. their properties follow the same distribution). In this paper, we explore the case where there are no entities that are directly suitable for grouping. We propose to use methods from Gaussian mixture fitting and merging to identify entity groups and keep the number of groups constant over time. Empirical results suggest that decrease in prediction accuracy is small, while the algorithm runtime decreases significantly.

[1]  T. Kirste,et al.  Computational State Space Models for Activity and Intention Recognition. A Feasibility Study , 2014, PloS one.

[2]  Kristian Kersting,et al.  Mixed Sum-Product Networks: A Deep Architecture for Hybrid Domains , 2018, AAAI.

[3]  Adrian E. Raftery,et al.  Model-Based Clustering, Discriminant Analysis, and Density Estimation , 2002 .

[4]  Nando de Freitas,et al.  Rao-Blackwellised Particle Filtering for Dynamic Bayesian Networks , 2000, UAI.

[5]  Hector Geffner,et al.  Goal Recognition over POMDPs: Inferring the Intention of a POMDP Agent , 2011, IJCAI.

[6]  Luc De Raedt,et al.  Stochastic relational processes: Efficient inference and applications , 2011, Machine Learning.

[7]  Paolo Milazzo,et al.  Maximally Parallel Probabilistic Semantics for Multiset Rewriting , 2011, Fundam. Informaticae.

[8]  Pedro M. Domingos,et al.  Probabilistic theorem proving , 2011, UAI.

[9]  H. Sorenson,et al.  Nonlinear Bayesian estimation using Gaussian sum approximations , 1972 .

[10]  Hazel R. Parry,et al.  A comparative analysis of parallel processing and super-individual methods for improving the computational performance of a large individual-based model , 2008 .

[11]  Marten Scheffer,et al.  Super-individuals a simple solution for modelling large populations on an individual basis , 1995 .

[12]  Thomas Kirste,et al.  Approximate Probabilistic Parallel Multiset Rewriting Using MCMC , 2018, KI.

[13]  A.R. Runnalls,et al.  A Kullback-Leibler Approach to Gaussian Mixture Reduction , 2007 .

[14]  Luca Scrucca,et al.  mclust 5: Clustering, Classification and Density Estimation Using Gaussian Finite Mixture Models , 2016, R J..

[15]  Kristian Kersting,et al.  Lifted Filtering via Exchangeable Decomposition , 2018, IJCAI.