Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
Wednesday, July 31, 2019
1. Intro & Activity Update
2. Community Open Mic
3. Alex Barbosa Coqueiro - Public
Sector SA Mana...
Why we do what we do?
2
Serverless is New Agile
Serverless Dev (gluing
other people’s APIs
and managed services)
We're obs...
Why AI/ML topics at Serverless meetup?
3
AWS Machine Learning Stack
#ServerlessTO
Sponsors
4
Manning Publications 2019 giveaways:
1. www.manning.com/books/serverless-applications-with-nodejs
2. www.manning.com/livev...
Venue Sponsor
6
As Certified B Corporation, Myplanet is purpose-driven and
creates benefit for all stakeholders, not just ...
Catering Sponsor
7
Also an active Serverless Community member – check out
their “Data Pipelines using Serverless Architect...
Devinity
8
Devinity… continued
9
Community Open Mic
10
10 seconds of freedom
to pitch yourself, or
your company
Feature Talk
Racing with AI
11
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Racing with Artificial Intelligence
Alex Coqueir...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Rubik’s cube challenge
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
43,252,003,274,489,856,000
43 QUINTILLION
UNIQUE...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Don’t code the patterns, let the
system learn th...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
F2 U' R' L F2 R L' U'
ModelData
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
F2 U' R' L F2 R L' U'
Confidence
1%
accuracy
R U...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Confidence
20%
accuracy
40%
accuracy
60%
accurac...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Confidence
95%
accuracy
?
F2 R F R′ B′ D F D′ B ...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
SOLVED IN 0.9 SECONDS
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Let’s apply it into the business
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
https://youtu.be/xC-tikvEvzo
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
10,000+ customers | 2x customer references | 85%...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
LEA (Robot Care Systems)
https://youtu.be/r-US8r...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
tuSimple (Autonomous Vehicle)
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Show me how to do it
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
© 2018, Amazon Web Services, Inc. or its affilia...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Our problem re-formulation
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Different problems require different learning st...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Autonomous Driving Development
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Robocar (Donkey Car Project)
Donkey Car Project
...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
SSD MultiBox — Real-Time Object Detection +
Beha...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Amazon SageMaker:
Build, Train, and Deploy ML Mo...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Different problems require different learning st...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Reinforcement learning in the real world
Reward ...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
• Build machine learning models in Amazon
SageMa...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Robotic autonomous
race car
DeepRacer: An exciti...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Track components
TRACK CENTER
TRACK WALL
TRACK S...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Action space
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
The reward function in a race grid
S G = 2
GOALA...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Incentivizing centerline behavior
0.1 0.1 0.1 0....
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Programming your own reward function
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
© 2018, Amazon Web Services, Inc. or its affilia...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
AWS DeepRacer Neural Network Architecture
An ove...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Amazon
Sagemaker RL
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Reinforcement Learning Algorithms Compared
Value...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Hyper parameters control the training algorithm
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Coordinate system and track waypoints
OUTER BOUN...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
AWS DeepRacer Car Specifications
CAR 18th scale ...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Robotic Arms
International Space Station
Drones
...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Robotics trends
in 2018
Robotics is undergoing f...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Robotic development cycle
2) Develop
robotics
ap...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Most widely used software framework for teaching...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Introducing AWS RoboMaker
A service that makes i...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
AWS RoboMaker service suite
Development
Environm...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
AWS RoboMaker
Sample Robot Applications
Hello
Wo...
AWS Cloud
AWS
DeepRacer
NAT gateway
VPC
AWS DeepRacer
Models
Simulation
video
Metrics
AWS DeepRacer Simulation Architecture
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
ROS Msg Node
Stored File
ROS Nodes
Model
Optimiz...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
DATA
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
We are not spectators,
but actors of the future
...
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
ml.aws
@alexbcbr
Thank you!
Future Talks
2019
1
Upcoming Serverless Toronto Meetups
2
1. August 12, 2019: 1. A look at Google Cloud Functions
– Matt Welke // 2. Firebase ...
Upcoming SlideShare
Loading in …5
×

Racing with Artificial Intelligence

144 views

Published on

#AI + #ML + #Robotics combination is a game-changer, so #ServerlessTO members were lucky to have Alex Barbosa Coqueiro - Public Sector Solutions Architect Manager at AWS Canada, introduce us to AWS Robomaker & AWS DeepRacer!

Alex also talked about managed #ReinforcementLearning (RL) with Amazon SageMaker, and compared DeepRacer to Donkey Car – open source project for small-scale self-driving cars.

Video at: https://youtu.be/t8bo9gOveoo

Published in: Science
  • Be the first to comment

  • Be the first to like this

Racing with Artificial Intelligence

  1. 1. Wednesday, July 31, 2019 1. Intro & Activity Update 2. Community Open Mic 3. Alex Barbosa Coqueiro - Public Sector SA Manager @ AWS Canada: “Racing with AI” 4. Networking 1 ServerlessToronto.org Meetup Agenda
  2. 2. Why we do what we do? 2 Serverless is New Agile Serverless Dev (gluing other people’s APIs and managed services) We're obsessed helping Startups (and creating meaningful MVPs & products) We build bridges between Serverless Community (“Dev leg”), and Front-end & Voice- First folks (“UX leg”), and empower UX developers Achieve agility NOT by “sprinting” faster (like in Scrum), but by working smarter (by using bigger building blocks and less Ops)
  3. 3. Why AI/ML topics at Serverless meetup? 3 AWS Machine Learning Stack
  4. 4. #ServerlessTO Sponsors 4
  5. 5. Manning Publications 2019 giveaways: 1. www.manning.com/books/serverless-applications-with-nodejs 2. www.manning.com/livevideo/production-ready-serverless 3. www.manning.com/livevideo/production-ready-serverless 4. www.manning.com/livevideo/serverless-applications-with-AWS 5. www.manning.com/livevideo/serverless-applications-with-AWS 6. www.manning.com/books/serverless-architectures-on-aws 7. www.manning.com/books/http2-in-action 8. www.manning.com/books/event-streams-in-action 9. www.manning.com/books/the-design-of-everyday-apis 10. www.manning.com/livevideo/graphql-in-motion 11. www.manning.com/books/voice-applications-for-alexa-and-google-assistant 12. www.manning.com/livevideo/machine-learning-for-mere-mortals 13. www.manning.com/books/classic-computer-science-problems-in-python 14. www.manning.com/books/getting-mean-with-mongo-express-angular-and-node 5Check out MEAP program from our Learning Sponsor!
  6. 6. Venue Sponsor 6 As Certified B Corporation, Myplanet is purpose-driven and creates benefit for all stakeholders, not just shareholders!
  7. 7. Catering Sponsor 7 Also an active Serverless Community member – check out their “Data Pipelines using Serverless Architectures” talk!
  8. 8. Devinity 8
  9. 9. Devinity… continued 9
  10. 10. Community Open Mic 10 10 seconds of freedom to pitch yourself, or your company
  11. 11. Feature Talk Racing with AI 11
  12. 12. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Racing with Artificial Intelligence Alex Coqueiro Head of Public Sector Solutions Architecture for Canada, Latin America and Caribbean AWS @alexbcbr
  13. 13. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Rubik’s cube challenge
  14. 14. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. 43,252,003,274,489,856,000 43 QUINTILLION UNIQUE COMBINATIONS
  15. 15. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Don’t code the patterns, let the system learn through data
  16. 16. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. F2 U' R' L F2 R L' U' ModelData
  17. 17. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. F2 U' R' L F2 R L' U' Confidence 1% accuracy R U r U R U2 r U2% accuracy Training Models Model
  18. 18. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Confidence 20% accuracy 40% accuracy 60% accuracy 80% accuracy 95% accuracy 2% accuracy Training Models Model
  19. 19. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Confidence 95% accuracy ? F2 R F R′ B′ D F D′ B D F Inference Model
  20. 20. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
  21. 21. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. SOLVED IN 0.9 SECONDS
  22. 22. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Let’s apply it into the business
  23. 23. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. https://youtu.be/xC-tikvEvzo
  24. 24. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. 10,000+ customers | 2x customer references | 85% of TensorFlow projects in the cloud happen on AWS
  25. 25. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. LEA (Robot Care Systems) https://youtu.be/r-US8rs8EY0
  26. 26. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. tuSimple (Autonomous Vehicle)
  27. 27. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Show me how to do it
  28. 28. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved.Use Case – Autonomous Driving
  29. 29. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Our problem re-formulation
  30. 30. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Different problems require different learning strategies labeled training data Complexityofdecisions Supervised learning Non-labeled training data
  31. 31. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Autonomous Driving Development
  32. 32. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Robocar (Donkey Car Project) Donkey Car Project https://github.com/sunilmallya/donkey/tree/master/sagetrain http://awsrobocar.s3-website-us-east-1.amazonaws.com/ https://github.com/tescal2/donkey
  33. 33. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. SSD MultiBox — Real-Time Object Detection + Behavioral Cloning
  34. 34. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon SageMaker: Build, Train, and Deploy ML Models at Scale 1 2 3
  35. 35. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Different problems require different learning strategies labeled training data Complexityofdecisions Supervised learning Unsupervised learning Reinforcement Learning Non-labeled training data
  36. 36. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Reinforcement learning in the real world Reward positive behavior Don’t reward negative behavior The result!
  37. 37. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. • Build machine learning models in Amazon SageMaker • Train, test, and iterate on the track using the AWS DeepRacer 3D racing simulator • Compete in the world’s first global autonomous racing league, to race for prizes and a chance to advance to win the coveted AWS DeepRacer Cup AWS DeepRacer A fully autonomous 1/18th-scale race car designed to help you learn about reinforcement learning through autonomous driving
  38. 38. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Robotic autonomous race car DeepRacer: An exciting way for developers to get hands-on experience with Reinforcement Learning Racing LeagueVirtual simulator, to train and experiment
  39. 39. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Track components TRACK CENTER TRACK WALL TRACK SURFACE aka ON-TRACK FIELD aka OFF-TRACK TRACK BOUNDARIES
  40. 40. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Action space
  41. 41. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. The reward function in a race grid S G = 2 GOALAGENT
  42. 42. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Incentivizing centerline behavior 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 S 2 2 2 2 2 2 G = 2 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 REWARD FUNCTION 8.6 9.5 8.5 7.5 6.3 5.0 3.5 1.9 S 10.4 9.4 8.2 6.9 5.4 3.8 G = 2 8.6 9.5 8.5 7.5 6.3 5.0 3.5 1.9 MAX VALUE OF EACH STATE AFTER LOTS OF EXPLORING Discount per step 0.9
  43. 43. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Programming your own reward function
  44. 44. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. © 2018, Amazon Web Services, Inc. or its affiliates. All rights reserved. Let’s go deeper Let’s go deeper…
  45. 45. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. AWS DeepRacer Neural Network Architecture An overview of the network architecture that AWS DeepRacer uses: Output
  46. 46. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Sagemaker RL
  47. 47. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Reinforcement Learning Algorithms Compared Value Approximation Policy Approximation Advantages More stable performance when it works, and tends to converge on global optimum Effective in continuous action spaces, can learn stochastic policies, and faster convergence Disadvantages Difficult to converge if too many (state, action) combinations, slower convergence in general, and can’t learn stochastic properties Typically converges to a local rather than global optimum, high variance in estimating the gradient adversely affects stability, and evaluating a policy is generally inefficient Examples Q-Learning, Deep Q Network, Deep Double Q Network Policy Gradient, Proximal Policy Optimization (PPO)
  48. 48. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Hyper parameters control the training algorithm
  49. 49. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Coordinate system and track waypoints OUTER BOUNDARY WAYPOINTS TRACK CENTER WAYPOINTS INNER BOUNDARY WAYPOINTS X Y TRACK WIDTH CAR DIRECTION
  50. 50. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. AWS DeepRacer Car Specifications CAR 18th scale 4WD with monster truck chassis CPU Intel Atom™ Processor MEMORY 4GB RAM STORAGE 32GB (expandable) WI-FI 802.11ac CAMERA 4 MP camera with MJPEG DRIVE BATTERY 7.4V/1100mAh lithium polymer COMPUTE BATTERY 13600mAh USB-C PD SENSORS Integrated accelerometer and gyroscope PORTS 4x USB-A, 1x USB-C, 1x Micro-USB, 1x HDMI SOFTWARE Ubuntu OS 16.04.3 LTS, Intel® OpenVINO™ toolkit, ROS Kinetic
  51. 51. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Robotic Arms International Space Station Drones Education Water Home Self-Driving Vehicles Autonomous Walker Rover Robot landscape
  52. 52. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Robotics trends in 2018 Robotics is undergoing fundamental change in collaboration, autonomous mobility, and increasing intelligence Source: IDTechEx • Logistics • Construction • Retail • Hospitality • Healthcare Robots are being put to work every day across many industries • Agriculture • Energy Management • Oil and Gas • Facilities Management • Household chores By 2023, it’s estimated that mobile autonomous robots will emerge as the standard for logistic and fulfillment processes By 2030, 70% of all mobile material handling equipment will be autonomous
  53. 53. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Robotic development cycle 2) Develop robotics application 1) Select robotics software framework 1) Deploy and manage application 3) Test and simulate application New application release and update
  54. 54. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Most widely used software framework for teaching and learning about robotics – over 16 million .deb (Linux Debian) packages downloaded in 2018, a 400% increase since 2014 Founded in Stanford labs over 10 year ago, now managed by the Open Source Robotics Foundation (OSRF) Global open-source community supports two products—Robot Operating System (ROS) and Gazebo ROS A set of software libraries and tools, from drivers to algorithms, that help developers build robot applications Gazebo Robust physics engine, high-quality graphics, and programmatic and graphical interfaces to help developers simulate robots Robot Operating System (ROS) primer
  55. 55. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. Introducing AWS RoboMaker A service that makes it easy for developers to develop, test, and deploy robotics applications, as well as build intelligent robotics functions using cloud services
  56. 56. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. AWS RoboMaker service suite Development Environment SimulationCloud Extensions for ROS Fleet Management
  57. 57. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. AWS RoboMaker Sample Robot Applications Hello World Navigation and Person Recognition Voice Commands Robot Monitoring Object- following using RL Self- driving using RL
  58. 58. AWS Cloud AWS DeepRacer NAT gateway VPC AWS DeepRacer Models Simulation video Metrics AWS DeepRacer Simulation Architecture
  59. 59. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. ROS Msg Node Stored File ROS Nodes Model Optimizer Video M-JPEG Web Server Video Inference Results Web Server Publisher Autonomous Drive Control Node Optimized Model Media engine Camera Model Inference engine Manual Drive Navigation Node Servo & Motor AWS DeepRacer Software Architecture
  60. 60. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. DATA
  61. 61. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. We are not spectators, but actors of the future Herb Simon, 2000
  62. 62. © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved. ml.aws @alexbcbr
  63. 63. Thank you!
  64. 64. Future Talks 2019 1
  65. 65. Upcoming Serverless Toronto Meetups 2 1. August 12, 2019: 1. A look at Google Cloud Functions – Matt Welke // 2. Firebase Crash Course – Kudz Murefu 2. September 12, 2019: Serverless Design Patterns and Best Practices – Mike Apted, Startup SA at AWS 3. October, 2019: Serverless CI/CD pipelines with AWS CodePipeline and CodeBuild, vs CircleCI, vs Travis, vs Seed – Frank Wang & Jay V 4. November or December: Serverless Heroes (authors of “Serverless Applications with Node.js”) Slobodan Stojanović & Aleksandar Simović will be here! 5. December or January, 2020: re:Invent recap – Jonathan Dion, Senior Technical Evangelist at AWS

×