Slide 1

Slide 1 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Demystifying the AI Black Box ©2018 Amazon Web Services, Inc. or its affiliates, All rights reserved Adrian Hornsby, Senior Technical Evangelist @adhorn

Slide 2

Slide 2 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Myth: AI is dark magic

Slide 3

Slide 3 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. What are we talking about? AI Machine Learning Deep Learning

Slide 4

Slide 4 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. One of the ”Founding Father" of Artificial Intelligence John McCarthy 1955

Slide 5

Slide 5 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Photo from the 1956 Dartmouth Conference with Marvin Minsky, Ray Solomonoff, Claude Shannon, John McCarthy, Trenchard More, Oliver Selfridge and Nathaniel Rawchester

Slide 6

Slide 6 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Frank Rosenblatt, 1957 Perceptron

Slide 7

Slide 7 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Paul Werbos, 1975 Backpropagation

Slide 8

Slide 8 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. LeCun, 1989 First application of backpropagation https://www.youtube.com/watch?v=FwFduRA_L6Q

Slide 9

Slide 9 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. The curse of dimensionality

Slide 10

Slide 10 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. The Advent of AI Data GPUs & Acceleration Cloud Computing Algorithms

Slide 11

Slide 11 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. What is a Deep Learning?

Slide 12

Slide 12 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Predicting the price of a house with humans Price City ZipCode Life Quality Parking Size # Room Accessibility Family Friendly

Slide 13

Slide 13 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Predicting the price of a house with neural network Price City ZipCode Life Quality Parking Size # Room Accessibility Family Friendly Input Output Discovered by the neural network

Slide 14

Slide 14 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. https://ml4a.github.io/ml4a/neural_networks/ Deep Learning Training Learning

Slide 15

Slide 15 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Learning = Minimizing the loss (error) function Backpropagation

Slide 16

Slide 16 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Deep Learning is a big deal It’s able to do better than other ML and Humans

Slide 17

Slide 17 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. https://github.com/precedenceguo/mx-rcnn https://github.com/zhreshold/mxnet-yolo Object Detection

Slide 18

Slide 18 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Text Detection and Recognition https://github.com/Bartzi/stn-ocr

Slide 19

Slide 19 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. https://github.com/tornadomeet/mxnet-face Face Detection

Slide 20

Slide 20 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. FDA-approved medical imaging https://www.periscope.tv/AWSstartups/1vAGRgevBXRJl https://www.youtube.com/watch?v=WE81dncwnIc Object Segmentation

Slide 21

Slide 21 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Real-Time Pose Estimation https://github.com/dragonfly90/mxnet_Realtime_Multi-Person_Pose_Estimation

Slide 22

Slide 22 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Generative Adversarial Networks (GAN) The future at work (already) today Generating new ”celebrity” faces https://github.com/tkarras/progressive_growing_of_gans

Slide 23

Slide 23 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Data Visualization & Analysis Business Problem – ML problem framing Data Collection Data Integration Data Preparation & Cleaning Feature Engineering Model Training & Parameter Tuning Model Evaluation Are Business Goals met? Model Deployment Monitoring & Debugging Yes No Data Augmentation Feature Augmentation The ML Process Re-training Predictions

Slide 24

Slide 24 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Amazon ML

Slide 25

Slide 25 text

No content

Slide 26

Slide 26 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.

Slide 27

Slide 27 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.

Slide 28

Slide 28 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.

Slide 29

Slide 29 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.

Slide 30

Slide 30 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. FRAMEWORKS AND INTERFACES PLATFORM SERVICES APPLICATION SERVICES Amazon Rekognition Amazon Polly Amazon Lex ML for everyone. Amazon Rekognition Video Amazon Transcribe Amazon Comprehend Amazon SageMaker AWS DeepLens Amazon EMR Deep Learning AMI Amazon Translate

Slide 31

Slide 31 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Where do you start? The Low Hanging Fruits

Slide 32

Slide 32 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. FRAMEWORKS AND INTERFACES PLATFORM SERVICES APPLICATION SERVICES Amazon Rekognition Amazon Polly Amazon Lex ML for everyone. Amazon Rekognition Video Amazon Transcribe Amazon Comprehend Amazon SageMaker AWS DeepLens Amazon EMR Deep Learning AMI Amazon Translate

Slide 33

Slide 33 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Amazon Rekognition Deep learning-based image & video analysis

Slide 34

Slide 34 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. © 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. http://timescapes.org/trailers/

Slide 35

Slide 35 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Marinus Analytics uses facial recognition to stop human trafficking “Now with Traffic Jam’s FaceSearch, powered by Amazon Rekognition, investigators are able to take effective action by searching through millions of records in seconds to find victims.” http://www.marinusanalytics.com/articles/2017/10/17/amazon-rekognition-helps-marinus-analytics-fight-human-trafficking

Slide 36

Slide 36 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.

Slide 37

Slide 37 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Amazon Polly “What's the difference between a hippo and a zippo? One is really heavy, and the other is a little lighter.” Amazon Polly: Text In, Life-like Speech Out The Text-To-Speech technology behind Amazon Polly takes advantage of bidirectional long short-term memory (LSTM)* 52 voices across 25 languages

Slide 38

Slide 38 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. The price of this book is €45 A Focus On Voice Quality & Pronunciation Support for Speech Synthesis Markup Language (SSML) Version 1.0 https://www.w3.org/TR/speech-synthesis

Slide 39

Slide 39 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. “With Amazon Polly our users benefit from the most lifelike Text-to-Speech voices available on the market.” Severin Hacker CTO, Duolingo

Slide 40

Slide 40 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Amazon Lex “What’s the weather forecast?” “It will be sunny and 25°C” Weather Forecast Amazon Lex Build Conversational Chatbots

Slide 41

Slide 41 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.

Slide 42

Slide 42 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Amazon Comprehend Discover insights from text Entities Key Phrases Language Sentiment Amazon Comprehend Topic Modeling

Slide 43

Slide 43 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.

Slide 44

Slide 44 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. « Amazon Comprehend helps us analyze the key sentiments, objects, and geos in our 30 million plus reviews & testimonies […] so our customers can make the best decision possible for their travel.” Matt Fryer, VP and Chief Data Science Officer Hotels.com https://aws.amazon.com/solutions/case-studies/expedia/

Slide 45

Slide 45 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. FRAMEWORKS AND INTERFACES PLATFORM SERVICES APPLICATION SERVICES Amazon Rekognition Amazon Polly Amazon Lex ML for everyone. Amazon Rekognition Video Amazon Transcribe Amazon Comprehend Amazon SageMaker AWS DeepLens Amazon EMR Deep Learning AMI Amazon Translate

Slide 46

Slide 46 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. https://www.youtube.com/watch?v=qGotULKg8e0 • Over 10 million images from 300,000 hotels • Fine-tuned a pre-trained Convolutional Neural Network using 100,000 images • Hotel descriptions now automatically feature the best available images Expedia Ranking hotel images using deep learning https://news.developer.nvidia.com/expedia-ranking-hotel-images-with-deep-learning/

Slide 47

Slide 47 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.

Slide 48

Slide 48 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. There’s Never Been A Better Time To Build New Businesses https://aws.amazon.com/machine-learning/

Slide 49

Slide 49 text

© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Thank You. Adrian Hornsby, Cloud Architecture Evangelist @adhorn