Introduction to AI on AWS Boaz Ziniman - Technical Evangelist, AWS - - PowerPoint PPT Presentation

▶
introduction to ai on aws
SMART_READER_LITE
LIVE PREVIEW

Introduction to AI on AWS Boaz Ziniman - Technical Evangelist, AWS - - PowerPoint PPT Presentation

Dec 31, 2022 126 likes •561 views

Introduction to AI on AWS Boaz Ziniman - Technical Evangelist, AWS @ziniman GPU Technology Conference Tel Aviv October 2017 2006 EC2 & S3 2017 90+ Managed Services Kinesis Streaming Analytics, Mobile Testing, Redshift

slide-1
SLIDE 1

@ziniman

Introduction to AI on AWS

GPU Technology Conference – Tel Aviv October 2017

Boaz Ziniman - Technical Evangelist, AWS

slide-2
SLIDE 2

2006

EC2 & S3

2017

90+ Managed Services

Kinesis Streaming Analytics, Mobile Testing, Redshift Datawarehouse, Code Deploy/Build Tools, Elastic Container Service, Application Load Balancer, Lambda, API Gateway, DynamoDB, Elastic Map, AI, Reduce (Hadoop/Spark/Presto/etc), Elastic Beanstalk, Elastic Transcoder, RDS, Elasticsearch, IoT, more..

slide-3
SLIDE 3

AWS Rapid Pace Of Innovation

slide-4
SLIDE 4

Most Fully Featured Technology Platform

slide-5
SLIDE 5

The difficulty comes in writing software that will make sense of the data

slide-6
SLIDE 6

A system or service which can perform tasks that usually require human intelligence

Artificial Intelligence

slide-7
SLIDE 7
slide-8
SLIDE 8
slide-9
SLIDE 9
slide-10
SLIDE 10

25,000 skills

slide-11
SLIDE 11

And a few more examples…

Fr Fraud detecti ction

  • n

Detecting fraudulent transactions, filtering spam emails, flagging suspicious reviews, …

Pe Personaliza zation

Recommending content, predictive content loading, improving user experience, …

Ta Targeted marketing

Matching customers and offers, choosing marketing campaigns, cross-selling and up-selling, …

Co Conten ent t classif lassific icatio tion

Categorizing documents, matching hiring managers and resumes, …

Chur Churn n pr predic edictio tion

Finding customers who are likely to stop using the service, free- tier upgrade targeting, …

Cus Customer mer suppo support

Predictive routing of customer emails, social media listening, …

slide-12
SLIDE 12

Machine Learning On AWS Today

slide-13
SLIDE 13

Amazon AI Ecosystem

slide-14
SLIDE 14

Amazon AI Ecosystem

slide-15
SLIDE 15

General-purpose GPU compute applications. Features:

  • High Frequency Intel Xeon E5-2686v4 (Broadwell)
  • High-performance NVIDIA K80 GPUs, each with 2,496

parallel processing cores and 12GiB of GPU memory

  • Supports GPUDirect™
  • Enhanced Networking
  • EBS-optimized by default at no additional cost

Amazon EC2 P2 Instances

Instance Size GPUs GPU Peer to Peer vCPUs Memory (GiB) Network Bandwidth* p2.xlarge 1

  • 4

61 1.25Gbps p2.8xlarge 8 Y 32 488 10Gbps p2.16xlarge 16 Y 64 732 20Gbps

slide-16
SLIDE 16

Expedia - Ranking Hotel Images

  • Leading online travel company
  • How to improve hotels listing images?
  • Use GPUs and deep learning to rank hotel images
  • Build a model based on 100K images
  • How to rank 10M images?
  • Use data parallelization across multiple GPUs on AWS
  • Improve ranking time from more than a week to 1 day
slide-17
SLIDE 17

Amazon AI Ecosystem

slide-18
SLIDE 18

On One-Cl Click ck De Deep Le Lear arnin ing

AWS Deep Learning AMIs Amazon Linux & Ubuntu Up to~40k CUDA cores

Ap Apach che MX MXNet

TensorFlow Theano Keras Caffe CNTK Torch Pre-configured CUDA drivers Anaconda, Python3 Out-of-the-box Tutorials

+ + CloudF udFormation n templ plate + + Cont ntaine ner Image ge

Av Available in the AW AWS Ma Mark rketplace

slide-19
SLIDE 19

Amazon AI Ecosystem

slide-20
SLIDE 20

Can We Help Customers Put Intelligence At The Heart Of Every Application & Business?

slide-21
SLIDE 21

Amazon AI Ecosystem

slide-22
SLIDE 22

Pol

  • lly

Te Text-to to-Speec Speech

Artificial Intelligence Services on AWS

slide-23
SLIDE 23

Amazon Polly “Today in Seattle, WA it’s 11°F” “Today in Seattle Washington it’s 11 degrees Fahrenheit”

Text In, Life-like Speech Out

slide-24
SLIDE 24

“Today in Seattle, WA, it’s 11°F” ‘"We live for the music" live from the Madison Square Garden.’

1. . Automatic, , Accurate Text Pr Processing

A Focus On Voice Quality & Pronunciation

slide-25
SLIDE 25

2. . Intelligi gible and Easy to Un Understand

  • 1. Automatic, Accurate Text Processing

A Focus On Voice Quality & Pronunciation

slide-26
SLIDE 26

https://www.w3.org/TR/speech-synthesis/ <speak> The spelling of my name is <prosody rate='x-slow'> <say-as interpret-as="characters">Boaz</say-as> </prosody> </speak>

A Focus On Voice Quality & Pronunciation

slide-27
SLIDE 27
  • 2. Intelligible and Easy to Understand

3. . Add Se Semantic Meaning g to Text

“Richard’s number is 2122341237“ “Richard’s number is 2122341237“

Te Telephone Number

  • 1. Automatic, Accurate Text Processing

A Focus On Voice Quality & Pronunciation

slide-28
SLIDE 28
  • 2. Intelligible and Easy to Understand
  • 3. Add Semantic Meaning to Text

4. . Customized Pr Pronunciation

“My daughter’s name is Kaja.” “My daughter’s name is Kaja.”

  • 1. Automatic, Accurate Text Processing

A Focus On Voice Quality & Pronunciation

slide-29
SLIDE 29

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

  • n the market.

Se Seve verin Ha Hacker CTO, Duolingo ”

“

  • Spoken language crucial for language learning
  • Accurate pronunciation matters
  • Faster iteration thanks to TTS
  • As good as natural human speech
slide-30
SLIDE 30

Rek ekog

  • gnition
  • n

Im Imag age A e Analy nalysis is

Artificial Intelligence Services on AWS

slide-31
SLIDE 31

Amazon Rekognition

Object and scene detection Facial analysis Face comparison Celebrity recognition Image moderation

Amazon Rekognition: Images In, Rich Metadata Out

slide-32
SLIDE 32

Object & Scene Detection

slide-33
SLIDE 33

Facial Analysis

slide-34
SLIDE 34

Celebrity Recognition

slide-35
SLIDE 35

Facial Search

slide-36
SLIDE 36

Explicit Nudity Nudity Graphic Male Nudity Graphic Female Nudity Sexual Activity Partial Nudity Suggestive Female Swimwear or Underwear Male Swimwear or Underwear Revealing Clothes

Image moderation

slide-37
SLIDE 37

Amazon Rekognition Customers

  • Digital Asset Management
  • Media and Entertainment
  • Travel and Hospitality
  • Influencer Marketing
  • Systems Integration
  • Digital Advertising
  • Consumer Storage
  • Law Enforcement
  • Public Safety
  • eCommerce
  • Education
slide-38
SLIDE 38

https://aws.amazon.com/solutions/case-studies/cspan/

slide-39
SLIDE 39

Lex ex

Co Conv nver ersatio sation Bo Bots ts

Artificial Intelligence Services on AWS

slide-40
SLIDE 40

Ama Amazon

  • n Lex

Automatic Speech Recognition Natural Language Understanding

“Wha “What’s t ’s the he w weathe her r foreca cast?”

Weather Forecast

Speech Recognition & Natural Language Understanding

slide-41
SLIDE 41

Ama Amazon

  • n Lex

Automatic Speech Recognition Natural Language Understanding

“Wha “What’s t ’s the he w weathe her r foreca cast?” “I “It w will be be sunny sunny an and 25°C”

Weather Forecast

Speech Recognition & Natural Language Understanding

slide-42
SLIDE 42

Wrap up

slide-43
SLIDE 43

@ziniman

Thank You!

Boaz Ziniman - Technical Evangelist, AWS

@AWScloud for Global AWS News & Announcements Local Events: https://aws.amazon.com/events/aws-israel/