Importance sampling policy gradient algorithms in reproducing kernel Hilbert space

@article{Le2017ImportanceSP,
  title={Importance sampling policy gradient algorithms in reproducing kernel Hilbert space},
  author={Tuyen Pham Le and Vien Anh Ngo and P. Marlith Jaramillo and TaeChoong Chung},
  journal={Artificial Intelligence Review},
  year={2017},
  pages={1-21}
}
Modeling policies in reproducing kernel Hilbert space (RKHS) offers a very flexible and powerful new family of policy gradient algorithms called RKHS policy gradient algorithms. They are designed to optimize over a space of very high or infinite dimensional policies. As a matter of fact, they are known to suffer from a large variance problem. This critical… CONTINUE READING