Skip to content

Implementation of methods in "Teaching Robot Policies using Erroneous Examples" undergraduate thesis work.

License

Notifications You must be signed in to change notification settings

HARPLab/erroneous_example_augmented_taxi

 
 

Repository files navigation

Improving the Transparency of Agent Decision Making Using Demonstrations

This repository contains the raw code for all of the methods and environments (e.g. an augmented variant of the Taxi domain [Dietterich, JAIR2000]) described in the following thesis:

  • Michael S. Lee, Improving the Transparency of Agent Decision Making Using Demonstrations, Carnegie Mellon University, 2024,

with the goal of helping humans generalize to the agent's overall decision making in a particular domain by showing key agent decisions and behaviors to humans.

Please find the code for the individual methods and user studies the comprise chapters/papers of the thesis on Mike's website or Github, which will likely be more digestible.

Some of the required packages include numpy, matplotlib, pypoman to perform computational geometry with polytopes, and pygame if you want to visualize some MDPs.

About

Implementation of methods in "Teaching Robot Policies using Erroneous Examples" undergraduate thesis work.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 88.3%
  • Jupyter Notebook 11.7%