Multimodal Misinformation Detection Across Social Media Platforms

This repository contains the implementation and experiments conducted on building a scalable misinformation detection system using DistilBERT and multiple real-world fake news datasets.
The project aims to detect misinformation by leveraging textual content (news body + headline) across different social media news sources while maintaining efficiency and scalability using distilled transformer models.
Built a lightweight and optimized deep learning model using transformer-based encoders and PyTorch, achieving 80% accuracy while reducing computational load for deployment in low-resource environments.
Fine tuned BERT for better results.

Name		Name	Last commit message	Last commit date
Latest commit History 42 Commits
DistilBert.ipynb		DistilBert.ipynb
DistilBert_on_gossipcop.ipynb		DistilBert_on_gossipcop.ipynb
Distil_Text+Headline_1ep_10K_Dataset.ipynb		Distil_Text+Headline_1ep_10K_Dataset.ipynb
Distil_Text+Headline_2ep_5K_Dataset.ipynb		Distil_Text+Headline_2ep_5K_Dataset.ipynb
GossipCopDistil_Text+Headline_Dataset.ipynb		GossipCopDistil_Text+Headline_Dataset.ipynb
GossipCop_Dataset_URL_Extraction.ipynb		GossipCop_Dataset_URL_Extraction.ipynb
PoliDataSet.ipynb		PoliDataSet.ipynb
PolitifactDistil_Text+Headline_Dataset.ipynb		PolitifactDistil_Text+Headline_Dataset.ipynb
README.md		README.md
bert.py		bert.py
gossipDistil.ipynb		gossipDistil.ipynb
gossip_on_ISOTDataDistil.ipynb		gossip_on_ISOTDataDistil.ipynb
gossip_on_poliDataDistil.ipynb		gossip_on_poliDataDistil.ipynb
poliDataDistil.ipynb		poliDataDistil.ipynb
poli_on_ISOTDataDistil.ipynb		poli_on_ISOTDataDistil.ipynb
poli_on_gossipDataDistil.ipynb		poli_on_gossipDataDistil.ipynb

Provide feedback