Transferable Skills - Your Roadmap - Part 1 and 2 - Dirk Spencer Senior Recru...
Deep Learning Session 1 : bright future for you summary
1. Artificial Intelligence
for everyone
Be a leader in
applying Deep Learning
https://sites.google.com/view/AIforEveryone
AI for Good Workshop
March 2019
Session 1
2. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
Approach
•Day 1: Big ideas, more conceptual thinking, think
a idea-a-thon (thinkers and colloborators)
•Day 2: Next level details, more code, make it
happen, hack-a-thon (thinkers and coders)
3. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
Workshop goals
• State of art Deep Learning made friendly
• Special: Made Simple, We won’t miss on advanced approaches
(includes Future of AI)
• Thinking Quiz , Conceptual understanding / Hands-on / Group Activity
• AI for Social Good
• Special: Brings out the hidden potential in you !
• Idea-a-thon :
•Find a problem that you want to solve, and solve it with AI
4. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
Day 1
• 9.30 am: Deep Learning Big ideas
• 10:00 am: How to solve a problem with AI // Research & Architect stream
• 10:30 am: How to architect // Research & Architect stream
• 11:15 am: How to design a computer that can learn like a child // Math & job stream
• 11:45 am: ‘Hello World’ // Coder stream // make it happen stream
• 12noon: Special talk on “AI for Good” (Live from Mumbai Wadhwani AI)
• 2 pm: Current and Future trends in AI // Innovator stream
• 3pm: Working session as Team / Goal: Find an idea you want to solve //Leader stream
5. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
Day 2
• 10:00 am : How to architect – Thinking activity - Recap // Research & Architect stream
• 10:50 break
• 11:00 am: Special talk on “Applying Tech for Good” (Live from Bangalore)
• 11.30 am: ‘Hello World’ // Coder stream // make it happen stream
• 2:00 pm: Functional API // Coder stream // make it happen stream
• 3:00 pm: Current and Future trends in AI // Research & Architect stream
• 3.30 pm: Hack-a-thon Team : Code / AI for Good idea //Entrepreneur stream
• 4:00pm : Session ends
6. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
Pre-requisites = mathematics / None
•Don’t worry!
• This workshop is for you !!
• Do you need to have a PhD in Mathematics to learn AI?
• Pre-requisites: None
7. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
Our inspiration
Friendly approaches :
1) KERAS.io
François Chollet’s
Book on “Deep Learning with Python”
2) Deeplearning.ai (Coursera.org)
Andrew Ng
3) Fast.ai
Jeremy
8. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
More inspiration
Excellent Resources
• Stanford cs231 n
http://cs231n.stanford.edu
• MIT Deep Learning
http://introtodeeplearning.com/
https://deeplearning.mit.edu
• IIT Madras
my classes notes with Prof. Anurag (Deep Learning)
9. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
Learning objectives
Time slots Sub-topic What will you gain What will you learn / Keywords
Day 1 morning Introduction
to Deep Learning
· Get familiar with the art of Deep Learning
· Get started to 1st Hello World Artificial
Intelligence project
· Get started to code & experiment
your Deep Learning projects with friendly library
· Introduction
to Deep Learning concepts
· How to easily get started
on Deep Learning projects using Keras on
Google Tensorflow
· How to apply Deep Learning for
Computer Vision
Day 1 evening Invent the next
breakthrough by learning
the secrets from pioneering
AI researchers
· Think like a AI guru as we learn from leading AI
research papers
· Learn the design an
new deep learning architecture to how to solve a
healthcare challenge
· Get exposed to the art of solving real world
challenges with advanced Deep Learning
· Art of architecting a neural network
· Grasp how to develop cutting edge AI
by learning from state of art AI research in
the last 3 years (Image classification, Auto-
encoder, GAN , AutoML, Federated
Learning)
· Hands-on practice to Keras Functional
API on python
Day 2 morning Gain the art of solving a
problem with AI
· Hands-on practice
on Deep Learning techniques
· How to solve problems in healthcare
innovation or agricultural productivity
Day 2 evening AI for Good Hack-a-thon
Solve a real world problem
and develop a working
prototype solution
· Team formation and Design thinking
· Practice to convert an idea to reality
· Get started to build a AI product
· Hack-a-thon and win prizes as a team
· Present your ideas and results in 60
seconds
10. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
Goal
1. What is the future of AI?
2. How can you apply AI?
3. How can you invent AI?
You will be a leader in Deep Learning
11. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
Why you will be a leader?
“Though AI research is moving
forward amazingly fast,
most of the research is
NOT yet applied”
Your opportunity
to be a leader
12. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
13. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
The future of AI
1. Deep Learning
“Biggest leap since invention
of computers”
Output of AI: Scalar values
(numbers)
2. Generative
Deep Learning
GAN “the most interesting idea
in the last 10 years in ML.”
Output of AI: Vectors , Images
3. AI that creates
another AI
Architecture Search/
Progressive
Output of AI: Neural netwoks
14. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
The future of AI
1. Deep Learning
“Biggest leap since invention
of computers”
Output of AI: Scalar values
(numbers)
2. Generative
Deep Learning
GAN “the most interesting idea
in the last 10 years in ML.”
Output of AI: Vectors , Images
3. AI that creates
another AI
Neural network auto-evolution
Output of AI: Neural netwoks
Output = 2 Output = Output =
15. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
1. Deep Learning
visual recognition – image classification
•
Credits: Google Tensorflow Dev Summit 2017
16. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
1. Deep Learning
Predicting the future
17. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
2. Generative Deep Learning
Convert Text to image
• Generative Deep Learning
• VAE
• GAN
3. music synthesis
18. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
2. Generative Deep Learning
Create a artificial face
19. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
3. Artificial General Intelligence
A neural network creates another on
Parent / controller
20. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
3. Artificial General Intelligence
A neural network creates another one!
• Output: neural network
Credits: AI Learns Video Frame Interpolation | Two Minute Papers #197
21. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
Examples project: AI for Good
22. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
Special talks on AI for good projects
• Food
• A way to measure levels of pest in agriculture
farms
• Deep Learning Architecture aspects:
• Detect type of pest (image classification)
• Count number of pests (object detection/ Region-CNN,
SSD)
• Uplift children / educate blind
• Improve education for blind children
• Architecture aspects:
• Detect location of finger and edges of pictures
Dr. Mythili & team , MSRIT
23. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
Key Learning objective
for Day 1
How to solve a problem with Deep Learning ?
How to design an neural network architecture?
24. Acknowledgments & Credits are mentioned to inspirational resources presented in Slide 7 & 8. For more like this, https://sites.google.com/view/AIforEveryone
Thinking assignment
Monitor your heart 24/7 and preventive health alerts
heart sounds health alert