RE-Constrained-Latent-Graph

Code for the paper: Relation Extraction by Constrained Latent Graph

Prerequisites

We get and pre-process the data similar to dsre-vae.

Environment

conda create -n CLG python=3.8
conda activate CLG
pip install -r requirements.txt

Pretrained word embeddings

mkdir embeds
cd embeds 
wget http://nlp.stanford.edu/data/glove.6B.zip
unzip glove.6B.zip

Dataset Pre-processing

cd data

# To get the data
python get_data.py --dataset [nyt10, nyt10_570k, wiki_distant]      # nyt10 before nyt10_570k for sharing file

# Pre-process the data
python preprocess.py --max_sent_len 50 [30 for wikidata] --lowercase --max_bag_size 500 --path LANG/[nyt10, nyt10_570k, wiki_distant]  --dataset [nyt10, nyt10_570k, wiki_distant]

KB Pre-processing

To get the Knowledge Base prior, we need to get and train the KB embeddings.

# To get the KB
cd data/KB
python make_data.py --data [Wikidata, Freebase] \
                    --train_file PATH_TO_TRAIN_FILE
                    --val_file PATH_TO_VAL_FILE
                    --test_file PATH_TO_TEST_FILE

Training KB embeddings

In order to train Knowledge Base embeddings, we will use the DGL-KE package. The following script will train TransE entity and relation embeddings for Freebase and Wikidata.

cd data/KB
sh train_embeds.sh

Embeddings will be saved in the Freebase/ckpts_64/ and Wikidata/ckpts_64/ directories, respectively. Collect priors for your own KB:

python calculate_priors.py --kg_embeds Freebase/ckpts_64/TransE_l2_Freebase_0/Freebase_TransE_l2_entity.npy \
                           --e_map Freebase/entities.tsv \  # [Or Wikidata\...]
                           --data ../LANG/nyt10/nyt10_train.txt # [Or nyt10_570k Or wiki_distant]\
                           --kg [Freebase, Wikidata]

These can be directly downloaded here. (Thanks for Fenia Christopoulou‘s open source) #TODO: Clean unnecessary config and code------------------------