Uncertainty-Based-Offline-RL-with-Diversified-Q-Ensemble

This is a blog post based on the paper Uncertainty-Based Offline Reinforcement Learning with Diversified Q-Ensemble, which proposes the offline RL algorithm EDAC (Ensemble Diversified Actor Critic).

There are several implementations:

• official
• CORL (Clean Offline Reinforcement Learning) or our fork
• our own implementation

run blog locally

install

• install Quarto
• clone this repo

pip install -r blog/requirements.txt

run

quarto preview blog/blog.ipynb

writing

The main content is in blog/blog.ipynb, the references are in blog/references.bib and the images are in blog/figures/.