Skip to content

temetski/moral_agent

Repository files navigation

Setup

Create and activate a virtual env [optional]

python -m venv venv
source venv/bin/activate

And then install all prerequisite packages

pip install -r requirments.txt

Pretraining a model

Models can be pretrained using the following command:

python algorithms/ppo.py --seed 42 --env-id environments.milk:FindMilk-v4 --num-envs 8

Note that the --env-id parameter uses the format <path.to.module>:<name_of_env>.

Citing

If your find any part of this work useful to your research, your are encouraged to cite our paper:

@article{amuled2026,
  author = {Dubey, Rohit K. and Dailisan, Damian and Mahajan, Sachit},
  title = {Addressing Moral Uncertainty using Large Language Models for Ethical Decision-Making},
  journal = {Frontiers in Artificial Intelligence},
  year = {2026},
  doi = {10.48550/ARXIV.2503.05724},
  note = {Accepted for publication},
  keywords = {Computers and Society (cs.CY), Artificial Intelligence (cs.AI)},
  copyright = {Creative Commons Attribution Non Commercial No Derivatives 4.0 International}
}

About

This repository hosts the code for the paper "AMULED: Addressing Moral Uncertainty using Large Language Models for Ethical Decision-Making"

Resources

Stars

Watchers

Forks

Contributors