Inspiration

Our inspiration for NotiScam was to help "under-represented demographics" (in terms of internet usage) use the internet more safely. This demographic includes the elderly and those with mental disabilities, who may not be as familiar with common red flags that indicate malicious intent on the internet.

What it does

NotiScam takes text and indicates whether it is spam or fake news (whichever is applicable) using cutting edge transformer neural networks for classification. It allows the user to realize whether the text inputted could be harmful to them, so they can safely navigate away from the page containing the text if needed.

How we built it

We built NotiScam's backend using Python, from the the robust Tensorflow library and accessory libraries for our AI classification models, to Flask for our API server. Our front end was built using up-to-date javaScript ES6, HTML, and CSS. Over multiple hours, we were able to build a streamlined front end capable of sending messages to our AI model, and receive its answers.

Challenges we ran into

Some of the challenges we endured and overcame were:

  • Networking - This was by far one of the most difficult things we faced near the end, it took hours to get the server and posts requests working together. Finally we figured out how to send requests to servers of our own design.

  • Formatting - The format of how multiple aspects our code would be written was something difficult to manage, but ultimately we combined our code efficiently and in a cohesive manner.

    Accomplishments that we're proud of

    The biggest accomplishments we have made are:

  • Using Artificial Intelligence to accurately determine whether or not a news article was fake or true.

  • Using Artificial Intelligence to correctly summarize a news article given to it.

    What we learned

    We learned a ton in this project, here's a few:

  • We learned how to set up servers in python, using flask.

  • We learned how to create, and train Artificial Intelligence models using tenser-flow.

What's next for NotiScam

To continue NotiScam, we plan to improve our neural network models even more, in hopes of gaining even higher accuracy and versality with our models' predictions.

Built With

Share this project:

Updates