Collaborative decision-oriented dialogue environments for humans and LLM agents.
These environments are associated with the paper Decision-Oriented Dialogue For Human-AI Collaboration.
Get started and play the games with GPT-3 as an assistant:
pip install -e .
vim dialop/.api_key
python dialop/play_gpt3.py --game {optimization, planning, mediation}
games: define game logic, environment data generationenvs: text-based interfaces for gamesapps: web app interfaces for games for human-human play
| Name | Description | Human UI | Agent Text Interface | Hosted UI |
|---|---|---|---|---|
Optimization |
Assign the best reviewers for conference papers | ![]() |
![]() |
Play now |
Planning |
Plan the best itinerary with the least travel and price | ![]() |
![]() |
Play now |
Mediation |
Coordinate a set of group flights that works well for everyone | ![]() |
![]() |
Play now |
To play against yourself with the text-based environment:
python dialop/play.py --game {optimization, planning, mediation}
To play against yourself with the web UI:
cd dialop/apps
game={optimization, planning, mediation} flask run
For Planning, you'll need to input a MapboxGL access key in static/client.js to show the map on the agent side.
To see human-human dialogues:
unzip dialop/data.zip
Coming soon!
@article{lin2023decision,
title={Decision-Oriented Dialogue for Human-AI Collaboration},
author={Lin, Jessy and Tomlin, Nicholas and Andreas, Jacob and Eisner, Jason},
journal={Transactions of the Association for Computational Linguistics}
year={2024}
}
This project may contain trademarks or logos for projects, products, or services. Authorized use of Microsoft trademarks or logos is subject to and must follow Microsoft’s Trademark & Brand Guidelines. Use of Microsoft trademarks or logos in modified versions of this project must not cause confusion or imply Microsoft sponsorship. Any use of third-party trademarks or logos are subject to those third-party’s policies.







