Yet another PyTorch implementation of the model described in the paper An Unsupervised Neural Attention Model for Aspect Extraction by He, Ruidan and Lee, Wee Sun and Ng, Hwee Tou and Dahlmeier, Daniel, ACL2017.

NOTA BENE: as of December 2021, gensim>=4.1.0 and hydra-core>=1.1.0 are required.

For a working example of a whole pipeline please refer to example_run.sh

Let's get some data:

wget http://snap.stanford.edu/data/amazon/productGraph/categoryFiles/reviews_Cell_Phones_and_Accessories_5.json.gz
gunzip reviews_Cell_Phones_and_Accessories_5.json.gz    
python3 custom_format_converter.py reviews_Cell_Phones_and_Accessories_5.json

Then we need to train the word vectors:

python3 word2vec.py reviews_Cell_Phones_and_Accessories_5.json.txt

And run

python main.py model.aspects_number=35 data.path=$DATA_NAME.json.txt model.log_progress_steps=1000

Please see all passable params in the configs/ directory. For a working example of a whole pipeline please refer to example_run.sh

I acknowledge the implementation is raw, code modification requests and issues are welcome.

@alexdevmotion

• Evaluation: PMI, NPMI, LCP, L1/L2/coord/cosine (Nikolenko SIGIR'16), ...
• Aspects prediction on text + visualization
• Saving the model, aspects, etc.

Greatly appreciated, if you use this work.

@misc{abaepytorch2019alekseev,
  title     = {{alexeyev/abae-pytorch: ABAE, PyTorch implementation}},
  year      = {2019},
  url       = {https://github.com/alexeyev/abae-pytorch},
  language  = {english}
}

• Reference Keras implementation
• Updated Keras version, compliant with Keras 2.x and Python 3.x.