Fake news detector

The brief

A fact-checking group wants help sorting the flood of articles shared on social media. Build a classifier that flags likely fake news, test it on topics it was not trained on, and explain its decisions. Good looks like: honest results on unseen topics, explanations, and a discussion of the risks of automated fact-checking.

What you will build

1. Explore the articles and labels
2. Build a text-classification baseline
3. Fine-tune a language model
4. Test on topics not seen in training
5. Explain decisions and discuss risks

Free, open data to build it with

• Getting Real about Fake News ↗ Kaggle
• News Category Dataset ↗ Kaggle

Check each dataset's licence before you publish your work. Your mentor gives your team its own version of the brief (data, scope or region), so every team's project is its own.