AWS offers a family of intelligent services that provide cloud-native machine learning and deep learning technologies to address your different use cases and needs. For developers looking to add managed AI services to their applications, AWS brings natural language understanding (NLU) and text-to-speech (TTS) with Amazon Polly, visual search and image recognition with Amazon Rekognition, and developer-focused machine learning with Amazon Machine Learning. In this talk you will learn about these services and see demos of their capabilities
AWS Speaker: Denis V. Batalov, Solutions Architect - Amazon Web Services
Customer Speaker: Tom Wells - Synthesis Software Technologies
4. Thousands Of Employees Across The Company Focused on AI
Discovery &
Search
Fulfilment &
Logistics
Enhance
Existing Products
Define New
Product
Categories
Bring Machine
Learning To All
Artificial Intelligence At Amazon (2017)
8. AI Applications on AWS
Zillow
• Zestimate (using Apache Spark)
Howard Hughes Corp
• Lead scoring for luxury real estate
purchase predictions
FINRA
• Anomaly detection, sequence matching,
regression analysis, network/tribe analysis
Netflix
• Recommendation engine
Pinterest
• Image recognition search
Fraud.net
• Detect online payment fraud
DataXu
• Leverage automated & unattended ML at
large scale (Amazon EMR + Spark)
Mapillary
• Computer vision for crowd sourced maps
Hudl
• Predictive analytics on sports plays
Upserve
• Restaurant table mgmt & POS for
forecasting customer traffic
TuSimple
• Computer Vision for Autonomous Driving
Clarifai
• Computer Vision APIs
9. AI Services
AI Platform
Amazon
Rekognition
Amazon
Polly
Amazon
Lex
More to come
in 2017
Amazon
Machine Learning
Amazon Elastic
MapReduce
Spark &
SparkML
More to come
in 2017
Amazon AI: Democratized Artificial Intelligence
AI Engines
Apache
MXNet
Caffe Theano KerasTorch CNTKTensorFlow
P2 ECS Lambda GreenGrass FPGAEMR/Spark
More to
come
in 2017
Hardware
17. “How many successful innovation projects have you
run, and how much additional revenue did they
make?”
à
“How many failed innovation projects have you run,
and how much did they cost?”
55. API
ECS Cluster
t2.micro t2.micro
Containers
luno-trade-feed
(service)
DynamoDB
luno-trades
extract-to-s3
(once-off task)
S3
luno-trades
Extract
Store
CloudWatch Event
Midnight UTC
Trigger
Lambda
Run ECS Task
Run Task
WS://
European Central Bank
Website
(.xml)
Lambda
Extract ECB Rates
Trigger
Store
Lambda
Extract BTC/USD Prices
blockchain.info
(.csv)
Scrape Store
Trigger
Scrape
DynamoDB
currency-rates
DynamoDB
btc-usd-trades
Extract
Extract
bitfinex-trade-feed
(service)
API
WS://
Store
EMR Cluster
(Spot Instances)
c4.xlarge c4.xlarge
Jobs
fancy predictive stuff
…
56. API
ECS Cluster
t2.micro t2.micro
Containers
luno-trade-feed
(service)
extract-to-s3
(once-off task)
S3
luno-trades
Extract
Store
CloudWatch Event
Midnight UTC
Trigger
Lambda
Run ECS Task
Run Task
WS://
European Central Bank
Website
(.xml)
Lambda
Extract ECB Rates
Trigger
Store
Lambda
Extract BTC/USD Prices
blockchain.info
(.csv)
Scrape Store
Trigger
Scrape
DynamoDB
currency-rates
DynamoDB
btc-usd-trades
Extract
Extract
bitfinex-trade-feed
(service)
API
WS://
Store
tradebot-prediction
(service)
Lambda
Dynamo Stream to Kinesis
Kinesis Stream
luno-trades
Kinesis Stream
currency-rates
Kinesis Stream
btc-usd-trades
New Item
New Item
New Item
Put
Put
Put
Get Records
DynamoDB
predicIonsStore
Extract
DynamoDB
luno-trades
57.
58.
59. What did we learn
Ability to predict BTC ZAR price based on USD seems possible..?
Ability to quickly dump and store data for later analysis hugely valuable
Spark not so great for time-series data (?)
Future plans incl. AWS ML, build actual trade-execution, build lots more
ML models and race them against each other
65. Amazon AI Services
• Leveraging Amazon internal experiences with AI / ML
• Managed API services with embedded AI for maximum
accessibility and simplicity
• Full stack of platforms and engines for specialized deep
learning applications
66. Converts text
to life-like speech
47 voices 24 languages Low latency,
real time
Fully managed
Polly: Life-like Speech Service
Voice Quality & Pronunciation
1. Automatic, Accurate Text Processing
2. Intelligible and Easy to Understand
3. Add Semantic Meaning to Text
4. Customized Pronunciation
Articles and Blogs
Training Material
Chatbots (Lex)
Public Announcements
67. Amazon Polly: Quality
Natural sounding speech
A subjective measure of how close TTS output is to human speech.
Accurate text processing
Ability of the system to interpret common text formats such as abbreviations, numerical
sequences, homographs etc.
Today in Las Vegas, NV it's 54°F.
"We live for the music", live from the Madison Square Garden.
Highly intelligibile
A measure of how comprehensible speech is.
”Peter Piper picked a peck of pickled peppers.”
68. Amazon Polly: Language Portfolio
Americas:
• Brazilian Portuguese
• Canadian French
• English (US)
• Spanish (US)
A-PAC:
• Australian English
• Indian English
• Japanese
EMEA:
• British English
• Danish
• Dutch
• French
• German
• Icelandic
• Italian
• Norwegian
• Polish
• Portuguese
• Romanian
• Russian
• Spanish
• Swedish
• Turkish
• Welsh
• Welsh English
69. Recording Data for TTS
Tons of text
Recording script:
Few weeks of
recordings
Automatic
selection of
texts
Recording script:
• Covers all combinations of diphones
and significant features in a
language
70. an error occurred while searching for your route
because snaps weren't all so obedient anymore,
now we say apple again. and we say apple,
general electric soars today. information on general electric
quick breads, zucchini, holiday, crock pot, cake,
so are you still keeping tabs on your old team,
that weighs more than four tons, disrupts the herring's swim
…
An apple a day, keeps …
71. Duolingo voices its language learning service Using Polly
Duolingo is a free language learning service where
users help translate the web and rate translations.
With Amazon Polly our users
benefit from the most lifelike
Text-to-Speech voices
available on the market.
Severin Hacker
CTO, Duolingo
”
“ • Spoken language crucial for
language learning
• Accurate pronunciation matters
• Faster iteration thanks to TTS
• As good as natural human speech
72. Amazon Rekognition
Deep learning-based image recognition service
Search, verify, and organize millions of images
Object and Scene
Detection
Facial
Analysis
Facial
Recognition
Image Moderation
Integrated with S3, Lambda, Polly, Lex
73. Object and Scene Detection
Generate labels for thousands of objects, scenes, and
concepts, each with a confidence score
• Search, filter, and
curate image
libraries
• Smart searches for
user generated
content
• Photo, travel, real
estate, vacation
rental applications
Maple
Plant
Villa
Garden
Water
Swimming Pool
Tree
Potted Plant
Backyard
74. Facial Analysis
Locate faces within images and analyze face attributes to
detect emotion, pose, facial landmarks, and features
• Avoid faces when cropping
images and overlaying ads
• Capture user demographics
and sentiment
• Recommend the best photos
• Improve online dating match
recommendations
• Dynamic, personalized ads
75. Face Comparison
Measure the likelihood that faces in two images are of the
same person
• Add face verification to
applications and devices
• Extend physical security
controls
• Provide guest access to
VIP-only facilities
• Verify users for online
exams and polls
76. Facial Recognition
Identify people in images by finding the closest match for an
input face image against a collection of stored face vectors
• Add friend tagging to
social and messaging apps
• Assist public safety officers
find missing persons
• Identify employees as they
access sensitive locations
• Identify celebrities in
historical media archives
77. Media Case Study
Identify who is on camera at what time for each of 8 networks
so that recorded video streams can be indexed and searched
Video frame-sampling facial recognition solution using
Amazon Rekognition:
• Indexed 97,000 people into a face collection in 1 day
• Sample frames every 6 secs and test for image variance
• Upload images to S3 and call Rekognition to find best facial match
• Store time stamp and faceID metadata
81. Significantly improve many applications on multiple domains
“deep learning” trend in the past 10 years
image understanding speech recognition natural language
processing
…
Deep Learning
autonomy
85. Deploy Everywhere
Beyond
BlindTool by Joseph Paul Cohen, demo on Nexus 4
Fit the core library with all dependencies
into a single C++ source file
Easy to compile on …
Amalgamation
Runs in browser
with Javascript
The first image for
search “dog” at
images.google.com
Outputs “beagle”
with prob = 73%
within 1 sec
86. TX1 on Flying Drone
TX1 with customized board
Drone
Realtime detection and tracking on TX1
~10 frame/sec with 640x480 resolution
87. New P2 Instance | Up to 16 GPUs
§This new instance type incorporates up to 8 NVIDIA Tesla
K80 Accelerators, each running a pair of
NVIDIA GK210 GPUs.
§Each GPU provides 12 GiB of memory (accessible via 240
GB/second of memory bandwidth), and 2,496 parallel
processing cores.
§Available in PDX, IAD and DUB RegionsInstance Name GPU Count vCPU Count Memory
Parallel
Processing Cores
GPU Memory
Network
Performance
p2.xlarge 1 4 61 GiB 2,496 12 GiB High
p2.8xlarge 8 32 488 GiB 19,968 96 GiB 10 Gigabit
p2.16xlarge 16 64 732 GiB 39,936 192 GiB 20 Gigabit
88. One-Click GPU
Deep Learning
AWS Deep Learning AMI
Up to~40k CUDA cores
MXNet
TensorFlow
Theano
Caffe
Torch
Pre-configured CUDA drivers
Anaconda, Python3
+ CloudFormation template
+ Container Image
89. AWS Marketplace
Discover, Procure, Deploy, and Manage Software In the Cloud
• 3,800+ software listings
• Over 1,200 participating ISVs
• Open source and commercial
software
• Bring-your-own-license
• Procure new
• Deployed in Most AWS
Regions
• 135,000+ active customers
• Over 370M of deployed EC2
instances per month