The Equidistant Logo

The Equidistant Project - Unbiased News Platform

screenshot The Problem: Misinformation and bias in news reporting have become critical issues in today's world. Access to accurate and unbiased news is essential for an informed and enlightened public discourse. Traditional news outlets often struggle to provide truly objective information.

Our Solution: The Equidistant Project is a comprehensive platform that addresses these challenges head-on. Our project combines data scraping, machine learning, and a scalable backend infrastructure to create a reliable source of unbiased news. We aim to empower users to access and interpret factual news in an era filled with misinformation.

Key Features: flow-diagram

Data Scraping: We scrape a diverse range of news articles from sources like Fox News and CNN.
Text Embeddings: Using advanced NLP models (GPT-3 Ada), we generate text embeddings to capture the factual essence of articles.
Cosine Similarity: We calculate cosine similarity to pair articles with similar content, helping users access multiple perspectives on a topic.
Instruct Model: We employ GPT-3.5-turbo-instruct, fine-tuned using Reinforcement Learning with human feedback, to generate balanced and unbiased articles.
Frontend: Our frontend, built with Next.js, Tailwind CSS, Algolia, and React Server Components, provides an intuitive user interface.
Backend: Our backend uses MongoDB Atlas, Algolia, FastAPI, and Cloudflare Workers to store, serve, and search for articles on an international scale.

Tech Stack:

Frontend: Next.js, Tailwind CSS, Algolia, React Server Components
Backend: FastAPI, MongoDB Atlas, Algolia, Cloudflare Workers
NLP Models: GPT-3 Ada, GPT-3.5-turbo-instruct

Deployment Architecture:

Our deployment architecture is designed for scalability and reliability:

Frontend: Hosted on our server in California. Scaled with Cloudflare.
Backend: Deployed on a scalable server infrastructure, powered by Cloudflare Workers, to ensure high availability and performance on an international scale.
Database: MongoDB Atlas, a cloud-based database service, provides scalability and reliability for storing articles.

Why It Matters:

Fights Misinformation: By generating unbiased articles, we help combat the spread of misinformation.
Encourages Informed Discourse: Our platform empowers users to make well-informed decisions by providing multiple perspectives on news topics.
Scalable Worldwide: The Cloudflare Workers-backed infrastructure ensures global scalability.

The Equidistant Project is more than just a hackathon project; it's a commitment to a more informed and unbiased future. Join us in the fight against misinformation!

What's next?

WebSockets: Real-time updates will keep our users informed with the latest news as it unfolds. By implementing WebSockets, we enable a live, interactive experience, making our platform more engaging and up-to-date.
Internationalization and Localization: As a global platform, we aim to offer our services in multiple languages to cater to a diverse user base. We'll work on translating our interface and providing news articles in various languages, adapting to the cultural and regional nuances.
Machine Learning Feedback Loop: As our platform grows, we will collect more data which will be used to further train and refine our machine learning models. This feedback loop will help us continually improve the quality of our news aggregation and summarization processes.
Mobile Applications: To reach a wider audience, we will develop mobile applications for both iOS and Android. These applications will bring our unbiased news platform to users’ fingertips wherever they are.