Skip to content

sfujim/LAP-PAL

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 

Repository files navigation

An Equivalence between Loss Functions and Non-Uniform Sampling in Experience Replay

PyTorch implementation of Loss-Adjusted Prioritized (LAP) experience replay and Prioritized Approximation Loss (PAL). LAP is an improvement to prioritized experience replay which eliminates the importance sampling weights in a principled manner, by considering the relationship to the loss function. PAL is a uniformly sampled loss function with the same expected gradient as LAP.

The paper will be presented at NeurIPS 2020. Code is provided for both continuous (with TD3) and discrete (with DDQN) domains.

Bibtex

@article{fujimoto2020equivalence,
  title={An Equivalence between Loss Functions and Non-Uniform Sampling in Experience Replay},
  author={Fujimoto, Scott and Meger, David and Precup, Doina},
  journal={Advances in Neural Information Processing Systems},
  volume={33},
  year={2020}
}

About

Author's PyTorch implementation of LAP and PAL with TD3 and DDQN

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages