This project provides the data and models described on the paper "NeuralREG: An end-to-end approach to referring expression generation".

webnlg/

Original and Delexicalized versions of WebNLG corpus

preprocessing.py

Script for extracting the referring expression collection from the WebNLG corpus. Update the variable paths in the script and run the command:

python2.7 preprocessing.py

data/

Training, development and test referring expressions sets and vocabularies.

only_names.py OnlyNames model

Update variable paths in the script and run the following command:

python2.7 only_names.py

ferreira/ Ferreira model

Update variable paths in the scripts and execute them in the following order to train the model and to generate the referring expressions:

python2.7 reg_train.py

python2.7 reg_main.py

seq2seq.py NeuralREG+Seq2Seq model

Update the variable paths in the script and run the following command:

python3 seq2seq.py --dynet-autobatch 1 --dynet-mem 8192 --dynet-gpu

attention.py NeuralREG+CAtt model

Update the variable paths in the script and run the following command:

python3 attention.py --dynet-autobatch 1 --dynet-mem 8192 --dynet-gpu

hierattention.py NeuralREG+HierAtt model

Update the variable paths in the script and run the following command:

python3 hierattention.py --dynet-autobatch 1 --dynet-mem 8192 --dynet-gpu

eval/

Automatic evaluation scripts to extract information about the referring expression collection (corpus.py), to obtain the results depicted in the paper (evaluation.py) and to test statistical significance (statistics.R)

humaneval/ Human evaluation scripts to obtain results depicted in the paper (stats.py) and to test statistical significance (statistics.R)

Author: Thiago Castro Ferreira

Date: 15/12/2017