Learning to Execute Actions or Ask Clarification Questions

Here are the dataset and codes for the Findings of NAACL paper titled "Learning to Execute Actions or Ask Clarification Questions".

Dependencies

python==3.7
torch==1.7
tensorboardX
prettytable

Introduction

An intelligent agent should not only understand and execute the instructor's requests but also be able to take initiatives, e.g., asking clarification questions, in case the instructions are ambiguous.

A simple example of builder task.

Code and Dataset

• CollaborativeBuilding: Codes for collaborative building task;
• LearnToAsk: Codes for learning to ask task and joint learning task;
• builder_utterance_labels.json: Annotations of all builder utterances. Please ignore builder_utterance_labels.txt, which is our draft version.
• The raw dataset Minecraft Dialogue Corpus is from the repository.

1. Dataset Preparation

Please download the original dataset. Then

unzip data.zip
cd data
wget https://nlp.stanford.edu/data/glove.42B.300d.zip
unzip glove.42B.300d.zip
cd ../CollaborativeBuilding/builder
python vocab.py --lower --use_speaker_tokens --oov_as_unk --all_splits --add_builder_utterances
cd ../..

2. Code

Please run codes in CollaborativeBuilding and LearnToAsk for Collaborative Building task and Learning to Ask task respectively.

Citation

Please cite our work if it is helpful.

@inproceedings{Shi2022learning,
title = {Learning to Execute Actions or Ask Clarification Questions},
author = {Shi, Zhengxiang and Feng, Yue and Lipani, Aldo},
year = {2022},
address = {Seattle, Washington, USA},
booktitle = {Findings of the North American Chapter of the Association for Computational Linguistics},
publisher = {Association for Computational Linguistics},
keywords = {Conversational System, Clarification Questions},
url = {https://arxiv.org/abs/2204.08373}
}