Citation

BibTex format

@inproceedings{Eleftheriadis:2017,
author = {Eleftheriadis, S and Nicholson, TFW and Deisenroth, MP and Hensman, J and Eleftheriadis, S and Nicholson, TFW and Deisenroth, M and Hensman, J},
pages = {5310--5320},
publisher = {Neural Information Processing Systems Foundation, Inc.},
title = {Identification of Gaussian Process State Space Models},
url = {http://arxiv.org/abs/1705.10888v1},
year = {2017}
}

RIS format (EndNote, RefMan)

TY  - CPAPER
AB - The Gaussian process state space model (GPSSM) is a non-linear dynamicalsystem, where unknown transition and/or measurement mappings are described byGPs. Most research in GPSSMs has focussed on the state estimation problem.However, the key challenge in GPSSMs has not been satisfactorily addressed yet:system identification. To address this challenge, we impose a structuredGaussian variational posterior distribution over the latent states, which isparameterised by a recognition model in the form of a bi-directional recurrentneural network. Inference with this structure allows us to recover a posteriorsmoothed over the entire sequence(s) of data. We provide a practical algorithmfor efficiently computing a lower bound on the marginal likelihood using thereparameterisation trick. This additionally allows arbitrary kernels to be usedwithin the GPSSM. We demonstrate that we can efficiently generate plausiblefuture trajectories of the system we seek to model with the GPSSM, requiringonly a small number of interactions with the true system.
AU - Eleftheriadis,S
AU - Nicholson,TFW
AU - Deisenroth,MP
AU - Hensman,J
AU - Eleftheriadis,S
AU - Nicholson,TFW
AU - Deisenroth,M
AU - Hensman,J
EP - 5320
PB - Neural Information Processing Systems Foundation, Inc.
PY - 2017///
SN - 1049-5258
SP - 5310
TI - Identification of Gaussian Process State Space Models
UR - http://arxiv.org/abs/1705.10888v1
UR - http://hdl.handle.net/10044/1/52526
ER -