Yuer867 / Replication-and-Extension-of-MAC

This is the final project of DS-GA 1011: Natural Language Processing with Representation Learning

Replication-and-Extension-of-MAC

This is the repository for final project of DS-GA 1011: Natural Language Processing with Representation Learning

Replication and Extension of Hierarchical Multi-head Attentive Network for Evidence-aware Fake News Detection

Usage

1. Install required packages

We use Pytorch 1.13.0 and python 3.8.6.

nltk==3.7
numpy==1.23.5
scikit_learn==1.1.3
torch==1.13.0
tqdm==4.64.1
transformers==4.24.0

2. Data Preprocessing

1. download pre-trained word embeddings file

We use glove.840B.300d (Downloading) with a dimension of 300 as pre-trained word embeddings.

2. convert json file into pkl

For PolitiFact dataset, run

python preprocess.py --dataset="PolitiFact" --embeddings_file="glove.840B.300d.txt"

For Snopes dataset, run

python preprocess.py --dataset="Snopes" --embeddings_file="glove.840B.300d.txt"

For FakeNewsNet+ dataset, run

python preprocess_tweet.py --dataset='FakeNewsNet_plus' --embeddings_file="glove.840B.300d.txt"

3. Running experiment

For PolitiFact dataset, run

python mac.py --dataset="PolitiFact" --use_post_sources --use_article_sources --mac_nhead_1=3 --mac_nhead_2=1

For Snopes dataset, run

python mac.py --dataset="Snopes" --use_post_sources --use_article_sources --mac_nhead_1=5 --mac_nhead_2=2

For FakeNewsNet+ dataset, run

python mac_tweet.py --dataset='FakeNewsNet_plus'