Pub Date : 2006-09-01DOI: 10.1109/NSSPW.2006.4378807
P. Fearnhead, Z. Liu
We review work on how to perform exact online inference for a class of multiple changepoint models. These models have a conditional independence structure, and require you to be able to integrate out (either analytically or numerically) the parameters associated within each segment. The computational cost per observation increases linearly with the number of observations. This algorithm is closely related to a particle filter algorithm, and we describe how efficient resampling algorithms can be used to produce an accurate particle filter for this class of models.
{"title":"Efficient Online Inference for Multiple Changepoint Problems","authors":"P. Fearnhead, Z. Liu","doi":"10.1109/NSSPW.2006.4378807","DOIUrl":"https://doi.org/10.1109/NSSPW.2006.4378807","url":null,"abstract":"We review work on how to perform exact online inference for a class of multiple changepoint models. These models have a conditional independence structure, and require you to be able to integrate out (either analytically or numerically) the parameters associated within each segment. The computational cost per observation increases linearly with the number of observations. This algorithm is closely related to a particle filter algorithm, and we describe how efficient resampling algorithms can be used to produce an accurate particle filter for this class of models.","PeriodicalId":388611,"journal":{"name":"2006 IEEE Nonlinear Statistical Signal Processing Workshop","volume":"82 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130397574","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2006-09-01DOI: 10.1109/NSSPW.2006.4378847
F. Gustafsson, T. Schon, R. Karlsson, P. Nordlund
The marginalized particle filter is a powerful combination of the particle filter and the Kalman filter, which can be used when the underlying model contains a linear sub-structure subject to Gaussian noise. This paper surveys state of the art for theory and practice.
{"title":"State-of-the-Art for the Marginalized Particle Filter","authors":"F. Gustafsson, T. Schon, R. Karlsson, P. Nordlund","doi":"10.1109/NSSPW.2006.4378847","DOIUrl":"https://doi.org/10.1109/NSSPW.2006.4378847","url":null,"abstract":"The marginalized particle filter is a powerful combination of the particle filter and the Kalman filter, which can be used when the underlying model contains a linear sub-structure subject to Gaussian noise. This paper surveys state of the art for theory and practice.","PeriodicalId":388611,"journal":{"name":"2006 IEEE Nonlinear Statistical Signal Processing Workshop","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124726589","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2006-09-01DOI: 10.1109/NSSPW.2006.4378853
G. Poyiadjis, Sumeetpal S. Singh, A. Doucet
This paper proposes novel particle methods for online parameter estimation for partially observed diffusions. We consider diffusions observed with error under a non-linear mapping and multivariate diffusions where only a subset of the components is observed. The proposed methods rely on the commonly used idea of data augmentation and are based on obtaining particle approximations to the derivatives of the optimal filter. The performance of our algorithms is assessed using several financial applications.
{"title":"Online Parameter Estimation for Partially Observed Diffusions","authors":"G. Poyiadjis, Sumeetpal S. Singh, A. Doucet","doi":"10.1109/NSSPW.2006.4378853","DOIUrl":"https://doi.org/10.1109/NSSPW.2006.4378853","url":null,"abstract":"This paper proposes novel particle methods for online parameter estimation for partially observed diffusions. We consider diffusions observed with error under a non-linear mapping and multivariate diffusions where only a subset of the components is observed. The proposed methods rely on the commonly used idea of data augmentation and are based on obtaining particle approximations to the derivatives of the optimal filter. The performance of our algorithms is assessed using several financial applications.","PeriodicalId":388611,"journal":{"name":"2006 IEEE Nonlinear Statistical Signal Processing Workshop","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114332534","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2006-09-01DOI: 10.1109/NSSPW.2006.4378822
J. Salmi, A. Richter, V. Koivunen
In this paper we describe the application of extracting the MIMO radio channel propagation parameters from channel sounding measurements using the Extended Kalman Filter. This approach allows to capture the dynamics of the radio propagation channels and enables recursive, computationally low-complexity (compared with traditional iterative maximum likelihood based methods) estimation of the parameters. We also discuss the selection of the state dimension, i.e., the appropriate number of propagation paths to track.
{"title":"MIMO Propagation Parameter Tracking using EKF","authors":"J. Salmi, A. Richter, V. Koivunen","doi":"10.1109/NSSPW.2006.4378822","DOIUrl":"https://doi.org/10.1109/NSSPW.2006.4378822","url":null,"abstract":"In this paper we describe the application of extracting the MIMO radio channel propagation parameters from channel sounding measurements using the Extended Kalman Filter. This approach allows to capture the dynamics of the radio propagation channels and enables recursive, computationally low-complexity (compared with traditional iterative maximum likelihood based methods) estimation of the parameters. We also discuss the selection of the state dimension, i.e., the appropriate number of propagation paths to track.","PeriodicalId":388611,"journal":{"name":"2006 IEEE Nonlinear Statistical Signal Processing Workshop","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125509245","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2006-09-01DOI: 10.1109/NSSPW.2006.4378833
W. Ng, Jack Li, S. K. Pang, S. Godsill
In this paper we propose an online tracking algorithm for multiple manoeuvring targets using variable rate particle filters (VRPFs). Unlike conventional particle filters, VRPFs combined with an intrinsic dynamical model enables us to track the manoeuvring behaviour of an object even though only a single dynamical model is employed. Furthermore a Markov Random Field motion model is included for modelling target interactions. In this paper we propose to integrate a data-dependent importance sampling method with the framework to generate more representative state particles. A Poisson observation model is also used to model both targets and clutter measurements, avoiding the data association difficulties associated with traditional tracking approaches. Finally computer simulations demonstrate the potential of the proposed method for tracking multiple highly manoeuvrable targets in a hostile environment with high clutter density and low detection probability.
{"title":"On Tracking Applications using Variable Rate Particle Filters","authors":"W. Ng, Jack Li, S. K. Pang, S. Godsill","doi":"10.1109/NSSPW.2006.4378833","DOIUrl":"https://doi.org/10.1109/NSSPW.2006.4378833","url":null,"abstract":"In this paper we propose an online tracking algorithm for multiple manoeuvring targets using variable rate particle filters (VRPFs). Unlike conventional particle filters, VRPFs combined with an intrinsic dynamical model enables us to track the manoeuvring behaviour of an object even though only a single dynamical model is employed. Furthermore a Markov Random Field motion model is included for modelling target interactions. In this paper we propose to integrate a data-dependent importance sampling method with the framework to generate more representative state particles. A Poisson observation model is also used to model both targets and clutter measurements, avoiding the data association difficulties associated with traditional tracking approaches. Finally computer simulations demonstrate the potential of the proposed method for tracking multiple highly manoeuvrable targets in a hostile environment with high clutter density and low detection probability.","PeriodicalId":388611,"journal":{"name":"2006 IEEE Nonlinear Statistical Signal Processing Workshop","volume":"78 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115204491","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2006-09-01DOI: 10.1109/NSSPW.2006.4378859
Edmund S. Jackson, W. Fitzgerald
A significant and stubbornly intractable problem in genome sequence analysis has been the de-novo identification of transcription factor binding sites in promoter regions. Probabilistic methods have faced difficulties from prior ignorance and poor models of the biological sequence. These problems result in inference in an extremely irregular, high dimensional space. We derive and demonstrate a novel method with improved convergence to the global mode utilising an iterated particle optimisation in place of the standard Gibbs sampling approach.
{"title":"A Sequential Monte Carlo EM Solution to the Transcription Factor Binding Site Identification Problem","authors":"Edmund S. Jackson, W. Fitzgerald","doi":"10.1109/NSSPW.2006.4378859","DOIUrl":"https://doi.org/10.1109/NSSPW.2006.4378859","url":null,"abstract":"A significant and stubbornly intractable problem in genome sequence analysis has been the de-novo identification of transcription factor binding sites in promoter regions. Probabilistic methods have faced difficulties from prior ignorance and poor models of the biological sequence. These problems result in inference in an extremely irregular, high dimensional space. We derive and demonstrate a novel method with improved convergence to the global mode utilising an iterated particle optimisation in place of the standard Gibbs sampling approach.","PeriodicalId":388611,"journal":{"name":"2006 IEEE Nonlinear Statistical Signal Processing Workshop","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116936476","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2006-09-01DOI: 10.1109/NSSPW.2006.4378836
Ihor Smal, W. Niessen, E. Meijering
Motion analysis of subcellular structures in living cells is currently a major topic in molecular cell biology, for which computerized methods are desperately needed. In this paper we adopt and tailor particle filtering techniques for this purpose and present the results of robust and accurate tracking of multiple objects in real fluorescence microscopy image data acquired for specific biological studies. Experimental results demonstrate that the automated method produces results comparable to manual tracking but using only a fraction of the manual tracking time.
{"title":"Particle Filtering for Multiple Object Tracking in Molecular Cell Biology","authors":"Ihor Smal, W. Niessen, E. Meijering","doi":"10.1109/NSSPW.2006.4378836","DOIUrl":"https://doi.org/10.1109/NSSPW.2006.4378836","url":null,"abstract":"Motion analysis of subcellular structures in living cells is currently a major topic in molecular cell biology, for which computerized methods are desperately needed. In this paper we adopt and tailor particle filtering techniques for this purpose and present the results of robust and accurate tracking of multiple objects in real fluorescence microscopy image data acquired for specific biological studies. Experimental results demonstrate that the automated method produces results comparable to manual tracking but using only a fraction of the manual tracking time.","PeriodicalId":388611,"journal":{"name":"2006 IEEE Nonlinear Statistical Signal Processing Workshop","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125055689","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2006-09-01DOI: 10.1109/NSSPW.2006.4378861
O. Zoeter, A. Ypma, T. Heskes
In this article we study inference problems in non-linear dynamical systems. In particular we are concerned with assumed density approaches to filtering and smoothing. In models with uncorrelated (but dependent) state and observation, the extended Kalman filter and the unscented Kalman filter break down. We show that the Gaussian particle filter and the one-step unscented Kalman filter make less assumptions and potentially form useful filters for this class of models. We construct a symmetric smoothing pass for both filters that does not require the dynamics to be invertible. We investigate the characteristics of the methods in an interesting problem from mathematical finance. Among others we find that smoothing helps, in particular for the deterministic one-step unscented Kalman filter.
{"title":"Deterministic and Stochastic Gaussian Particle Smoothing","authors":"O. Zoeter, A. Ypma, T. Heskes","doi":"10.1109/NSSPW.2006.4378861","DOIUrl":"https://doi.org/10.1109/NSSPW.2006.4378861","url":null,"abstract":"In this article we study inference problems in non-linear dynamical systems. In particular we are concerned with assumed density approaches to filtering and smoothing. In models with uncorrelated (but dependent) state and observation, the extended Kalman filter and the unscented Kalman filter break down. We show that the Gaussian particle filter and the one-step unscented Kalman filter make less assumptions and potentially form useful filters for this class of models. We construct a symmetric smoothing pass for both filters that does not require the dynamics to be invertible. We investigate the characteristics of the methods in an interesting problem from mathematical finance. Among others we find that smoothing helps, in particular for the deterministic one-step unscented Kalman filter.","PeriodicalId":388611,"journal":{"name":"2006 IEEE Nonlinear Statistical Signal Processing Workshop","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128914039","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2006-09-01DOI: 10.1109/NSSPW.2006.4378825
Byron M. Yu, K. Shenoy, M. Sahani
Neural activity unfolding over time can be modeled using non-linear dynamical systems [1]. As neurons communicate via discrete action potentials, their activity can be characterized by the numbers of events occurring within short pre-defined time-bins (spike counts). Because the observed data are high-dimensional vectors of non-negative integers, non-linear state estimation from spike counts presents a unique set of challenges. In this paper, we describe why the expectation propagation (EP) framework is particularly well-suited to this problem. We then demonstrate ways to improve the robustness and accuracy of Gaussian quadrature-based EP. Compared to the unscented Kalman smoother, we find that EP-based state estimators provide more accurate state estimates.
{"title":"Expectation Propagation for Inference in Non-Linear Dynamical Models with Poisson Observations","authors":"Byron M. Yu, K. Shenoy, M. Sahani","doi":"10.1109/NSSPW.2006.4378825","DOIUrl":"https://doi.org/10.1109/NSSPW.2006.4378825","url":null,"abstract":"Neural activity unfolding over time can be modeled using non-linear dynamical systems [1]. As neurons communicate via discrete action potentials, their activity can be characterized by the numbers of events occurring within short pre-defined time-bins (spike counts). Because the observed data are high-dimensional vectors of non-negative integers, non-linear state estimation from spike counts presents a unique set of challenges. In this paper, we describe why the expectation propagation (EP) framework is particularly well-suited to this problem. We then demonstrate ways to improve the robustness and accuracy of Gaussian quadrature-based EP. Compared to the unscented Kalman smoother, we find that EP-based state estimators provide more accurate state estimates.","PeriodicalId":388611,"journal":{"name":"2006 IEEE Nonlinear Statistical Signal Processing Workshop","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130142309","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2006-09-01DOI: 10.1109/NSSPW.2006.4378858
K. Kalogeropoulos, G. Roberts, P. Dellaportas
This paper presents a Markov chain Monte Carlo algorithm suitable for a class of partially observed non-linear diffusions. This class is of high practical interest; it includes for instance stochastic volatility models. We use data augmentation, treating the unobserved paths as missing data. However, unless these paths are transformed, the algorithm becomes reducible. We circumvent the problem by introducing appropriate reparametrisations of the likelihood that can be used to construct irreducible data augmentation schemes.
{"title":"Irreducible Markov Chain Monte Carlo Schemes for Partially Observed Diffusions","authors":"K. Kalogeropoulos, G. Roberts, P. Dellaportas","doi":"10.1109/NSSPW.2006.4378858","DOIUrl":"https://doi.org/10.1109/NSSPW.2006.4378858","url":null,"abstract":"This paper presents a Markov chain Monte Carlo algorithm suitable for a class of partially observed non-linear diffusions. This class is of high practical interest; it includes for instance stochastic volatility models. We use data augmentation, treating the unobserved paths as missing data. However, unless these paths are transformed, the algorithm becomes reducible. We circumvent the problem by introducing appropriate reparametrisations of the likelihood that can be used to construct irreducible data augmentation schemes.","PeriodicalId":388611,"journal":{"name":"2006 IEEE Nonlinear Statistical Signal Processing Workshop","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117051936","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：