Runtime Terrors - Fake News Detector

• 

  Accurate Fake News Classification

Inspiration

While searching for an ML project to integrate into a web interface, I came across an article suggesting fake news detection. For there I was inspired to build a version of our own that goes beyond. Including full article detector, and displays the articles bias.

What it does

Determines if an article's headline in the users active tab is fake news.

How we built it

Using React and Bootstrap on the front we created an extension for the Chrome platform. On the backend we used Node.js, Python, and handled the detection feature using Google Gemini's api.

Challenges we ran into

Like all projects in their infancy we dreamed a little too big for what could be accomplished in 48 hours. While implementing the headline detector, we discovered that trying to detect an entire article would prove to be very expensive. And prior to this we had some trouble grabbing all of the text data from the active tab.

Accomplishments that we're proud of

We are proud that we completed the projects core features.

What we learned

We learned about integrating Gemini's api to a website using node.js.

What's next for Runtime Terrors - Fake News Detector

In the future we plan to improve our UI, allow for account creation and a subscription service. The subscription service would include full article detection, and display the articles bias. And a dedicated model trained of verified examples of real and fake headlines.

Built With

• gemini
• miro
• node.js
• python
• react