Hindi Fake News Detection and Abstractive Summarization

This project integrates Fake News Detection, Explainability, and Abstractive Summarization for Hindi news articles.

It uses BERT model to classify news as real or fake, explains its predictions using LIME, and generates a summary of the news article using the T5 model — all in one seamless pipeline.

---

Features

• Fake News Detection using fine-tuned BERT
• Real-time prediction: Real or Fake
• Model Explainability using LIME
• Abstractive Summarization using T5 Transformer
• Streamlit-based interactive UI
• Supports Hindi language inputs

---

Model Architecture

Task	Model
Classification	BERT
Explainability	LIME
Summarization	T5 Transformer

---

Dataset

• Fake News Detection: Custom Hindi dataset (balanced) containing real/fake labels
• Summarization: Hindi news article samples for training/testing the T5 model
• Datasets preprocessed using NLP techniques (cleaning, tokenization)

---

Tech Stack

• Python 3.13
• BERT (Classification)
• scikit-learn
• LIME (Explainability)
• PyTorch
• Pandas, NumPy
• T5 (Summarization)
• Streamlit (UI)

---

How It Works

1. User pastes a Hindi news article into the Streamlit UI.
2. The article is passed to a fine-tuned BERT model for classification.
3. If classified as real/fake, LIME explains the prediction via token-level highlights.
4. In parallel, a T5 model generates a concise summary of the article.

---

Contributors

This project was originally developed as part of a group academic project.

Original Contributors:

• Amil Gauri (Maintainer)
• Vikas Pandit
• Bhushan Salve
• Ajay Chaurasiya

Project restructured, documented, and maintained by Amil Gauri for public release.

---

License

This project is open-source under the MIT License.
You are free to use, modify, and distribute it with proper credit.

See the LICENSE file for full details.

---