Hend-Khaled-Aly/AI-vs-Human

A Natural Language Processing (NLP) project that compares Bidirectional RNNs and Transformer models to classify text as AI-generated or human-written. Developed as part of the Machine Learning for Data Analysis program under the Digital Egypt Youth initiative. # AI vs Human: An NLP Project This repository hosts the **AI vs Human** project, created as a capstone for the **Machine Learning for Data Analysis** track under the **Digital Egypt Youth** initiative. The project investigates two advanced Natural Language Processing (NLP) techniques to distinguish between text written by humans and that generated by AI. ## Project Summary As AI-generated text becomes increasingly indistinguishable from human writing, the need for reliable detection methods grows. This project tackles that problem by implementing and comparing two key NLP strategies: 1. **Bidirectional Recurrent Neural Networks (BRNNs)** 2. **Transformer Models** Two dedicated Jupyter Notebooks are provided, each showcasing the complete pipeline—from data preparation to model training and evaluation. ## Repository Structure * `BRNNs.ipynb`: Demonstrates the classification approach using Bidirectional Recurrent Neural Networks. * `Transformers.ipynb`: Explores the Transformer-based model for identifying AI vs human-written text. ## Highlights * **Text Preprocessing**: Comprehensive text cleaning and formatting to support model input requirements. * **Training & Evaluation**: Step-by-step model training with detailed evaluation metrics for comparison. * **Performance Visualization**: Graphs and plots illustrating model performance (e.g., accuracy, loss). * **Modular Codebase**: Clean, well-documented code to support easy customization and learning. ## Setup & Dependencies To run the notebooks and reproduce the results, ensure the following tools and libraries are installed: * Python 3.7 or higher * Jupyter Notebook * Essential Python Libraries: * TensorFlow or PyTorch * Hugging Face Transformers * NumPy * Pandas * Matplotlib or Seaborn