sm_cnn

SM model

References:

1. Aliaksei _S_everyn and Alessandro _M_oschitti. 2015. Learning to Rank Short Text Pairs with Convolutional Deep Neural Networks. In Proceedings of the 38th International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR '15). ACM, New York, NY, USA, 373-382. DOI: http://dx.doi.org/10.1145/2766462.2767738

Setup

Clone and create the dataset:

git clone https://github.com/castorini/data.git
git clone https://github.com/castorini/Castor.git

You should you see the following tree:

.
├── Castor
│   ├── README.md
│   ├── baseline_results.tsv
│   ├── idf_baseline
│   ├── kim_cnn
│   ├── mp_cnn
│   ├── setup.py
│   ├── sm_cnn
└── data
    ├── GloVe
    ├── ParagramEmbeddings
    ├── README.md
    ├── SimpleQuestions_v2
    ├── TrecQA
    ├── WikiQA
    ├── msrvid
    ├── requirements.txt
    ├── sick
    ├── twitterPPDB
    ├── utils
    └── word2vec

Parse the TrecQA datset:

cd ../../data/TrecQA/
python parse.py
cd -

Parse the WikiQA datset:

cd ../../data/WikiQA/
unzip WikiQACorpus.zip
python create-train-dev-test-data.py 
cd -

Your repository root should be in your PYTHONPATH environment variable:

export PYTHONPATH=$(pwd)

To create the dataset:

cd Castor/sm_cnn/
./create_dataset.sh

We use trec_eval for evaluation:

cd ../utils/
./get_trec_eval.sh
cd ../sm_cnn

Training

Download the word2vec model from here and copy it to the data/ folder.

You can train the SM model for the 4 following configurations:

1. random - the word embedddings are initialized randomly and are tuned during training
2. static - the word embeddings are static (Severyn and Moschitti, SIGIR'15)
3. non-static - the word embeddings are tuned during training
4. multichannel - contains static and non-static channels for question and answer conv layers

To train on GPU 0 with static configuration:

python train.py --mode static --gpu 0

NB: pass --no_cuda to use CPU

The trained model will be save to:

saves/static_best_model.pt

Testing the model

python main.py --trained_model saves/TREC/multichannel_best_model.pt 

Evaluation

The performance on TrecQA dataset:

TrecQA:

Best dev

Metric	rand	static	non-static	multichannel
MAP	0.8096	0.8162	0.8387	0.8274
MRR	0.8560	0.8918	0.9058	0.8818

Test

Metric	rand	static	non-static	multichannel
MAP	0.7441	0.7524	0.7688	0.7641
MRR	0.8172	0.8012	0.8144	0.8174

WikiQA:

Best dev

Metric	rand	static	non-static	multichannel
MAP	0.7109	0.7204	0.7049	0.7245
MRR	0.7169	0.7234	0.7075	0.7259

Test

Metric	rand	static	non-static	multichannel
MAP	0.6313	0.6378	0.6455	0.6476
MRR	0.6522	0.6542	0.6689	0.6646

NB: The results on WikiQA are based on the SM model hyperparameters.

To create your own word2vec.pt file

• Download word2vec from here to the data/ folder

python $PYTHONPATH/utils/build_w2v.py --input data/aquaint+wiki.txt.gz.ndim=50.bin

Note that $PYTHONPATH holds the location of the repository root.