Search Results for author: Felipe Tobar

Found 16 papers, 3 papers with code

Streaming computation of optimal weak transport barycenters

no code implementations26 Feb 2021 Elsa Cazelles, Felipe Tobar, Joaquin Fontbona

The concept of weak barycenter and our computation approaches are illustrated on synthetic examples, validated on 2D real-world data and compared to the classical Wasserstein barycenters.

Bayesian Reconstruction of Fourier Pairs

no code implementations9 Nov 2020 Felipe Tobar, Lerko Araya-Hernández, Pablo Huijse, Petar M. Djurić

Our aim is to address the lack of a principled treatment of data acquired indistinctly in the temporal and frequency domains in a way that is robust to missing or noisy observations, and that at the same time models uncertainty effectively.

Gaussian process imputation of multiple financial series

no code implementations11 Feb 2020 Taco de Wolff, Alejandro Cuevas, Felipe Tobar

In Financial Signal Processing, multiple time series such as financial indicators, stock prices and exchange rates are strongly coupled due to their dependence on the latent state of the market and therefore they are required to be jointly analysed.

Imputation Time Series

MOGPTK: The Multi-Output Gaussian Process Toolkit

1 code implementation9 Feb 2020 Taco de Wolff, Alejandro Cuevas, Felipe Tobar

We present MOGPTK, a Python package for multi-channel data modelling using Gaussian processes (GP).

Gaussian Processes Imputation

The Wasserstein-Fourier Distance for Stationary Time Series

1 code implementation11 Dec 2019 Elsa Cazelles, Arnaud Robert, Felipe Tobar

The WF distance operates by calculating the Wasserstein distance between the (normalised) power spectral densities (NPSD) of time series.

Data Augmentation Dimensionality Reduction +2

Band-Limited Gaussian Processes: The Sinc Kernel

no code implementations NeurIPS 2019 Felipe Tobar

We propose a novel class of Gaussian processes (GPs) whose spectra have compact support, meaning that their sample trajectories are almost-surely band limited.

Gaussian Processes

Compositionally-Warped Gaussian Processes

no code implementations23 Jun 2019 Gonzalo Rios, Felipe Tobar

The Gaussian process (GP) is a nonparametric prior distribution over functions indexed by time, space, or other high-dimensional index set.

Gaussian Processes

Low-pass filtering as Bayesian inference

no code implementations9 Feb 2019 Cristobal Valenzuela, Felipe Tobar

We propose a Bayesian nonparametric method for low-pass filtering that can naturally handle unevenly-sampled and noise-corrupted observations.

Bayesian Inference Gaussian Processes +1

Bayesian Nonparametric Spectral Estimation

1 code implementation NeurIPS 2018 Felipe Tobar

Spectral estimation (SE) aims to identify how the energy of a signal (e. g., a time series) is distributed across different frequencies.

Time Series

Bayesian Learning with Wasserstein Barycenters

no code implementations28 May 2018 Julio Backhoff-Veraguas, Joaquin Fontbona, Gonzalo Rios, Felipe Tobar

We introduce a novel paradigm for Bayesian learning based on optimal transport theory.

Learning non-Gaussian Time Series using the Box-Cox Gaussian Process

no code implementations19 Mar 2018 Gonzalo Rios, Felipe Tobar

Gaussian processes (GPs) are Bayesian nonparametric generative models that provide interpretability of hyperparameters, admit closed-form expressions for training and inference, and are able to accurately represent uncertainty.

Gaussian Processes Time Series

Spectral Mixture Kernels for Multi-Output Gaussian Processes

no code implementations NeurIPS 2017 Gabriel Parra, Felipe Tobar

Early approaches to multiple-output Gaussian processes (MOGPs) relied on linear combinations of independent, latent, single-output Gaussian processes (GPs).

Gaussian Processes

Recovering Latent Signals from a Mixture of Measurements using a Gaussian Process Prior

no code implementations19 Jul 2017 Felipe Tobar, Gonzalo Rios, Tomás Valdivia, Pablo Guerrero

The proposed model is validated in the recovery of three signals: a smooth synthetic signal, a real-world heart-rate time series and a step function, where GPMM outperformed the standard GP in terms of estimation error, uncertainty representation and recovery of the spectral content of the latent signal.

Bayesian Inference Time Series

Improving Sparsity in Kernel Adaptive Filters Using a Unit-Norm Dictionary

no code implementations13 Jul 2017 Felipe Tobar

Kernel adaptive filters, a class of adaptive nonlinear time-series models, are known by their ability to learn expressive autoregressive patterns from sequential data.

Time Series

Initialising Kernel Adaptive Filters via Probabilistic Inference

no code implementations11 Jul 2017 Iván Castro, Cristóbal Silva, Felipe Tobar

We present a probabilistic framework for both (i) determining the initial settings of kernel adaptive filters (KAFs) and (ii) constructing fully-adaptive KAFs whereby in addition to weights and dictionaries, kernel parameters are learnt sequentially.

Time Series

Learning Stationary Time Series using Gaussian Processes with Nonparametric Kernels

no code implementations NeurIPS 2015 Felipe Tobar, Thang D. Bui, Richard E. Turner

We introduce the Gaussian Process Convolution Model (GPCM), a two-stage nonparametric generative procedure to model stationary signals as the convolution between a continuous-time white-noise process and a continuous-time linear filter drawn from Gaussian process.

Denoising Gaussian Processes +2

Cannot find the paper you are looking for? You can Submit a new open access paper.