Data-Efficient Reinforcement Learning in Continuous-State POMDPs
By Rowan Thomas McAllister, Carl E. Rasmussen. 2016
12 citations recorded by the source corpus.
We present a data-efficient reinforcement learning algorithm resistant to observation noise. Our method extends the highly data-efficient PILCO algorithm (Deisenroth & Rasmussen, 2011) into partially observed Markov decision processes (POMDPs) by considering the filtering process during policy evaluation. PILCO conducts policy search, evaluating each policy by first predicting an analytic distribution of possible system trajectories. We additionally predict trajectories w.r.t. a filtering process, achieving significantly higher performance than combining a filter with a policy optimised by the original (unfiltered) framework. Our test setup is the cartpole swing-up task with sensor noise, which involves nonlinear dynamics and requires nonlinear control.
Subfields: Artificial Intelligence, Computer Networks and Communications, Electrical and Electronic Engineering
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
- Rowan Thomas McAllister
- Carl E. Rasmussen
- Sergey Levine
- Pieter Abbeel
- Marc Peter Deisenroth
- Jan Peters
- Andrew Gordon Wilson
- Hannes Nickisch
In Computer Networks and Communications
In Electrical and Electronic Engineering
Citations in the corpus
- Improving PILCO with Bayesian Neural Network Dynamics Models (2016, methodology): “Second, we can consider observation noise and incorporate a filter, analogous to [14].”