machine learning and analytics in the cloud on iot data

28
1 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners. 1 Machine Learning and Analytics in the Cloud on IoT Data Ramesh Peri Sr. Principal Engineer Intel Corporation Austin, TX 78746

Upload: intel-software

Post on 21-Jan-2018

179 views

Category:

Technology


4 download

TRANSCRIPT

Page 1: Machine Learning and Analytics in the Cloud on IoT Data

1 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.1

Machine Learning and Analytics in the Cloud on IoT Data

Ramesh PeriSr. Principal EngineerIntel CorporationAustin, TX 78746

Page 2: Machine Learning and Analytics in the Cloud on IoT Data

2 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

• A typical IoT System

• Characteristics of IoT Data & Machine Learning techniques

• Intel IoTDevkit Overview

• A Complete end-to-end IoT Solution

Agenda

Page 3: Machine Learning and Analytics in the Cloud on IoT Data

3 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

IoT System

Page 4: Machine Learning and Analytics in the Cloud on IoT Data

4 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

A Real IoT System

Solar panels Power Mains

Envoy gatewayfrom enphase

Wifi router and cable modem

Austin energy power meter

Enphase mobile app

Enphase website

Austin energy website

Page 5: Machine Learning and Analytics in the Cloud on IoT Data

5 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

• A typical IoT System

• Characteristics of IoT Data & Machine Learning techniques

• Intel IoTDevkit Overview

• A Complete end-to-end IoT Solution

Agenda

Page 6: Machine Learning and Analytics in the Cloud on IoT Data

6 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

Characteristics of IoT Data

Streaming Time Series Data

Inference results in some real-time actuation

Data is Correlated

Lack of training data or Highly Imbalanced data

Page 7: Machine Learning and Analytics in the Cloud on IoT Data

7 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

Machine Learning Techniques for IoT Data

• Rule Based SystemsRB Systems

• One class Support Vector MachinesOCSVM

• Probabilistic Exponentially Weighted Moving AveragePEWMA

Markov Models

….

• Used to predict sequences of events based on historical data

Page 8: Machine Learning and Analytics in the Cloud on IoT Data

8 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

• A typical IoT System

• Characteristics of IoT Data & Machine Learning techniques

• Intel® IoTDevkit Overview

• A Complete end-to-end IoT Solution

Agenda

Page 9: Machine Learning and Analytics in the Cloud on IoT Data

9 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

Intel® IoT Platforms in Market today

Name Processor/SoC Target Classification

Galileo (X1000) Quark Maker SBC

Edison Atom Maker SBC

D1000 x86 IoT Edge MCU

D2000 x86 IoT Edge MCU

C1000 Curie (x86 + Arc) IoT Edge MCU++

Arduino101 Curie (x86 + Arc) Maker MCU++

Minnow Board Atom Reference SBC

Joule Atom Maker++/IoT SBC

Gateways from PC OEMs Atom IoT Gateway Gateway

DefinitionsSBC – Single Board Computer with exposed I/OsMCU – MicroControllerMCU++ - MicroController but with more stuffGateway – small form factor PC with no exposed I/Os

Page 10: Machine Learning and Analytics in the Cloud on IoT Data

10 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

• Intel® System Studio IoT Edition

• Intel ® XDK IoT Edition

• Wind River* Helix* Cloud

• MCU development tools

• C, C++, JavaScript, Python, and Java programming language support

Intel® IoTDevKit Overview

Supports Intel® IoT Gateways, Compute Modules, and MCUs with Middleware libraries, tools, and cloud connectors

IoT Cloud and Edge Analytics for data collection, visualization

Page 11: Machine Learning and Analytics in the Cloud on IoT Data

11 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

Intel® IoTDevkit Goals

Good OOB experience

Support for Sensors/Peripherals

Uniform Dev Environment

Robust Developer Ecosystem

Cloud Support

Page 12: Machine Learning and Analytics in the Cloud on IoT Data

12 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

Intel® IoTDevkits

https://software.intel.com/en-us/iot/hardware/edison/dev-kit

https://software.intel.com/en-us/iot/home

Intel® NUC kitThin Canyon

Arduino101

GroveSensor

Kit

Page 13: Machine Learning and Analytics in the Cloud on IoT Data

13 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

mraa.io & UPMAPI providing low-level and high-level wrappers to low-level communication protocols development platforms supporting many OSes

UPM is a software suite supporting over 250 sensors and actuators http://upm.mraa.io

Page 14: Machine Learning and Analytics in the Cloud on IoT Data

14 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

Intel® IoT Developer Kit IDEs

Intel® XDK IoT Edition(Node.js applications)

Intel® System Studio IoT Edition(C, C++, Java applications)

Page 15: Machine Learning and Analytics in the Cloud on IoT Data

15 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

Robust Developer Eco System

Intel® IoT Developer Kit, with easy to use hardware, software, tools, and cloud services.

IoT Roadshows, Hackathons, and Workshops providing hands-on, F2F training and practical experience.

Online communities, for learning, building, and sharing.

● Intel® Developer Zone for IoT

● Partner and 3rd party communities

ISV Enabling to build an ecosystem around the IoT Platform.

Intel® IoT Developer Program is a comprehensive program for makers, hobbyists, as well as professional developers, offering tools, kits, knowledge, and a community of experts to help you quickly and easily turn your innovative ideas into commercial IoT solutions.

Dream it, Build it with Intel® IoT

Page 16: Machine Learning and Analytics in the Cloud on IoT Data

16 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

Cloud Services on Intel® IoT Platforms

• Cloud Connector Recipes & Package (RPM)

• Cloud enabled kits for good OOB experience

• https://www.seeedstudio.com/Grove-IoT-Commercial-Gateway-Kit-p-2724.html

• https://www.seeedstudio.com/Grove-IoT-Developer-Kit-Microsoft-Azure-Edition-p-2631.html?ref=newInBazaar

• https://www.amazon.com/dp/B0168KU5FK

• Tools integration

• Sample Projects

Cloud Services for Intel® IoT Platforms include the following:

Page 17: Machine Learning and Analytics in the Cloud on IoT Data

17 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

• A typical IoT System

• Characteristics of IoT Data & Machine Learning techniques

• Intel IoTDevkit Overview

• A Complete end-to-end IoT Solution

Agenda

Page 18: Machine Learning and Analytics in the Cloud on IoT Data

18 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

A Complete end-to-end IoT Solution

Develop a setup where temperature sensor data is plotted in the cloud and when the temperature goes above a certain value

send out an email

Page 19: Machine Learning and Analytics in the Cloud on IoT Data

19 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

The Architecture

Temp Sensor EdisonATT Cloud

Sensor physically connected to Edison Analog port

• Time series sensor datam2x

• Node-red based flowsflow

Page 20: Machine Learning and Analytics in the Cloud on IoT Data

20 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

On Intel ® Edison

Page 21: Machine Learning and Analytics in the Cloud on IoT Data

21 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

Setup the cloud service

• Open an account at developer.att.com

• This will create accounts for you at

• m2x.att.com (a time series database service)

• flow.att.com (a cloud service programmable in node-red visual language)

• Setup m2x

• Edison device in m2x.att.com

• Add a stream to it called Temperature

• Make a note of deviceid and apikey

Page 22: Machine Learning and Analytics in the Cloud on IoT Data

22 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

Node Flow on Flow DesignerFrom previous slide

Page 23: Machine Learning and Analytics in the Cloud on IoT Data

23 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

Spin up the service in flow Pressing this will spin up a machine in cloud with appropriate services

After a few mins (some times it may take longer) a machine instance is created for you at bottom you will see the details

This goes into the node-red programon Edison in node send-to-cloud

Page 24: Machine Learning and Analytics in the Cloud on IoT Data

24 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

View in m2x after data gets streamed into cloud

Page 25: Machine Learning and Analytics in the Cloud on IoT Data

25 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

Add a trigger in m2x to detect hightemp condition

A simple rule based system usinga threshold value

Page 26: Machine Learning and Analytics in the Cloud on IoT Data

26 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

Flow to handle the Trigger from m2x

Page 27: Machine Learning and Analytics in the Cloud on IoT Data

27 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

Summary

New opportunities for Machine Learning at Cloud and Edge of IoT Systems

Collect and send to cloud

Massage the data in Cloud

Send to storage

Apply ML Model to detect

anomaly and notify cloud

Cloud takes needed action

Page 28: Machine Learning and Analytics in the Cloud on IoT Data

28 Copyright© 2013, Intel Corporation. All rights reserved. *Other brands and names are the property of their respective owners.

INFORMATION IN THIS DOCUMENT IS PROVIDED “AS IS”. NO LICENSE, EXPRESS OR IMPLIED, BY ESTOPPEL OR OTHERWISE, TO ANY INTELLECTUAL PROPERTY RIGHTS IS GRANTED BY THIS DOCUMENT. INTEL ASSUMES NO LIABILITY WHATSOEVER AND INTEL DISCLAIMS ANY EXPRESS OR IMPLIED WARRANTY, RELATING TO THIS INFORMATION INCLUDING LIABILITY OR WARRANTIES RELATING TO FITNESS FOR A PARTICULAR PURPOSE, MERCHANTABILITY, OR INFRINGEMENT OF ANY PATENT, COPYRIGHT OR OTHER INTELLECTUAL PROPERTY RIGHT.

Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark and MobileMark, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products.

Copyright © 2013, Intel Corporation. All rights reserved. Intel and the Intel logo are trademarks of Intel Corporation in the U.S. and other countries.

Optimization Notice

Intel’s compilers may or may not optimize to the same degree for non-Intel microprocessors for optimizations that are not unique to Intel microprocessors. These optimizations include SSE2, SSE3, and SSSE3 instruction sets and other optimizations. Intel does not guarantee the availability, functionality, or effectiveness of any optimization on microprocessors not manufactured by Intel. Microprocessor-dependent optimizations in this product are intended for use with Intel microprocessors. Certain optimizations not specific to Intel microarchitecture are reserved for Intel microprocessors. Please refer to the applicable product User and Reference Guides for more information regarding the specific instruction sets covered by this notice.

Notice revision #20110804

Legal Disclaimer & Optimization Notice