graph-based-hs-cn

Exploring the usage of episodic knowledge graphs to create Counter-narratives (CN) for Hate Speech (HS)

Overview

This project uses the CONAN dataset. You can find the dat ain the data folder, and the preprocessing script in match_conan.py

Getting started

In order to run the code, follow these steps:

1. Create a virtual environment for the project (conda, venv, etc)

conda create --name graph-based-hs-cn python=3.8
conda activate graph-based-hs-cn

1. Install the required dependencies in requirements.txt

pip install -r requirements.txt --no-cache

1. Install the latest versions of the required cltl packages. We are using cltl.knowledgeextraction. Please clone the repositories, pull the latest versions and install the packages into the virtual environment like this:

conda activate graph-based-hs-cn
cd cltl-knowledgeextraction
git pull
pip install -e .

Usage

Authors

• Selene Báez Santamaría