Path Integral Control by Reproducing Kernel Hilbert Space Embedding
By Konrad Rawlik, Marc Toussaint, Sethu Vijayakumar. 2012
27 citations recorded by the source corpus.
We present an embedding of stochastic optimal control problems, of the so called path integral form, into reproducing kernel Hilbert spaces. Using consistent, sample based estimates of the embedding leads to a model-free, non-parametric approach for calculation of an approximate solution to the control problem. This formulation admits a decomposition of the problem into an invariant and task dependent component. Consequently, we make much more efficient use of the sample data compared to previous sample based approaches in this domain, e.g., by allowing sample re-use across tasks. Numerical examples on test problems, which illustrate the sample efficiency, are provided.
Subfields: Artificial Intelligence, Statistical and Nonlinear Physics, Statistics and Probability
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
- Sergey Levine
- Gerhard Neumann
- Pieter Abbeel
- Jan Peters
- Herke Van Hoof
- Byron Boots
- Chelsea Finn
- Le Song
In Statistical and Nonlinear Physics
- Yu Nishiyama
- Byron Boots
- Jan Peters
- Le Song
- Gerhard Neumann
- Herke Van Hoof
- Kenji Fukumizu
- Arthur Gretton
In Statistics and Probability
- Jan Peters
- Yu Nishiyama
- Le Song
- Amir Hossein Afsharinejad
- Shunsuke Naruse
- Gerhard Neumann
- Byron Boots
- Herke Van Hoof
Citations in the corpus
- The Nonparametric Kernel Bayes Smoother (2016, methodology): “Examples include the nonparametric kernel Bayes’ filter (nKB-filter) on discrete-time time-invariant state space models [10] and the Semiparametric kernel Monte-Carlo filter (KMC-filter) [16], as well as methods for dynamical system modeling and reinforcement learning applications [14, 21, 23, 4, 3,”
- Non-parametric Policy Search with Limited Information Loss (2017, methodology): “Alternative non-parametric methods were proposed by Rawlik et al. (2013a) and Deisenroth and Rasmussen (2011).”