- Dr Harsh Verma
harshverma2000
Insights on Career Opportunities in
AI and Machine Learningfor IOT
Intel Innovator & IOT Technology Leader Vice-President (Global Innovation) & Head, IOT Group, R Systems Chairman, ACM Sacramento Chapter
Dedicated toStudents exploring different careers related to Computer Science Pathway
50 Billion Connected Things by 2020
References & Thanks to
The Coming Era of AI and ML
All Logos & Trademarks belong to the respective organizations
Artificial Intelligence is the capability of a
machine to imitate intelligent human behavior.
John McCarthy proposed the term ARTIFICIAL INTELLIGENCE in 1956
● Artificial Intelligence (AI) / Narrow or Weak AI: “Expert systems that match or exceed human intelligence in a narrowly defined area, but not in broader areas”
e.g. GPS Navigation System, Siri, Alexa.I
● Machine Learning: A field of computer science that gives computers the ability to learn without being explicitly programmed.
Definition of AI
Intelligence exhibited by Machines
All Logos & Trademarks belong to the respective organizations
AI already in your home…
• Amazon uses AI to power Alexa• Google Search Assistant is heavily dependent on AI• Apple uses AI to power Siri• Microsoft uses AI to power Cortana• Facebook uses AI for targeted advertising and photo tagging
All Logos & Trademarks belong to the respective organizations
Internet of Things – The Vision
A vision put forward by Mark Weiser to extend the Internet to include Objects or Things in the Real World. Internet of Things makes computing ubiquitous and Smart Physical Objects can be controlled remotely!
Internet of Things – Everywhere
• Connecting People and Things - To create a
fully connected lifestyle
• Smart-phones, Voice-capable tablets,Connected Cars and even Thermostats are Internet-connected
What kinds of experience and preparation?
About Me - Passion for Intelligent Transportation Systems
BITS Pilani, India
All Logos & Trademarks belong to the respective organizations
My current work
Today
www.rsystems.com
R Systems – A Global Software Services Company
13
Integrated Product Lifecycle Management (iPLM) organization
CMMI Level 5, PCMM Level 5, ISO 9001:2008, ISO 27001:2013 certified
Global presence - INDIA, APAC, USA, CANADA and EUROPE
2350+ employees worldwide
All Logos & Trademarks belong to the respective organizations
www.rsystems.com
• Telecom, Digital Media & Mobility Solutions• BFSI• Manufacturing & Logistics • HealthCare
• Software Development Services • Quality Assurance & Control Services• BPO Solutions – Technical Help Desk
Services Domains
Intel® Software Innovator program supports innovative, independent development
to create and demonstrate forward-looking projects.
Intel Software Innovators demonstrate a spirit of ingenuity, experimentation, and
forward thinking that inspire communities through technical expertise & innovation
with cutting-edge technology.
All Logos & Trademarks belong to the respective organizations
www.intel.com
ACM is a vibrant international community of 100,000 members in over 100 countries.
ACM brings together computing educators, researchers, and professionals to inspire
dialogue, share resources, and address the field's challenges.
https://www.facebook.com/ACMSacramento/
Launched 2015
All Logos & Trademarks belong to the respective organizations
1969 1990 2005 2015 2020
Research Networks
World Wide Web
TCP/IP http:// VoIP
Web 2.0
Web3.0/ Internet of Things
Mobile Web Social Media
M2M Smart Cities
How did we get to the Internet of Things?
It’s going to be a World of Internet of Things
Predicted ~50 billion devices and objects to be connected by 2020
$19 Trillion Opportunity, according to Cisco Systems
Smart Cities -Multiple Building Blocks
Intelligent System of Systems
Mobile is re-defining everything
Intelligent Devices
Back-End Analytics
Predicted ~50 billion devices and objects to be connected by 2020
Industrial IoT
Industrial IoT : 4th Industrial Revolution
All Logos & Trademarks belong to the respective organizations
• IoT all about Sensors and Data
• Less than 1% of Data is currently used
• Mostly for Alarms or Real-time control
• Potential for Optimization and Prediction
IoT is Transforming our Lives
Source: McKinsey Global Institute
What kind of training or background is necessary?
www.rsystems.com
A variety of Sectors for IoT – Multiple Clusters!
Home
Chore Automation & Security $200B − 350B
Offices
Security and Energy
$70B − 150B
Factories
Operations and
Equipment Optimization
$1.2T − 3.7T
Retail Environments
Automated Checkout
$410B − 1.2TWorksites
Operations optimization/
Health and Safety
$160B − 930B
Human
Health and Fitness
$170B − 1.6T
Outside
Logistics and Navigation
$560B − 850B
Cities
Public Health
and Transportation
$930B − 1.7T
Vehicles
Autonomous vehicles and
condition-based maintenance
$210B − 740B
Cross-sector viewof a total potential impact of$3.9 trillion – 11.1 trillion
per year in 2025
Source: McKinsey Global Institute
www.rsystems.com
Areas where IoT Creates Value
22
Setting Description Example
HumanDevices attached to or inside the
human body
Devices (wearable's and ingestible) to monitor and
maintain human health and wellness; diseasemanagement, increased fitness, higher productivity
Home Buildings where people live Home controllers and security systems
RetailEnvironments
Spaces where consumers engage in commerce
Stores, banks, restaurants, arenas — anywhereconsumers consider and buy; self-checkout, in-storeoffers, inventory optimization
OfficesSpaces where knowledge workers work
Energy management and security in office buildings;improved productivity, including for mobileemployees
FactoriesStandardized production environments
Places with repetitive work routines, includinghospitals and farms; operating efficiencies,optimizing equipment use and inventory
SOURCE: McKinsey Global Institute analysis
www.rsystems.com
Areas where IoT Creates Value (Contd…)
23
Setting Description Example
Worksites Custom production environmentsMining, oil and gas, construction; operatingefficiencies, predictive maintenance, health andsafety
Vehicles Systems inside moving vehiclesVehicles including cars, trucks, ships, aircraft, andtrains; condition based maintenance, usage-baseddesign, pre-sales analytics
Cities Urban environmentsPublic spaces and infrastructure in urban settings;adaptive traffic control, smart meters, environmentalmonitoring, resource management
OutsideBetween urban environments (andoutside other settings)
Outside uses include railroad tracks, autonomousvehicles (outside urban locations), and flightnavigation; real-time routing, connected navigation,shipment tracking
SOURCE: McKinsey Global Institute analysis
• All these areas have huge potential for AI & Machine Learning – Find an Area that interests you!!
• AI is any technique that enables computers to mimic human intelligence
• Machine Learning enables machines to improve at Tasks with Experience
• Deep learning consists of algorithms that permit software to use artificial neural networks and train itself to perform tasks efficiently
AI, Machine Learning and Deep Learning
World Distribution of Data Scientists on DSC
Source: IIM Americas, Pinnacle 2017, AI & ML - Sanjiv R Das
● Rule-Based AI
Rule-based systems are constructed using automatic rule inference, from a starting set of data and rules.
● Data-driven AI
Data-driven systems can measure, understand and take decisions based on previous data,
which goes on increasing.
Rule-based AI can never be more intelligent than its creators, but data-driven AI can!
Two Types of AI
Training Steps for Machine Learning
https://www.youtube.com/watch?v=IpGxLWOIZy4
Performing Linear Regression
Least Squares Method to minimize Error
Convolutional Neural Networks
• A CNN consists of an input and an output layer, as well as multiple hidden layers.
• Convolutional layers apply a convolution operation to the input, passing the result to the next layer.
• Distinct layers transform the input into an output through a differentiable function.
Each Node represents a Pattern
CNNs approximate the response of individual Neurons in the Human Brain
• If a network has more than 3 layers, it is Deep• Patterns get more complex as number of layers go up
Source: Google
www.rsystems.com
IBM Watson beats Humans in Jeopardy!
Watson is a question answering computer system capable of answering questions
https://en.wikipedia.org/wiki/Watson_(computer)
Practice School
Capstone Project
Students pursue independent research on a question or problem
What has given me the most help at the start of my career?
Start with your House –IOT Project for Home Automation
More Info - https://github.com/intel-iot-devkit/path-to-product/tree/master/smart-home
Intel NUC Smart Home:
• Intel NUC, Arduino 101,
• Grove Kit
• Sensor -> Gateway -> Mobile App
• Cloud Analytics - IBM Bluemix
How do you make it Intelligent?
1. Define the opportunity2. Design a Proof-of-Concept3. Build and refine the prototype4. Learn and Have fun!
All Logos & Trademarks belong to the respective organizations
Programming Languages & Databases
AI and ML Toolkits
Toolkits driven by Academia & Industry for scripting & training ML algorithms
High Performance Math Primitives granting low level of control
Broad Data Analytics Acceleration object oriented library supporting distributed ML at the algorithm level
Open Source single platform from Data Science to Application Development
How do you get Processing at the Edge ?Various architectural patterns, including Fog and Edge computing, have been proposed to localize some of the processing near the IOT sensors
All Logos & Trademarks belong to the respective organizations
Intel Movidius Neural Compute Stick
https://developer.movidius.com/start
Intel Movidius Neural Compute Stick
• Enables rapid prototyping and deployment of Neural Networks
• Developing Inference applications at the edge.
• Very low-power VPU Architecture
• Processing at the Edge and takes away dependencies of the Cloud and Communication Network Bandwidth
All Logos & Trademarks belong to the respective organizations
Intel Movidius Neural Compute Stick
https://developer.movidius.com/start
Intel Movidius Neural Compute Stick
• Requires Ubuntu 16.04 and NC SDK
• NCS plugs into the USB Port
• NCS Forum, GitHub, NCS Application Zoo
Creating and Training the Model
• Training Data Sets available
All Logos & Trademarks belong to the respective organizations
Movidius NCS Workflow
https://developer.movidius.com/start
All Logos & Trademarks belong to the respective organizations
Enabling Wireless Services for IoT
Wireless Connectivity
WiFi
3G
4G
BT
ZigBee
6LoWPAN
NFC
DSRC
Visions & Contradicting Opinions on AI…
Elon Musk –
AI is a fundamental risk to the existence of human civilization
Founded OpenAI – for Safe AIFounded Neuralink to merge human brains with machines.
Mark Zuckerberg –
Very Optimistic on the Applications and Benefits of AI
Source: https://www.cnbc.com/
AI in Autonomous Vehicles
Intel
Autonomous Driving will End Human Driving Errors
and Lead to Safer Roads for Everyone
One of the big promises of AI is our driverless future.
Nearly 1.3 million people die in road crashes worldwide every year – an average
3,287 deaths a day. Nearly 90% of those collisions are caused by human error.- https://newsroom.intel.com/editorials/waymo-intel-announce-collaboration-driverless-car-technology/
AI in Self-Driving Cars qualify as Legal Driver
National Highway
Transportation and Safety
Administration AI system in
Google Self-driving Car that
controls its self-driving car
can be considered a driver
under federal law!
Ford:● Invests in new AI software company, Argo AI
Google’s Self-Driving Cars:● Waymo collaboration with Intel announced
Tesla: • AI will let cars predict your Destination
BMW: • Announced BMW will have an Intel Xeon Server-Grade Processor
All Logos & Trademarks belong to the respective organizations
AI in Self-Driving and Autonomous Cars
IoT and Autonomous World will not work without AI
The Dark Side of AI
Can Super-AI eliminate Humans?
● How do we manage uncontrollable AI?
● Encrypted Deep Learning: Build a super-intelligent AI to detect and cure
cancer (for example), but encrypt it so that it cannot replicate itself, nor can it
be used without human control.
● Encryption prevents it from being stolen by a malevolent player.
● Homomorphic encryption: Only permit a limited range of encrypted
predictions.
https://iamtrask.github.io/2017/03/17/safe-ai/
The biggestproblem is going to be evident in the labor
markets…
Source: IIM Americas, Pinnacle 2017, AI & ML - Sanjiv R Das
7 Challenges for IoT
Security
Security
Security
Standards
Privacy
Reliability
Scalability
5 Skills Hiring Managers Look for
1. Technical Skills
2. Communication Skills
3. Interpersonal Skills
4. Problem Solving and Critical Thinking
5. Enthusiasm, Motivation and Commitment
https://www.engineering.com/JobArticles/ArticleID/13894/5-Skills-Hiring-Managers-Look-for-in-Engineering-Grads.aspx
Summary
● AI and ML are a happening transformation
● AI is everywhere and Super AI is nowpossible.
● IOT needs AI for it to work and there is immense potential
● AI has a dark side
● There is also a great potential for Cyber-Security in today’s world
Read more at Stuart Russell’s terrific collection of links:
https://people.eecs.berkeley.edu/~russell/research/future/
Top Related