Confidential Health Predictions using Blind AI

[pbkompasz]

Show and tell project type

Builder Bounty Submission

Github Repo Link

https://github.com/pbkompasz/blind-ai

Video Walkthough Link

https://web-nu-eight-27.vercel.app/

Project Description

Our model was train on the Heart Failure Prediction Dataset to predict the possibility of developing heart disease based on different inputs e.g. age, sex, bio-markers, etc.

What problems does your project solve? How does it preserve privacy for users?

These are sensitive information, which a user might not want to share on any web platform. The solution to this is using Blind AI. The model is trained from a public dataset and the health predictions are confidential, without revealing a user's presonal information.

How does the project use Nillion? Describe and link to any Nada programs

Deployment link: https://web-nu-eight-27.vercel.app/
A logistic regression model was trained on the Heart Failure Prediction Dataset.
This model was uploaded to the Nillion Testnet (3rgqxWd47e171EUwe4RXP9hm45tmoXfuF8fC52S7jcFoQTnU8wPiL7hqWzyV1muak6bEg7iWhudwg4t2pM9XnXcp/heart-disease) and through a web interface a user can input their private information and get health predictions.

Is there anything else you want to share?

I tried to implement different models but the compilation time took too long. After 1 hour of compilation of a small neural network I had to stop it.

Optional - Link your project and team members' social handles

No response

Optional - Team ETH Address(es)

0x26BfbD8ED2B302ec2c2B6f063C4caF7abcB062e0