This is the repository for the Value Aggregation project. It is chiefly maintained by Jared Moore ([email protected]).

Repository Structure

• env-aggregation
  • Run make init to set up.
• external_data
  • Run make data to set up
• experiments
  • intuition: Files and data to run experiments on humans (mturk) and llms. See the README.
  • prevalence: Testing for the prevalence of different models in natural data sets.
    • nlp_positionality.ipynb: An example of processing the NLPositionality data and comparing formal models.
    • moral_machines.ipynb: An example of processing the Moral Machines data and comparing formal models.
    • taylor_data.ipynb: An example of processing the processed Common Sense Normbank data from Taylor's Value Kaleidoscope project.
• src: The files for the value_aggregation python package
• demos.ipynb: Various examples of moral uncertainty between different models.
• Makefile
  • make test: runs the test suite
  • make init: initializes the virtual environment and installs dependencies
  • make data: downloads external data into ./external_data
• pyproject.toml: For setting up the model as a package.
• README.md
• requirements.txt: Python dependencies
• setup.py: For setting up the module as a package.

src (the value_aggregation package)

• value_aggregation: The implementation of the formal models
  • __init__.py
  • Game.py: Lays out the various formal models and their parameters.
  • GameState.py: Constructors for various kinds of game states, used in different game set ups. For example: MEC, NBS, MFT, MFO, and the (discarded) expecti-max and equilibria selection.
  • utils.py: Utility functions for the package.
• tests: Various tests of the functionality in parliament
  • __init__.py
  • context.py
  • TestGameNodes.py
  • TestGameState.py
  • TestOtherTheories.py
  • TestRunParliament.py
  • TestUtils.py

Running jupyter notebooks

Install jupyter.

Run source env-parliament/bin/activate; jupyter notebook and open any .ipynb file.