[REQUEST] Support for REM (Random Ensemble Mixture) variant of DQN

takuseno / d3rlpy

An offline deep reinforcement learning library

https://takuseno.github.io/d3rlpy

MIT License

1.3k stars 232 forks source link

[REQUEST] Support for REM (Random Ensemble Mixture) variant of DQN #220

Open itsmesaisatish opened 2 years ago

itsmesaisatish commented 2 years ago

Is there any plan to support Random Ensemble Mixture(REM) variant of DQN . It helps for robust offline training and results are much better compared to other variants of DQN.

Please refer Google Deepmind's paper with code for reference - https://arxiv.org/abs/1907.04543v4

Regards, Sai

takuseno commented 2 years ago

@itsmesaisatish Thanks for the issue. Currently, I'm focusing on publishing the journal paper about this library. Once we get it published, we'll start adding new algorithms again.