Implementation for the EMNLP 2021 paper.
apt-get -y update
apt-get install -y unzip zip parallel
conda create -p /tmp/imrc python=3.6 numpy scipy cython nltk
conda activate /tmp/imrc
pip install --upgrade pip
pip install numpy==1.16.2
pip install gym==0.15.4
pip install tqdm pipreqs pyyaml pytz visdom
conda install pytorch torchvision cudatoolkit=9.2 -c pytorch
pip install transformers
pip install allennlpThe original SQuAD dataset does not provide its test set, we take 23 wiki articles from its training set as our validation set. We then use the SQuAD dev set as our test set.
# download SQuAD from official website, then
python utils/split_original_squad.pyTo speed up training, we parse (tokenization and SRL) the dataset in advance.
python utils/preproc_squad.pyThis will result squad_split/processed_squad.1.1.split.[train/valid/test].json, which are used in iMRC tasks.
python utils/get_wiki_filter_squad.py
python utils/split_wiki_data.pyThis will result wiki_without_squad/wiki_without_squad_[train/valid/test].json, which are used to pre-train the continuous belief graph generator.
To train the agent equipped with different types of graphs, run:
# without graph
python main.py configs/imrc_none.yaml
# co-occurrence graph
python main.py configs/imrc_cooccur.yaml
# relative position graph
python main.py configs/imrc_rel_pos.yaml
# SRL graph
python main.py configs/imrc_srl.yaml
# continuous belief graph
# in this setting, we need a pre-trained graph generator.
# we provide our pre-trained graph generator at
# https://drive.google.com/drive/folders/1zZ7C_-xaYsfg2Ms7_BO5n3Qzx69UqMKD?usp=sharing
# one can choose to train their own version by:
python pretrain_observation_infomax.py configs/pretrain_cont_bnelief.yaml
# then using the downloaded/saved model checkpoint
python main.py configs/imrc_cont_belief.yamlTo change the task settings/configurations:
general:
naozi_capacity: 1 # capacity of agent's external memory queue (1, 3, 5)
generate_or_point: "point" # "qmpoint": q+o_t, "point": q, "generate": vocab
disable_prev_next: False # False: Easy Mode, True: Hard Mode
model:
recurrent: True # recurrent component described in Section 3.3 and Section 4.Additional Results@inproceedings{Yuan2021imrc_graph,
title={Interactive Machine Comprehension with Dynamic Knowledge Graphs},
author={Xingdi Yuan},
year={2021},
booktitle="EMNLP",
}