Skip to content

Example code for the paper "RL Exploration of T-Mazes with FEP-Based Intrinsic Rewards"

Notifications You must be signed in to change notification settings

oist-cnru/curious_maze

Repository files navigation

curious_maze

Example code for the paper "Intrinsic Rewards for Exploration without Harm from Observational Noise: A Simulation Study Based on the Free Energy Principle".

Train an agent without intrinsic rewards with

python run main.py

or adjust arguments like

python main.py --curiosity hidden_state --alpha None

About

Example code for the paper "RL Exploration of T-Mazes with FEP-Based Intrinsic Rewards"

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages