Spatiotemporal Learning via Infinite-Dimensional Bayesian Filtering and Smoothing: A Look at Gaussian Process Regression Through Kalman Filtering

By Simo Särkkä, Arno Solin, Jouni Hartikainen. 2013

279 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Statistical and Nonlinear Physics

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Michael A. Osborne
2. Stephen J. Roberts
3. Suzanne Aigrain
4. Steven Reece
5. Zhidi Lin
6. Yue Xu
7. Feng Yin
8. Deshi Li

In Statistical and Nonlinear Physics

1. Michael A. Osborne
2. Stephen J. Roberts
3. Vinesh Rajpaul
4. Suzanne Aigrain

Citations in the corpus

• Recurrent Neural Filters: Learning Independent Bayesian Filtering Steps for Time Series Prediction (2019, neutral): “…they have similar benefits to Bayes filters (i.e. predictive uncertainties, natural multistep prediction etc.), inference at each time step has at least an O ( T ) complexity in the number of past observations – either via sparse GP approximations or Kalman filter formulations (Sarkka et al., 2013).”
• FedLoc: Federated Learning Framework for Data-Driven Cooperative Localization and Location Data Processing (2020, methodology): “In [41, 42], Gaussian processes implemented via recursive Kalman filtering are used to model spatio-temporal data with rather low computational complexity.”
• A Gaussian process framework for modelling stellar activity signals in radial velocity data (2015, methodology): “In the case of time-series analysis, there exist very efficient O ( N ) approaches for Gaussian process regression based on approximate re-interpretation as filters (Reece & Roberts 2010; Sarkka et al. 2013; Reece et al. 2014).”