Skip to content

Code for experiments and figures in "Reward Bases: Instanteous Reward Revaluation with Temporal Difference Learning"

License

Notifications You must be signed in to change notification settings

BerenMillidge/Reward_Bases

Repository files navigation

README

Install Python and make sure that the python command can be ran from your command line.

Install require python packages with:

pip install seaborn pandas matplotlib numpy scipy

Then run

python sea_salt_experiment.py

to produce:

Run

python papageorgiou_results.py

to produce:

Code for experiments and figures in "Reward Bases: Instanteous Reward Revaluation with Temporal Difference Learning". Paper can be found here: https://www.biorxiv.org/content/biorxiv/early/2022/04/14/2022.04.14.488361.full.pdf

Run

python room_task_experiments.py

to produce:

Further information

The files envs.py contains the environment classes for the various experiments and the learners.py file contains the classes for the reward basis, temporal difference, and successor representation agent

About

Code for experiments and figures in "Reward Bases: Instanteous Reward Revaluation with Temporal Difference Learning"

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages