Decentralized Machine Learning

The DML Infrastructure

On-Device Machine Learning without Data Extraction

Utilize untapped private data in individual devices for machine learning with privacy protected

Unleash Innovation from Periphery

Unlock innovation by creating a developer community and competitions in algorithm marketplace

Connect Idle Processing Power of Billions Devices

Leverage processing power of all connected devices for running machine learning algorithms

Mass Participation in Machine Learning Training

Create an algorithm trainer community to improve algorithms through collective efforts

Multi-Blockchain Adoption and Interoperability

Communicate across multi-chains to achieve blockchain-agnostic protocol

Return Power to Ecosystem Owners

Avoid centralization and control by a few oligopolies via blockchain and tokenization

DML Prototype Demo

Key Features of DML Prototype:

Searching, requesting and listing for machine learning algorithms in DML Algo Marketplace;
Creation of machine learning application requests and selection of relevant individual device owners for private data access;
Deployment of machine learning algorithms into smart devices;
Running of the machine learning algorithms with idle processing power in the DML App;
Submission of machine learning results (and raw data remains within the devices) from the DML App;
Generation of aggregated machine learning prediction analytics reports;
Smart contract transactions in DML Tokens for machine learning models.

Read Our One Pager

Read Our Explainer Deck

Roadmap

Feb 2016

Google published the research paper on federated learning
Mar 2016

AlphaGo beat Lee Sedol in Go
Mar 2017

Idea generation of decentralization in machine learning
Apr 2017

Google published research blog in federated learning
May 2017

Development of proof of concept
Sep 2017

Idea generation of decentralization in algorithms
Dec 2017

Whitepaper published and DecentralizedML.com online
Feb 2018

Release of DML Protocol Gen 0 (DML Algo Marketplace) Prototype
Apr 2018

Token Generation Event and Launch of DML Protocol Gen 0 (DML Algo Marketplace) Beta
Apr-May 2018

DML Algo Marketplace online
Jun-Jul 2018

Release of DML Protocol Gen 1 alpha (decentralized machine learning on-device private data)
Jun 2018

Research of state channels for increasing DML scalability
Jul-Aug 2018

First DML Algo competition to grow and support developers’ community
Jul-Aug 2018

Release of DML Protocol Gen 1 beta
Sep 2018

DML Protocol Gen 1 online
Dec 2018

Release of customized state channels for increasing DML scalability
Q1 2019

Release of DML Protocol Gen 2 beta (decentralized machine learning on-device private data with third-party service and data access)
Q1 2019

Research of multi-chain support and interoperability
Q2 2019

DML Protocol Gen 2 online
Jul-Aug 2019

Release of DML Protocol Gen 3 beta (decentralized machine learning on-device private data with third-party service and data access and mobile sensors/ IoT connection capability)
Sep-Oct 2019

DML Protocol Gen 3 online
Q4 2019

Research of general purpose API start for expanding usage of DML marketplace from machine learning to general applications
Q1 2020

Research of new blockchain supporting mass adaption of general purpose decentralized applications and data privacy
Q2-Q3 2020

Release of DML Protocol Gen 4 beta (Support deployment of general applications)
Q4 2020

DML Protocol Gen 4 online