DeepVerse, Machine Learning on Blockchain

Team photo, Fred (Skipper), Lupeng (First mate), Siyu (Crew)

Inspiration

Bridging digital world and real world

What it does

DeepVerse is to integrate machine learning and blockchain to do various useful things without being explicited programmed to and yet remain open, transparent, and decentralized. It leverages oracles and APIs to perform complex machine learning algorithms which is unable to do on the chain, and bring the verifiable results from the heavy calculation back to the blockchain.

How we built it

https://github.com/F14r3/deepverse-chainlink

The 5 minutes video gives an overview of it. Some supplementary information is also provided from 5' to 6'37.

Challenges

The blockchain field is similar to the fast growing machine learning area, with zillions of versions flying around, the compatibility issue is concerning as I spend half of the time to making things work in the same environment.

Accomplishments that we're proud of

We are the first to train machine learning models collaboratively on blockchain while protecting users' privacy in 2019, and now we believe we are the first to create the link to solve the last-mile problem where blockchain falls down.