This repository contains the code for the UCL (Unbiased and Fair Toxic Speech Detection) project, which is intended for submission to the ICDE 2025 conference.
UCL is a state-of-the-art toxic speech detection system that addresses both explicit and implicit biases in text data. The system aims to improve the accuracy and fairness of toxic speech detection by employing advanced debiasing techniques.
- Bias Mitigation: Reduces both explicit and implicit biases in toxic speech detection.
- High Accuracy: Utilizes cutting-edge algorithms to achieve high accuracy in identifying toxic content.
- Fairness: Ensures equitable treatment of different demographic groups on social platforms.
- Scalability: Designed to handle large-scale datasets and real-time detection.
To install the UCL project, follow these steps:
-
Clone the repository:
git clone https://github.com/Tislab-NLP/UCL.git cd UCL
-
Create a virtual environment:
python3 -m venv venv source venv/bin/activate # On Windows use `venv\Scripts\activate`
-
Install the required dependencies:
pip install -r requirements.txt
Train the model:
bash python main.py --config configs/train_config.yaml
- Python 3.8 or higher
- pip
- Additional requirements are listed in
requirements.txt
This project is licensed under the MIT License. See the LICENSE file for details.