We’ve been divided politically, socially, culturally and economically by technology platforms like Facebook, Twitter and Reddit. It’s hard for us to step out of our biases and understand differing perspectives. And so we built a platform with machine learning (NLP doc2vec models) to solve this.
What it does
Logos is a web platform for having discussions based on common interests. After filling out their descriptions, users see a feed of conversation posts created by other users and personalized with machine learning. From there, users can start a conversation they care about (e.g. income inequality, programming, philosophy) with another user.
How we built it
The frontend is built with React.js and hosted on AWS S3. The real-time chat is built with Firebase Cloud Messaging. The backend is built with Python Django, and our machine learning doc2vec are built in Gensim (NLP) with text processing done in Google Cloud Platform.
Challenges we ran into
Integrating the doc2vec machine learning model. Implementing the Firebase Cloud Messaging
Accomplishments that we’re proud of
We’re happy that our doc2vec machine learning model can rank the conversation posts based on their similarities to the user’s descriptions. We are also happy that we had time to integrate Firebase Cloud Messaging.
What we learned
We learned how to apply machine learning to text similarity and create real-time chat between users with Firebase Cloud Messaging.
What's next for LOGOS
Polishing the design more Checking that the latency is near real-time for webpage loads Integrating FCM more directly into the app