EquiScan

Inspiration

Everyday when I go to school, there are always kids saying something or the other without really knowing the meaning. For example they could say, "You dont look indian." Or maybe, "Are you having curry for lunch?" These are what you call microagressions. This really affected me especially in early middle school and made me feel hurt. This is why I built the Bias and Microagression detector.

What it does

It checks sentences or paragraphs and finds the underlying bias, toxicity, and microagressions.

How we built it

Using a prexisting hugging face model (Toxic-Bert), I imported it in. What this model does is recognizes the negative tones and words and gives it a pecentage reating based on that. From there I added if and elif statements to look for microagressions. Along with that I used Tkinker to make a GUI with python as well.

Challenges we ran into

Since I am a begginer at coding, it took me a while to figure out how to get the microagressions working, and how to integrate it with the prexisting model.

Accomplishments that we're proud of

I am proud of figuring out all the syntax of the GUI sotware on the spot.

What we learned

I learned a lot about coding skills, how AI models get trained, and how bias applies in our whole world.

What's next for EquiScan

First, I will add to the GUI and add more features. Next, I will make it a app on the app store and allow users to screenshot or paste text into to scan it for bias, toxicity and microagressions. Finally, I can make it a browser extension and make it so that when users are on news websites, or political forums it will scan and alert them if bias is found.

Built With

• huggingface
• python
• vscode