We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.ME

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Methodology

Title: State-space deep Gaussian processes with applications

Authors: Zheng Zhao
Abstract: This thesis is mainly concerned with state-space approaches for solving deep (temporal) Gaussian process (DGP) regression problems. More specifically, we represent DGPs as hierarchically composed systems of stochastic differential equations (SDEs), and we consequently solve the DGP regression problem by using state-space filtering and smoothing methods. The resulting state-space DGP (SS-DGP) models generate a rich class of priors compatible with modelling a number of irregular signals/functions. Moreover, due to their Markovian structure, SS-DGPs regression problems can be solved efficiently by using Bayesian filtering and smoothing methods. The second contribution of this thesis is that we solve continuous-discrete Gaussian filtering and smoothing problems by using the Taylor moment expansion (TME) method. This induces a class of filters and smoothers that can be asymptotically exact in predicting the mean and covariance of stochastic differential equations (SDEs) solutions. Moreover, the TME method and TME filters and smoothers are compatible with simulating SS-DGPs and solving their regression problems. Lastly, this thesis features a number of applications of state-space (deep) GPs. These applications mainly include, (i) estimation of unknown drift functions of SDEs from partially observed trajectories and (ii) estimation of spectro-temporal features of signals.
Comments: See reproducible codes in this https URL Permanent link this http URL
Subjects: Methodology (stat.ME); Signal Processing (eess.SP); Machine Learning (stat.ML)
Journal reference: Doctoral dissertation, Aalto University, 2021
Cite as: arXiv:2111.12604 [stat.ME]
  (or arXiv:2111.12604v1 [stat.ME] for this version)

Submission history

From: Zheng Zhao [view email]
[v1] Wed, 24 Nov 2021 16:25:43 GMT (6643kb,D)

Link back to: arXiv, form interface, contact.