IoT, Automation and AI to enrich Human Experience

IoT, Automation and AI to enrich Human Experience

Hassan SawafDirector of Applied Science & Artificial Intelligence

Amazon Web Services

Agenda

• My Motivation

• Amazon Software Services

• Alexa

• AWS

• Use Cases

• Q&A

Agenda

• My Motivation

Agenda

• My Motivation

Personal Background:

• Serial Entrepreneur since mid-90s

• Speech Recognition, Machine Translation

and Computer Vision since 1996

• Daimler Benz, AIXPLAIN AG, AppTek Inc., SAIC, eBay

• now Amazon (AWS AI)

Agenda

• My Motivation

Personal Experiences:

• Business ideas often require complex AI services

• E.g. “real-time speech translation”

• E.g. “personal voice assistant”

• Expensive R&D necessary to establish robust AI services

• Challenges can me prohibitive

for small and large enterprises

Alexa

• Goal:

Ubiquitous

Computing

• Alexa Skills Kit

• Alexa Voice Services

• Alexa Fund

• Smart Home

Amazon Web Services

• Goal:

Ubiquitous

Computing

• Check it out:https://aws.amazon.com

Amazon Web Services

• Goal:

Ubiquitous

Computing

• Check it out:https://aws.amazon.com

Amazon Web Services

• Goal:

Ubiquitous

Computing

• Check it out:https://aws.amazon.com

AWS Internet of Things

• AWS IoT service since October 2015

• Check it out on:

https://aws.amazon.com/iot-platform

AWS Internet of Things

• AWS IoT service since November 2016

• Check it out on:

https://aws.amazon.com/greengrass

Amazon Web Services

• Goal:

Ubiquitous

Computing

• Check it out:https://aws.amazon.com

AWS Machine Learning

• AWS ML service since November 2016

• Check it out on:

https://aws.amazon.com/machine-learning

AWS Lex

• AWS Lex service since November 2016

• Check it out on:

https://aws.amazon.com/iot-platform

The Advent Of Conversational Interactions

1st Gen: Machine-oriented interactions

The Advent Of Conversational Interactions

1st Gen: Machine-oriented interactions

2nd Gen: Control-oriented& translated

The Advent Of Conversational Interactions

1st Gen: Machine-oriented interactions

2nd Gen: Control-oriented& translated

3rd Gen: Intent-oriented

AI ServicesAmazon

Rekognition

Amazon AI: Democratized Artificial Intelligence

AI ServicesAmazon

Rekognition

Amazon

Polly

Amazon AI: Democratized Artificial Intelligence

AI ServicesAmazon

Rekognition

Amazon

Polly

Amazon

Lex

Amazon AI: Democratized Artificial Intelligence

AI ServicesAmazon

Rekognition

Amazon

Polly

Amazon

Lex

More to come

in 2017

Amazon AI: Democratized Artificial Intelligence

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 Services

AI Platform

AI Engines

Amazon

Rekognition

Amazon

Polly

Amazon

Lex

More to come

in 2017

Amazon

Machine Learning

Amazon Elastic

MapReduce

Spark &

SparkML

More to come

in 2017

Apache

MXNetTensorFlow Caffe Theano KerasTorch CNTK

Amazon AI: Democratized Artificial Intelligence

AI Services

AI Platform

AI Engines

Amazon

Rekognition

Amazon

Polly

Amazon

Lex

More to come

in 2017

Amazon

Machine Learning

Amazon Elastic

MapReduce

Spark &

SparkML

More to come

in 2017

Apache

MXNetCaffe Theano KerasTorch CNTK

Amazon AI: Democratized Artificial Intelligence

TensorFlow

P2 ECS Lambda GreenGrass FPGAEMR/Spark

More to

come

in 2017

Hardware

Autonomous Driving Systems

Computational

Knowledge Engine

Pinterest

Visual Search

Pinterest Lens

Recommendations & Ranking At Netflix

Personalized ranking,

page generation,

search, similarity, ratings

In 140 new countries,

simultaneously

Rekognition: Object & Scene Detection

Rekognition: Facial Detection

Model

Training

Amazon AI: Building Intelligent Systems

Model

Training

Amazon AI: Building Intelligent Systems

Inference

in the Cloud

Model

Training

Amazon AI: Building Intelligent Systems

Inference

in the Cloud

Inference

at the Edge

Apache MXNet

Programmable Portable High Performance

Near linear scaling

across hundreds of GPUs

Highly efficient

models for mobile

and IoT

Simple syntax,

multiple languages

Why Apache MXNet?

Most Open Best On AWS

Optimized for

deep learning on AWS

Accepted into the

Apache Incubator

(Integration with AWS)

0

4

8

12

16

1 2 4 8 16

Ideal

Inception v3Resnet

Alexnet91%Efficiency

Amazon AI: Scaling With MXNet

0

64

128

192

256

1 2 4 8 16 32 64 128 256

Amazon AI: Scaling With MXNet

Ideal

Inception v3Resnet

Alexnet

88%Efficiency

0

64

128

192

256

1 2 4 8 16 32 64 128 256

Amazon AI: Scaling With MXNet

Apache MXNet Background

MXNet Overview

• Founded by: U.Washington, Carnegie Mellon U. (~1.5yrs old)

• Recently Accepted to the Apache Incubator

• State of the Art Model Support: Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM)

• Ultra-scalable: Near-linear scaling equals fastest time to model

• Multi-language: Support for Scala, Python, R, etc.. for legacy code leverage and easy integration with Spark

• Ecosystem: Vibrant community from Academia and Industry

Open Source Project on Github | Apache-2 Licensed

Collaborations and Community

4th DL Framework in Popularity

(Outpacing Torch, CNTK and Theano)

0 27.5 55 82.5 110 137.5

TensorFlow

Caffe

Keras

MXNet

Theano

Deeplearning4j

CNTK

Torch7

Popularity

Diverse Community(Spans Industry and Academia)

0 15000 30000 45000 60000

Bing Xu (Apple)

Tianqi Chen (UW)

Mu Li (CMU/AWS)

Eric Xie (UW/AWS)

Yizhi Liu (Mediav)

Chiyuan Zhang (MIT)

Tianjun Xiao (Micrsoft)

Yutian Li (Face++)

Guo Jian (Tusimple)

Guosheng Dong (sogou)

Yu Zhang (MIT)

Depeng Liang (?)

Qiang Kou (Indiana U)

Xingjian Shi (HKUST)

Naiyan Wang (Tusimple)

Top Contributors

Roadmap / Areas of Investment

• NNVM Migration (complete)

• Apache project (Accepted and transitioning to Apache)

• Usability

• Keras Integration WIP (Expected by Q2)

• MinPy being merged (Dynamic Computation graphs, Std Numpyinterface)

• Documentation (installation, native documents, etc.)

• Tutorials, examples

• Platform support(Linux, Windows, OS X, mobile …)

• Language bindings(Python, C++, R, Scala, Julia, JavaScript …)

• Sparse datatypes and LSTM performance improvements

• Deploy your model your way: Lambda, EC2/Docker, Raspberry Pi

Application Examples | Python notebooks

• https://github.com/dmlc/mxnet-notebooks

• Basic concepts

• NDArray - multi-dimensional array computation

• Symbol - symbolic expression for neural networks

• Module - neural network training and inference

• Applications

• MNIST: recognize handwritten digits

• Check out the distributed training results

• Predict with pre-trained models

• LSTMs for sequence learning

• Recommender systems

• Train a state of the art Computer Vision model (CNN)

• Lots more..

Call to Action

MXNet Resources:

• MXNet Blog Post | AWS Endorsement

• Read up on MXNet and Learn More: mxnet.io

• MXNet Github Repo

• MXNet Recommender Systems Talk | Leo Dirac

Developer Resources:

• Deep Learning AMI |Amazon Linux

• Deep Learning AMI | Ubuntu – NEW!!!

• P2 Instance Information

• CloudFormation Template Instructions

• Deep Learning Benchmark

• MXNet on Lambda

• MXNet on ECS/Docker

• MXNet on Raspberry Pi | Wine Detector

Thank you!

[email protected]

Joseph Spisak

Manager | Product Mgmt

AI & Deep Learning

[email protected]

Hassan Sawaf

Director

AI & Applied Sciences