![Page 1: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/1.jpg)
June 2017, @ZimMatthias
#MachineLearning #Blockchain
@EclipseScout
![Page 2: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/2.jpg)
![Page 3: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/3.jpg)
Eclipse Scout
![Page 4: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/4.jpg)
What is Eclipse Scout?
Business Appliation Framework
Open Source Eclipse Project
Based on Java and HTML5
Multi Device support, Modular Apps, …
Goals
Long Term Sustainability (enterprise apps live > 10 years)
Boosts Productivity (producing software in Switzerland …)
Easy to learn (newbies productive in 1-2 weeks)
![Page 5: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/5.jpg)
Application Model (Very) Long Term Clean Business Code High Maintainability
Eclipse Scout The Java Story
![Page 6: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/6.jpg)
Rendering HTML5, CSS3, JavaScript Styling & theming
Eclipse Scout The HTML5 Story
![Page 7: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/7.jpg)
Eclipse Scout Commercial Application
![Page 8: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/8.jpg)
Eclipse Scout Hello World
![Page 9: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/9.jpg)
Machine Learning Deeplearning4j
![Page 10: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/10.jpg)
![Page 11: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/11.jpg)
Deep Learning Library Open Source Java (most are Python) Good documentation
Features Full GPU support Distributed deep learning Runs with Hadoop + Spark
Deeplearning4j
https://github.com/deeplearning4j/deeplearning4j
![Page 12: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/12.jpg)
The ML «Hello World» Recognition of handwritten digits
1998 Gradient-based Learing for Document Recognition, Y. LeCun
![Page 13: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/13.jpg)
Handwritten Data
![Page 14: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/14.jpg)
Training of a Neural Network Model
![Page 15: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/15.jpg)
Error Rate and Training Epochs
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
1 2 3 4 5 6 7
Numbers (BSI Data) 9’000 Training images (18 writers) 1’000 Test images (2 writers)
![Page 16: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/16.jpg)
Error Rate and Training Epochs
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
1 2 3 4 5 6 7
Numbers only
MNIST + Numbers
Numbers + MNIST Start with MNIST Modell Train on Numbers 40% Error reduction
![Page 17: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/17.jpg)
Blockchain web3j/Ethereum
![Page 18: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/18.jpg)
Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds smart contracts Remarkable achievements and much hype
Main Challenges Scalability Privacy Regulatory & legal
Blockchain «Micro-Intro»
![Page 19: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/19.jpg)
What is it? Piece of byte code (usually written in Solidity) Is executed by the Ethereum Virtual Machine (EVM) Has an owner Examples Greeter (the Ethereum «Hello World») «Truly» autonomous cars
Ethereum Smart Contracts
![Page 20: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/20.jpg)
Ethereum «Hello World»
![Page 21: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/21.jpg)
Smart contract: To order car to transport people (by paying to contract) Smart contract: Car pays for energy/services
«Truly» Autonomous Cars
![Page 22: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/22.jpg)
Ethereum and Application Integration
Ethereum Client Geth/TestRPC/…
Ethereum Peer-to-Peer Network
Interface http://localhost:8545 JSON-RPC
JavaScript
web3 web3j
Java
![Page 23: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/23.jpg)
Library to interact with Ethereum (its peer-to-peer clients) Open Source Java (default is JavaScript) Good documentation
Features JSON-RPC client API implementation Tool to generate Contract Wrappers in Java
web3j
https://github.com/web3j/web3j
![Page 24: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/24.jpg)
Web3j: Generated Contract Wrapper
![Page 25: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/25.jpg)
PostgreSQL
Ethereum Client TestRPC
web3
Eclipse Scout Backend
web3j JDBC
Eclipse Scout UI (web application)
![Page 26: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/26.jpg)
![Page 27: #MachineLearning #Blockchain @EclipseScout...Blockchain Bitcoin started the field in 2009 Main Features: Cheap, fast, efficient (traditional setup: T+3 and high fees) Ethereum adds](https://reader035.vdocuments.mx/reader035/viewer/2022081409/60842fec3cab9e18922ec6df/html5/thumbnails/27.jpg)
Thanks! @ZimMatthias