Inspiration
A paper called “Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings”
What it does
Learns the biased model weights that got from the biased data and regularizes the data points for that particular type of instance from the entire dataset by projecting them to a neutral axis.
How we built it
Python and Machine Learning techniques
Challenges we ran into
Initiating the Idea and Server for training the model
Accomplishments that we're proud of
Have completed the challenge using Machine learning way
What we learned
UI and Display approach is as important as the backend, we didn't get time to properly organize the slides and showcase the actual work done using diagrams.
What's next for AWS-Challenge
Can reduce bias in any model and rebuild the datasets with reduced bias.
Log in or sign up for Devpost to join the conversation.