EECS595 project: Improving Verifiability of TRIP with Data Augmentation and Graph Neural Networks (Group28)

This repo is cloned from Verifiable-Coherent-NLU and modify some codes to add two new approaches.

Getting Started

The data augmentation result can be reproduced using the jupyter notebook TRIP_Data_Augmentation.ipynb, which we ran in Colab with Python 3.7.

Incorporating LMs and GNNs

The conda virtual enviroment can be installed by the following commands:

conda env create -f trip_ours_env.yml

Graph construction and data preprocessing

Data preparation: To perform dependency parsing, we utilize the Stanford coreNLP dependency parser, and necessary files can be downloaded from https://drive.google.com/drive/folders/148bfSBczJhcHpgtPz98LA8am0MJAfTUW?usp=sharing (stanford-corenlp-4.2.2-models-english.jar and stanford-corenlp-4.2.2.zip). The ConceptNet Numberbatch embedding can also be downloaded from https://drive.google.com/drive/folders/148bfSBczJhcHpgtPz98LA8am0MJAfTUW?usp=sharing (numberbatch-en-19.08.txt). As the graph construction process takes several hours, we have provided the preprocessed TRIP data in https://drive.google.com/drive/folders/148bfSBczJhcHpgtPz98LA8am0MJAfTUW?usp=sharing (tiered_dataset.pickle). To preprocess the raw TRIP data, run the following commands:

python data_preprocessing_graph.py --drive_path DRIVE_PATH --save_pkl_path PATH_TO_SAVE_PICKLE_FILE --cn_nb_path numberbatch-en-19.08.txt --jar_path stanford-corenlp-4.2.2/stanford-corenlp-4.2.2.jar --models_jar_path stanford-corenlp-4.2.2-models-english.jar

The GNN result can be reproduced by the following commands:

python train_test_trip.py --drive_path DRIVE_PATH --pkl_file_path tiered_dataset.pickle --cn_nb_path numberbatch-en-19.08.txt

Python Dependencies

The required dependencies for Colab are installed within the notebook, while the exhaustive list of dependencies for any setup is given in requirements.txt. Out of these, the minimal requirements can be installed in a new Anaconda environment by the following commands:

conda create --name tripPy python=3.7
conda activate tripPy
pip install torch==1.7.1 torchvision==0.8.2 torchaudio==0.7.2
pip install transformers==4.2.2
pip install sentencepiece==0.1.96
pip install deberta==0.1.12
pip install spacy==3.2.0
python -m spacy download en_core_web_sm
pip install pandas==1.1.5
pip install matplotlib==3.5.0
pip install progressbar2==3.38.0
pip install ipykernel jupyter ipywidgets # For Jupyter Notebook setting