D. Koller and U. Lerner. Sampling in factored dynamic systems. In Doucet et al. [22].  Home/Search   Document Details and Download   Summary   Related Articles   Check

....a fundamental role in Assisted Cognition systems. We address the location estimation problem as an instance of Bayesian filtering, which represents the state of a dynamic system by random variables and estimates posterior distributions recursively using probabilistic models of sensors and dynamics [51, 84, 40, 91, 103]. The key problem in connecting location estimation to high level reasoning is to bridge the gap between continuous, noisy sensor data and discrete, high level representations. Consider, for example, the task of tracking a person s outdoor location using GPS data. Figure 1 shows raw GPS readings ....

D. Koller and U. Lerner. Sampling in factored dynamic systems. In Doucet et al. [41].

.... x 1 ,x2,x 3 based on Yl,Y2,Y3 and weight it by P(Y2 Ixl)P(Y3 Ix2) In this way, the samples generated will be driven by the evidence and should be more effective particularly when evidence likelihood is very small [19] A particle filtering with important sampling was proposed for DPNs [19][20]. Figure 1. A Simple DPN To deal with large networks with many evidence nodes or extremely unlikely evidence, an adaptive 528 importance sampling algorithm, AIS BN, was proposed [18] The algorithm shows promising convergence rates even under extreme conditions and seems to outperform the ....

D. Koller and U. Lerner, "Sampling in Factored Dynamic Systems," A book chapter in Sequential Mo,te Carlo Methods in Practice, A. Doucet, J.F.G. de Freitas, and N. Gordon, Eds., Springer-Verlag 2000.

....and discrete dynamics and have been explored for hybrid diagnosis in [16] However, autonomous transitions between modes triggered by the continuous dynamics have not been considered. Particle filtering has been applied also for a class of hybrid systems modeled by dynamic Bayesian networks in [12] where the autonomous transitions between discrete states are only defined using the so called softmax conditional probability distributions. Hybrid diagnosis based on timed discrete event representations has been studied also in [15] In these methodologies, the continuous state is quantized and ....

D. Koller and U. Lerner. Sampling in factored dynamic systems. In Doucet et al. [4], pages 445--464.

....on 3 2 1 , y y y and weight it by ) 2 3 1 2 x y p x y p . In this way, the samples generated will be driven by the evidence and should be more effective particularly when evidence likelihood is very small [19] A particle filtering with important sampling was proposed for DPNs [19][20]. Figure 1. A Simple DPN To deal with large networks with many evidence nodes or extremely unlikely evidence, an adaptive importance sampling algorithm, AIS BN, was proposed [18] The algorithm shows promising convergence rates even under extreme conditions and seems to outperform the existing ....

D. Koller and U. Lerner, "Sampling in Factored Dynamic Systems," A book chapter in Sequential Monte Carlo Methods in Practice, A. Doucet, J.F.G. de Freitas, and N. Gordon, Eds., Springer-Verlag 2000.

No context found.

D. Koller and U. Lerner. Sampling in factored dynamic systems. In Doucet et al. [22].

No context found.

D. Koller and U. Lerner. Sampling in factored dynamic systems. In Doucet et al. [3], pages 445--464.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC