Learning to Follow Instruction in Text-Based Games

This is the code repository for the paper Learning to Follow Instruction in Text-Based Games, which will appear in NeurIPS 2022.

Installing Dependencies

We used python-3.9 in our experiments.

The main dependencies can be installed as pip install -r requirements.txt

We use the Spot library for LTL progression, which can be installed by following the instructions here. It should be installed to <path-to-python-env>/lib/python??/site-packages/spot

Once that's complete, we can download spacy

python -m spacy download en

and also download the FastText Word Embeddings

cd src/support-files
curl -L -o crawl-300d-2M.vec.h5 "https://bit.ly/2U3Mde2"

The dataset itself can be downloaded as

cd src/support-files/rl.0.2
wget https://aka.ms/twkg/rl.0.2.zip
unzip rl.0.2.zip

Note finally that vocabularies can be found in src/support-files/vocabularies. The location of the dataset and vocabularies is not important, and can be specified in the config.yaml file.

Training Agents

Instructions for training and evaluating agents on is available in the src folder.

Citation