Fakefilter

The News Shouldn't Have the Power to Manipulate Us! Fakefilter validates a claim based on your question by testing it against multiple reliable news sources.

How we built it

Fakefilter is written 100% in Python using HTML request and scraping libraries, and an ASCII GUI. Within the process of validation, semantic sentence comparisons came up. We didn't want to use a convoluted net and wanted to create a novel solution that used little compute power. This resulted in the final solution of using subsets and subset ratios.

Accomplishments that we're proud of

We implemented fairly robust NLP model into a very lightweight Python app to open up the field of fact-checking. In this political climate, fact-checking will continue to become more and more important.

What we learned

We discovered a lot of new Python libraries in a very short time, and worked with natural language processing for the first time. Since NLP can be used for so many things, this is exciting!