Cooperative Learning of Disjoint Syntax and Semantics

This code was used to obtain the results described in the paper:

ListOps

Download ListOps dataset. URLs of the original dataset and an extrapolation test set can be found in data/listops/external/urls.txt file. Download each URL using your web browser; wget won't work.
Run python listops/data_preprocessing/split.py to split the dataset into the train, the valid and the test sets. Make sure that you have downloaded the dataset and the files are present in the data/listops/external folder. This should produce some files in a directory data/listops/interim.
Build vocabulary using python -m listops.data_preprocessing.build_vocab. This should produce a vocab.txt file in data/listops/processed.
Run python -m listops.ppo.train_ppo_model or python -m listops.reinforce.train_reinforce_model to train the model with PPO or REINFORCE estimators.

Run python sst/ppo/train_ppo_model.py to train the model using SST-2 or SST-5 datasets.

Download SNLI and MultiNLI datasets and extract corresponding archives to data/nli folder. URLs can be found in data/nli/external/urls.txt file.
Run python nli/data_preprocessing/preprocess.py to preprocess dataset and generate vocabulary files.
Run python nli/ppo/train_ppo_model.py to train the model using SNLI or MultiNLI datasets.

The code is tested with Python 3.6.3 and PyTorch 1.0.1.

Latent-TreeLSTM is MIT licensed. See the LICENSE file for details.