Skip to content

Analyzing policy entropy of reinforcement learning agents

License

Notifications You must be signed in to change notification settings

sukiboo/policy_entropy

Repository files navigation

Policy Entropy of RL Agents

This repository contains the source code for the numerical experiments presented in the paper "Analyzing Policy Entropy of Reinforcement Learning Agents for Personalization Tasks".

How to Use

Installation

Install the requirements via pip install -r requirements.txt.

Run Experiments

Run the experiments via python -m run_experiment -c config, where config is a configuration file in ./configs/ directory. The available values are {config_mnist, config_cifar10, config_spotify, config_recogym, config_personalization}, which could be specified to recreate each of the presented numerical experiments. Optionally, a custom experiment can be set up by changing or adding new configuration file.

Load Experiments

All previously performed experiments are stored in ./data/ directory and can be recreated by loading via python -m run_experiment -l exp_name, where exp_name is the name of the experiment as it is saved in ./data/.

Results

Reward values Entropy values

File Structure

  • run_experiment.py --- set up and run the experiment
  • agent.py --- set up selected RL agents
  • environment.py --- create the specified environment
  • environments/ --- data required to set up various environments
  • configs/ --- configuration files for the experiments
  • data/ --- store data from previously run experiments
  • images/ --- plots of various results from experiments
  • visualization.py --- save/load the experiment data, plot the results

License

This project is licensed under the MIT License.

Releases

No releases published

Packages

No packages published

Languages