Adversarial Decomposition of Text Representation

The code for the paper "Adversarial Decomposition of Text Representation", NAACL 2019 https://arxiv.org/abs/1808.09042

Installation

1. Clone this repo: https://github.com/text-machine-lab/adversarial_decomposition.git
2. Install NumPy: pip install numpy==1.16.3
3. Install PyTorch v1.1.0: pip install https://download.pytorch.org/whl/cu100/torch-1.1.0-cp37-cp37m-linux_x86_64.whl (for python3.6, use pip install https://download.pytorch.org/whl/cu100/torch-1.1.0-cp36-cp36m-linux_x86_64.whl)
4. Install dependencies: pip install -r requirements.txt
5. Download spacy models: python -m spacy download en_core_web_lg

Initial setup

1. Create dir mkdir -p data/experiments
2. Create dir mkdir -p data/datasets
3. Create dir mkdir -p data/word_embeddings
4. Download the Shakespeare data: git clone https://github.com/cocoxu/Shakespeare.git data/datasets/shakespeare
5. Download the Yelp data: git clone https://github.com/shentianxiao/language-style-transfer.git data/datasets/yelp
6. Download the pickled GloVe embeddings wget https://mednli.blob.core.windows.net/shared/word_embeddings/glove.840B.300d.pickled -O data/word_embeddings/glove.840B.300d.pickled
7. Download the pickled fastText embeddings wget https://mednli.blob.core.windows.net/shared/word_embeddings/crawl-300d-2M.pickled -O data/word_embeddings/crawl-300d-2M.pickled

Running the code

Global constants are set in the file settings.py. In general, you don't need to change this file. Experiment parameters are set in the config.py file.

First, run the preprocessing script: python preprocess.py This scipt will print the ID of the preprocessing experiment, for example preprocess.buppgpnf. Copy this ID and change parameter preprocess_exp_id of the TrainConfig class on the line 12 in the file config.py file accordingly.

After you set the preprocess experiment id, run the training: python train.py. This scirpt will also print the ID of the training experiment. You can paste it in the eval_generation.ipynb notebook to play with the model.

Chaning the form and meaning

The provided eval_generation.ipynb notebook shows how to use the model to swap the meaning and form vectors of the input sentences!

Citation

If you find this code helpful, please consider citing our paper:

A. Romanov, A. Rumshisky, A. Rogers, D. Donahue,Adversarial decomposition of text represen-tation, In Proceedings of NAACL 2019: Conference of the North American Chapter of the Association for Computational Linguistics, 2019

https://arxiv.org/abs/1808.09042

@inproceedings{romanov2019adversarial,
  title={Adversarial Decomposition of Text Representation},
  author={Romanov, Alexey and Rumshisky, Anna and Rogers, Anna and Donahue, David},
  booktitle={Proceedings of NAACL 2019: Conference of the North American Chapter of the Association for Computational Linguistics},
  year={2019}
}