Home

Awesome

Intelligent Go-Explore

arXiv Website Twitter

<p align="center"> <img src="figs/ige.png" style="width: 80%;" /> </p>

Intelligent Go-Explore (IGE) integrates the intelligence and internalized human notions of interestingness from giant pretrained foundation models into all stages of the Go-Explore algorithm, enabling foundation model agents to robustly explore in complex environments.

Intelligent Go-Explore: Standing on the Shoulders of Giant Foundation Models; Cong Lu, Shengran Hu, Jeff Clune.

<p align="center"> <a href=https://arxiv.org/abs/2405.15143>View on arXiv</a> </p>

Setup

To install, clone the repository and run the following:

conda create -n ige python=3.11
conda activate ige
pip install -r requirements.txt

# BabyAI-Text Instructions
git clone https://github.com/flowersteam/Grounding_LLMs_with_online_RL.git
cd Grounding_LLMs_with_online_RL
pip install blosc; cd babyai-text/babyai; pip install -e .; cd ..
cd gym-minigrid; pip install -e.; cd ..
pip install -e .

# TextWorld Instructions (see https://github.com/microsoft/TextWorld for system libraries)
pip install textworld

Running Instructions

Notes

Game of 24

python3 game_of_24/run.py

BabyAI-Text

python3 babyai/run.py

TextWorld

# Run Coin Collector
python3 textworld/run_coin.py

# Run The Cooking Game
python3 textworld/run_cooking.py

# Run Treasure Hunter
python3 textworld/run_treasure.py

N.B. In order to generate the game files we used in TextWorld found in tw_games, we used the text_world/misc/make_*.py scripts.

Citation

If you find this project useful, please consider citing:

@article{lu2024IntelligentGoExplore,
title={{Intelligent Go-Explore}: Standing on the Shoulders of Giant Foundation Models},
author={Lu, Cong, and Hu, Shengran and Clune, Jeff},
journal={arXiv preprint arXiv:2405.15143},
year={2024}
}

Contact

Please contact Cong Lu for any queries.