The Variational Coupled Gaussian Process Dynamical Model
نویسندگان
چکیده
We present a full variational treatment of the Coupled Gaussian Process Dynamical Model (CGPDM), which is a non-parametric, modular dynamical movement primitive model. Our work builds on similar developments in Gaussian state-space models, but we obviate the need for sampling, which results in a fast deterministic approximation for the posterior of latent states. We illustrate the performance of our model on synthetic data.
منابع مشابه
Gaussian Process Latent Variable Models for Dimensionality Reduction and Time Series Modeling
Time series data of high dimensions are frequently encountered in fields like robotics, computer vision, economics and motion capture. In this survey paper we look first at Gaussian Process Latent Variable Model (GPLVM) which is a probabilistic nonlinear dimensionality reduction method. Further we discuss Gaussian Process Dynamical Model (GPDMs) which are based GPLVM. GPDM is a probabilistic ap...
متن کاملVariational methods with coupled Gaussian functions for Bose-Einstein condensates with long-range interactions. I. General concept
The variational method of coupled Gaussian functions is applied to Bose-Einstein condensates with long-range interactions. The time dependence of the condensate is described by dynamical equations for the variational parameters. We present the method and analytically derive the dynamical equations from the time-dependent Gross-Pitaevskii equation. The stability of the solutions is investigated ...
متن کاملSupplementary Material: The Variational Coupled Gaussian Process Dynamical Model
To simplify the optimization of such terms, we would like to carry out the optimization with respect to the density q(u) analytically, so as to remove the dependency on q(u). To this end, we calculate for the optimal variational q∗(u) in the above equation. This approach was suggested in [1], however, it is not well described there. Here we give an extended derivation. A necessary condition for...
متن کاملVariational Gaussian Process State-Space Models
State-space models have been successfully used for more than fifty years in different areas of science and engineering. We present a procedure for efficient variational Bayesian learning of nonlinear state-space models based on sparse Gaussian processes. The result of learning is a tractable posterior over nonlinear dynamical systems. In comparison to conventional parametric models, we offer th...
متن کاملVariational Gaussian Process Dynamical Systems
High dimensional time series are endemic in applications of machine learning such as robotics (sensor data), computational biology (gene expression data), vision (video sequences) and graphics (motion capture data). Practical nonlinear probabilistic approaches to this data are required. In this paper we introduce the variational Gaussian process dynamical system. Our work builds on recent varia...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2017