JEDAI Explains Decision-Making AI

This repository contains the code for the tool JEDAI. In-detail review of its architecture and features is given in paper at ArXiv, appearing as -

Using Explainable AI and Hierarchical Planning for Outreach with Robots.
Daksh Dobhal*, Jayesh Nagpal*, Rushang Karia, Pulkit Verma, Rashmeet Kaur Nayyar, Naman Shah, and Siddharth Srivastava.
ArXiv 2404.00808, 2024
Paper

System Demonstration of JEDAI was published at AAMAS 2022 as:

JEDAI: A System for Skill-Aligned Explainable Robot Planning.
Naman Shah*, Pulkit Verma*, Trevor Angle, and Siddharth Srivastava.
21st International Conference on Autonomous Agents and MultiAgent Systems, 2022.

Paper | Demo Talk | Tutorial

Docker Image

The recommended way of using JEDAI is to use the pre-configured docker image that is available here: JEDAI

Once all the files are downloaded, run the following command to load docker images -

docker load --input jedai_public.tar
docker load --input novnc.tar

Running JEDAI

Now, to start a JEDAI instance, run the following script from the directory where all the files have been downloaded -

./start_instances.sh -n <instance_count>

Replace <instance_count> with the number of instances you want to start.

After running the script, you can access the JEDAI interface on the web by going to the link (in green) in the output text in the terminal.

To stop the JEDAI instance, run the following -

./stop_instances.sh -n <instance_id>

Replace <instance_id> with the id of the instance you want to stop.

When running the ./start_instances.sh script, the instance ID is given alongside the link to access the JEDAI interface. Also, the Instance ID is the integer following the "instance" in the docker image name - instance<ID>...

Contributors

Daksh Dobhal
Jayesh Nagpal
Rushang Karia
Pulkit Verma
Rashmeet Kaur Nayyar
Naman Shah
Trevor Angle
Kiran Prasad
Amruta Tapadiya
Kyle Atkinson
Chirav Dave
Judith Rosenke
Siddharth Srivastava

Citation

The latest version of this work can be cited as :

@article{dobhal2024using,
  title={Using Explainable AI and Hierarchical Planning for Outreach with Robots},
  author={Dobhal, Daksh and Nagpal, Jayesh and Karia, Rushang and Verma, Pulkit and Nayyar, Rashmeet Kaur and Shah, Naman and Srivastava, Siddharth},
  journal={arXiv preprint arXiv:2404.00808},
  year={2024}
}

The earlier version (AAMAS 2022) can be cited as :

@inproceedings{shah_2022_jedai,
    author = {Shah, Naman and Verma, Pulkit and Angle, Trevor and Srivastava, Siddharth},
    title = {{JEDAI: A System for Skill-Aligned Explainable Robot Planning}},
    booktitle = {Proceedings of the 21st International Conference on Autonomous Agents and MultiAgent Systems},
    year={2022}
}