CGSum

code and dataset for AAAI 2021 paper: Enhancing Scientific Papers Summarization with Citation Graph

PYROUGE Installation

we recommend using the following commands to install the PYROUGE environment:

sudo apt-get install libxml-perl libxml-dom-perl
pip install git+git://github.com/bheinzerling/pyrouge
export PYROUGE_HOME_DIR=the/path/to/RELEASE-1.5.5
pyrouge_set_rouge_path $PYROUGE_HOME_DIR
chmod +x $PYROUGE_HOME_DIR/ROUGE-1.5.5.pl

You can refer to https://github.com/andersjo/pyrouge/tree/master/tools/ROUGE-1.5.5 for RELEASE-1.5.5 and remember to build Wordnet 2.0 instead of 1.6 in RELEASE-1.5.5/data\

cd $PYROUGE_HOME_DIR/data/WordNet-2.0-Exceptions/
./buildExeptionDB.pl . exc WordNet-2.0.exc.db
cd ../
ln -s WordNet-2.0-Exceptions/WordNet-2.0.exc.db WordNet-2.0.exc.db

DataSet SSN

The whole dataset and its corresponding citation relationship can be download through this link

example of our dataset:

{
  "paper_id": "102498304", # unique id of this paper
  "title":"Weak Galerkin finite element method for Poisson’s ...", # title of this paper
  "abstract":"in this paper , the weak galerkin finite element method for second order eilliptc   problems employing polygonal or  ...", # human written abstract
  "text":[
    ["The weak galerkin finite element method using triangulated meshes was proposed by .."],
     ["Let @inlineform1 be a partition of the domain Ω consisting of polygons in two dimensional"], 
    ...
  ] # body text, 
  "section_names": ["Introduction", " Shape Regularity",  ...] # corresponding section names to sections
  "domain":"Mathematic", # class label
}
...

You can download our preprocessed dataset which can be directly loaded by dataloader.py via SSN (inductive) and SSN (transductive). Note that we divide the dataset in two ways. The transductive division indicates that most neighbors of papers in test set are from the training set, but considering that in real cases, the test papers may from a new graph which has nothing to do with papers we used for training, thus we introduce SNN (inductive), by splitting the whole citation graph into three independent subgraphs – training, validation and test graphs. Our preprocessed datasets are chunked to 500 words, for full document you can retrieve them from the whole dataset by paper_id

你也可以通过百度云下载我们的数据集 SSN完整数据集和引用关系 提取码 v4u8

SSN inductive 提取码 gk4j

SSN transductive 提取码 17kw

requirements for running our code

• python 3.6+
• PyTorch 1.1
• DGL 0.4.3
• rouge 1.0.0
• pyrouge 0.1.3
• fastNLP 0.5.0+

Train and Test

Hyperparameters in the train.py/test.py script has been set to default, we also provide the example to run our code in train.sh and test.sh. you can train/test our model using the following command:

• training

python train_CGSum.py  --visible_gpu 0  --model_dir  save_models/CGSum_1hop  --dataset_dir  SSN/inductive --setting inductive --n_hop 1

python train_CGSum.py  --visible_gpu 0  --model_dir  save_models/CGSum_1hop  --dataset_dir  SSN/transductive --setting transductive --n_hop 1

• testing

python test_CGSum.py  --visible_gpu 0  --model_dir  save_models/CGSum_1hop  --model_name CGSum_inductive_1hopNbrs.pt --setting inductive  --decode_dir decode_path  --result_dir results --n_hop 1  --min_dec_steps 130

python test_CGSum.py  --visible_gpu 0  --model_dir  save_models/CGSum_1hop  --model_name CGSum_transductive_1hopNbrs.pt --setting transductive  --decode_dir decode_path  --result_dir results --n_hop 1  --min_dec_steps 140

To test our model , remember to replace the pyrouge root set in data_util/utils.py to your own path. you can also download our trained model to reproduce our results: inductive 1hop, inductive 2hop, transductive 1hop, transductive 2hop

our dataset is retrieved from S2ORC, the implementation of BertSum can refer to PreSumm, thanks for their works.