GP-BayesFilters: Bayesian filtering using Gaussian process prediction and observation models

By Jonathan Ko, Dieter Fox. 2008

449 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Signal Processing

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

In Signal Processing

Citations in the corpus

Neural Network Dynamics for Model-Based Deep Reinforcement Learning with Model-Free Fine-Tuning (2017, methodology): “However, the most efficient model-based algorithms have used relatively simple function approximators, such as Gaussian processes [6], [21], [22], time-varying linear models [5], [23], [24], and mixtures of Gaussians [25].”
Algorithmic Framework for Model-based Deep Reinforcement Learning with Theoretical Guarantees (2018, methodology): “…2011b; Morimoto & Atkeson, 2003; Deisenroth et al., 2011) using dynamical models ranging from Gaussian process (Deisenroth & Rasmussen, 2011b; Ko & Fox, 2009), time-varying linear models (Levine & Koltun, 2013; Lioutikov et al., 2014; Levine & Abbeel, 2014; Yip & Camarillo, 2014), mixture of…”
Deep Dynamics Models for Learning Dexterous Manipulation (2019, background): “Other work has focused on learning these models using high-capacity function approximators [17, 18, 3, 19, 20, 21, 22] and probabilistic dynamics models [23, 24, 25, 26].”
Gaussian Processes for Data-Efficient Learning in Robotics and Control (2015, neutral): “matching [15], [44] and linearization of the posterior GP mean function [28].”
Learning to Adapt in Dynamic, Real-World Environments through Meta-Reinforcement Learning (2018, methodology): “Prior model-based approaches tackled this problem by incorporating model uncertainty using Gaussian Processes (GPs) (Ko and Fox, 2009; Deisenroth and Rasmussen, 2011; Doerr et al., 2017).”
Transfer from Simulation to Real World through Learning Deep Inverse Dynamics Model (2016, methodology): “Our approach is agnostic to how the policy πsource was obtained.”
A Practical and Conceptual Framework for Learning in Control (2010, methodology): “In a practical application, if idealized transition dynamics are known, they can be used as a prior mean function as proposed for example by Ko et al. (2007) and Ko and Fox (2008, 2009a,b) in the context of RL and mechanical control.”
Expectation Propagation in Gaussian Process Dynamical Systems (2012, methodology): “In about 20% of the test cases, the inference methods based on explicit linearization of the posterior mean function (GPEKS and EP-GPEKS) ran into numerical problems typical of linearizations [6], i.e., overconﬁdent posterior distributions that caused numerical problems.”
Probabilistic Recurrent State-Space Models (2018, methodology): “Gaussian Process State-Space Models (GP-SSMs) are a popular class of probabilistic SSMs (Wang et al., 2008; Ko & Fox, 2009; Turner et al., 2010; Frigola et al., 2013; 2014; Eleftheriadis et al., 2017).”
Learning to Adapt: Meta-Learning for Model-Based Control (2018, background): “Model Learning: Many model-based approaches aim to learn a single global model, using function approximators such as Gaussian processes [20, 10] and neural networks [24, 43, 13].”
Model learning for robot control: a survey (2011, background): “2007, 2008), and state estimation (Ko and Fox 2009).”
Multi-modal estimation with kernel embeddings for learning motion models (2013, methodology): “The GP-BayesFilter particle filter use a standard sequential importance re-sampling (SIR) particle filter but learns the state transition and observation models from data using a set of Gaussian processes [19].”
How to Construct Deep Recurrent Neural Networks (2013, background): “Ko and Dieter (2009) proposed deep transitions for Gaussian Process models.”
Data-Efficient Reinforcement Learning in Continuous State-Action Gaussian-POMDPs (2017, methodology): “Larger nonlinearities warrant the unscented Kalman filter (UKF) – a deterministic sampling technique to estimate moments – or particle methods [7, 12].”
Toward fast policy search for learning legged locomotion (2012, methodology): “As an alternative to moment matching, approximations based on linearization are conceivable [11].”