Parallel subspace sampling for particle filtering in dynamic bayesian networks

Thumbnail Image
Full text at PDC
Publication Date
Advisors (or tutors)
Journal Title
Journal ISSN
Volume Title
Springer-Verlag Berlin
Google Scholar
Research Projects
Organizational Units
Journal Issue
Monitoring the variables of real world dynamic systems is a difficult task due to their inherent complexity and uncertainty. Particle Filters (PF) perform that task, yielding probability distribution over the unobserved variables. However, they suffer from the Curse of dimensionality problem: the number of particles grows exponentially with the dimensionality of the hidden state space. The problem is aggravated when the initial distribution of the variables is not well known, as happens in global localization problems. We present a new parallel PF for systems whose variable dependencies can be factored into a Dynamic Bayesian Network. The new algorithms significantly reduce the number of particles, while independently exploring different subspaces of hidden variables to build particles consistent with past history and measurements. We demonstrate this new PF approach on some complex dynamical system estimation problems, showing that our method successfully localizes and tracks hidden states in cases where traditional PFs fail.
© Springer-Verlag Berlin Heidelberg 2009. This work was supported by the Spanish Grants DPI2006-15661-C02-01 and CAM S-0505/DPI 0391. Further, Dr. Besada-Portas was supported by the Spanish post-doctoral Grant EX-2007-0915, Dr. Plis by NIMH Grant 1 R01 MH076282-01, and Dr. Lane by NSF Grant IIS-0705681. The authors also thank the Aula Sun-UCM for providing access to their computational resources for doing parts of the experiments. Joint European Conference on Machine Learnin(ECML)/European Conference on Principles and Practice of Knowledge Discovery in Databases (PKDD)(2009. Bled, ESLOVENIA)
Unesco subjects
1. Doucet, A., Freitas, N., Gordon, N. (eds.): Sequential Monte Carlo methods in practice. Springer, Heidelberg (2001) 2. MacCormick, J., Isard, M.: Partitioned sampling, articulated objects, and interface-quality hand tracking. In: Vernon, D. (ed.) ECCV 2000. LNCS, vol. 1843, pp. 3–19. Springer, Heidelberg (2000) 3. Doucet, A., de Freitas, N., Murphy, K., Russell, S.: Rao-blackwellised particle filtering for dynamic bayesian networks. In: 16th Conference on Uncertainty in Artificial Intelligence, pp. 176–183 (2000) 4. Vaswani, N.: Particle filters for infinite (or large) dimensional state spaces-part 2. In: IEEE ICASSP (2006)5. Ng, B., Peshkin, L.: Factored particles for scalable monitoring. In: Proceedings of the 18th Conference on Uncertainty in Artificial Intelligence, pp. 370–377. Morgan Kaufmann, San Francisco (2002) 6. Das, S., Lawless, D., Ng, B., Pfeffer, A.: Factored particle filtering for data fusion and situation assessments in urban environments. In: 8th International Conference on Information Fusion (July 2005) 7. Park, S., Hwang, J., Rou, K., Kim, E.: A new particle filter inspired by biological evolution: Genetic Filter. Proceeding of World Academy of Science, Engineering and Technology 21, 57–71 (2007) 8. Brandao, B.C., Wainer, J., Goldenstein, S.K.: Subspace hierarchical particle filter. In: XIX Brazilian Symposium on Computer Graphics and Image Processing (2006) 9. Koller, D., Lerner, U.: Sampling in factored dynamic systems. In: Sequential Monte Carlo in Practice, pp. 445–464 (2001) 146 E. Besada-Portas et al. 10. Maccormick, J., Blake, A.: A probabilistic exclusion principle for tracking multiple objects. Intenational Journal of Computer Vision 39(1), 57–71 (2000) 11. Klaas,M., de Freitas, N., Doucet, A.: Toward practical n2 Monte Carlo: theMarginal Particle Filter. In: Proceedings of UAI 2005, Arlington, Virginia, pp. 308–331. AUAI Press (2005) 12. Rose, C., Saboune, J., Charpillet, F.: Reducing particle filtering complexity for 3D motion capture using dynamic bayesian networks. In: 23th AAAI Conference on Artificial Intelligence (2008) 13. Pitt, M.K., Shephard, N.: Filtering via simulation: Auxiliary Particle Filters. Journal of the American Statistical Association 94(446), 590–599 (1999) 14. Burge, J., Lane, T., Link, H., Qiu, S., Clark, V.P.P.: Discrete dynamic bayesian network analysis of fMRI data. Human Brain Mapping (November 2007) 15. Friston, K.J., Harrison, L., Penny,W.: Dynamic CausalModelling. NeuroImage 19(4), 1273–1302 (2003) 16. Lancaster, J.L., Woldorff, M.G., Parsons, L.M., Liotti, M., Freitas, C.S., Rainey, L., Kochunov, P.V., Nickerson, D., Mikiten, S.A., Fox, P.T.: Automated Talairach atlas labels for functional brain mapping. Human Brain Mapping 10(3), 120–131 (2000) 17. Hofmann, R., Tresp, V.: Discovering structure in continuous variables using bayesian networks. In: Touretzky, D.S., Mozer, M.C., Hasselmo, M.E. (eds.) Advances in Neural Information Processing Systems, vol. 8, pp. 500–506. MIT Press, Cambridge (1996)