Helpicto Architecture
DATA PREP BUILD, TRAIN & DEPLOY INTELLIGENT APPS
Stored on
Azure Blob and SQL DB
Xamarin
Helpicto Images
Cognitive Services,
Azure Functions
Bringing the best of AI to Azure and the best of Azure to AI
Pre-Built AI
Azure Cognitive Services
Conversational AI
Azure Bot Service
Custom AI
Azure Machine Learning
How Developers Use AI in Azure
Build Modern Applications with AI
Understand & interpret meaning of data
including text, voice, images
Build Conversational Applications
Engage with users in natural ways
Optimize Business Processes with AI
Reason, learn & form conclusions from data
Azure
Cognitive Services
Azure
Bot Services
Azure
Machine Learning
Powerful prebuilt AI models exposed as API services
Simple REST APIs with .NET, Java, Python, Node SDKs
Easily customize for highest accuracy
Train in the cloud and deploy anywhere
Vision
Speech
Language
Conversation
Bing Search
Knowledge
1. Object Detection & Recognition for thousands of objects
2. Video Indexer (Preview)
3. Speech Recognition with customization
4. Speech Synthesis with customizable voice
5. Speech to Speech Translation
6. Text analytics with entity detection
7. Language Understanding (LUIS) with new integrated offer
8. QnA Maker is Generally Available
9. Bing Visual Search with smart identification
10.Bing Search SDK is Generally Available
Vision
Speech
Language
Conversation
Bing Search
Knowledge
Vision
Speech
Language
Conversation
Bing Search
Knowledge
Azure Search with
Cognitive Services
Built-in enrichment of content using AI models
Enable immersive search experiences over any data
Brian Smith
Principal Engineering Manager
@BrianSmithMSFT
November 22, 1963
John F. Kennedy (JFK)
HOW DO WE DO THIS?
Ingest Enrich Explore
Ingest Enrich Explore
Data Cognitive Skills Azure Search
AN AI–FIRST APPROACH TO CONTENT UNDERSTANDING
Cognitive Search
JFK FILES
COGNITIVE SEARCH
ARCHITECTURE
Explore
Azure
Storage
Azure Functions
-Cryptonyms
-Redactions
Cognitive Skills
-OCR + Handwriting
-Computer Vision
-Entities
Azure ML
Search Index
Azure Search
The JFK Files Architecture
Content
Extraction
ANALYZING
LEGAL CONTRACTS
UNDERSTANDING
ENGINEERING PLANS
EXTRACTING
FORM INFORMATION
Bring AI to your data
AZURE SEARCH PORTAL JFK FILES ON GITHUB
https://aka.ms/jfk-fileshttps://portal.azure.com
Get started today
Azure Search with built-in AI
COGNITIVE SEARCH BREAKOUT SESSION
Learn more at session 3201
Learn more at //build
Conversational AI
American Express Briefing – Jan 17, 2018
Chris Mullins
Group Program Manager, Azure Bot Service
190,000 total bots registered
33000 active bots in service
>3200 bots created per week
What is a Bot?
{ Your Code }
Intelligence Cognitive Services
Tools Bot Framework
HTTP
REST Endpoint
Services Bot Service
Cortana OthersFacebookSkype
Conversational and
Business logic
Development
tools
Knowledge
Cognitive Search Custom Speech
Understanding
Custom Voice
Personality
Azure Bot Service
Conversational AI
Noelle LaCharite
Principal Program Manager
@NoelleLaCharite
Wilson Lee
Senior Software Engineer
@SimplyWilson
BRK 3235:
BRK 3208:
THR 3118:
Building your own AI models for Transforming Data into Intelligence
Prepare Data Build & Train Deploy
Quickly launch and scale Spark on demand
Rich interactive workspace and notebooks
Seamless integration with all Azure data services
Azure Databricks
Step 1: Prepare Data
Apache® Spark™ based analytics
platform optimized for Azure
Broad frameworks and tools support:
 TensorFlow, Cognitive Toolkit, Caffe2, Keras, MxNET, PyTorch
Scale training from 1  100,000’s of servers
Azure ML Python SDK
Azure ML packages:
 Computer Vision, Text, Forecasting
Azure Machine Learning
Step 2: Build and Train
Docker
Containers
Azure Kubernetes Service (AKS)
Azure Batch
Azure IoT EdgeAzure
Machine Learning
Step 3: Deploy
Any other container host…
Integrated with Azure Machine Learning
Create new deep learning projects easily
Monitor model training progress & GPU utilization
Visualize your model performance with integrated
open tools like TensorBoard
Get started quickly with the Samples Gallery
The Machine Learning framework made by and for .NET developers
Proven & Extensible
Open Source
Supported on Windows, Linux, and macOS
Developer Focused
Join at github.com/dotnet/machinelearning
ML.NET Preview
Cross-platform Open Source Machine learning framework for .NET
Extensively used across Microsoft: Windows, Bing, Azure
High productivity throughout the entire ML workflow
Extensible to other frameworks (TensorFlow, CNTK…)
Learn more: BRK 3203 on Tuesday, May 8th at 1:15pm.
ML.NET Usage at Microsoft
+ more!
Windows 10
Power Point
Excel
Bing
Forecasting
Text Analytics
Supported on Windows, Linux, and macOS
Computer Vision
Azure ML Packages Preview
Python packages for Computer Vision, Forecasting and Text Analytics
Easily build and deploy models on Azure Machine Learning
Provides high level APIs for data preparation, augmentation, training,
evaluating and deployment.
Model experimentation and comparison, run history, model
management and deployment through Azure ML
Join at BRK 3226 What's new with Azure Machine Learning (breakout session)
Mark Hamilton
Software Engineer
github.com/mhamilton723
The Machine Learning framework made by and for .NET developers
Supported on Windows, Linux, and macOS
Azure ML and Project Brainwave
Real-time AI at cloud scale with industry-leading performance and lowest cost
Models are very easy to create and deploy as webservices
Record-setting DNN performance with accelerated ResNet50
Record speed: ResNet 50 on FPGA in <1.8 ms per image
Lowest cost: Only 21 cents per million images during preview
More accelerated models coming soon
FPGAs for ultra-fast inferencing
DATA PREP BUILD TRAIN DEPLOY
Stored on
Azure Premium Storage
Azure Machine
Learning
Jabil Classification Model
ResNet-50
Circuit board images
Jupyter Notebook
Azure Machine Learning
Ultra-fast Inferencing
using FPGAs
Ted Way, PhD
Senior Program Manager
@tedwinway
Planet scale Real-Time Inferencing for GeoAI
svFreely Available Imagery Labeled Training Data Inferred Land Cover Map
415K
images/sec
FPGA
Real-Time Low-Latency Inferencing with FPGAs
Setup:
800 FPGAs on Azure
195 Million Images; 20TB
Real-time inferencing 1 image at a time
Results:
415K inferences/second @ 1.8ms latency
10.6 minutes total
Order of magnitude better Price/Perf
over CPU & GPU (V100 with TensorRT
BRK3202
BRK3205
BRK2508
What we demonstrated
Build Modern Applications with AI
1. Content based indexing, search and analysis
Build Conversational Applications
2. Engaging with users in natural ways through chats,
speech, voice and translation
Optimize Business Processes with AI
3. Building custom AI models for classification
4. Real-time AI with FPGAs and edge support
Cognitive Search
(Azure Search +
Azure Cognitive Services)
Azure
Bot Services
Azure
Machine Learning
Roger Luo,
President
Why you should develop your next AI app on Azure
1. Broadest set of Pre-Built AI capabilities
2. Customizable and flexible AI services
3. Most advanced conversational AI
4. Most differentiated support for AI @ the Edge + unique
hardware for AI
5. The strongest enterprise cloud for Data + AI
The Microsoft AI platform: a State of the Union
The Microsoft AI platform: a State of the Union
The Microsoft AI platform: a State of the Union
The Microsoft AI platform: a State of the Union

The Microsoft AI platform: a State of the Union

  • 6.
    Helpicto Architecture DATA PREPBUILD, TRAIN & DEPLOY INTELLIGENT APPS Stored on Azure Blob and SQL DB Xamarin Helpicto Images Cognitive Services, Azure Functions
  • 7.
    Bringing the bestof AI to Azure and the best of Azure to AI Pre-Built AI Azure Cognitive Services Conversational AI Azure Bot Service Custom AI Azure Machine Learning
  • 8.
    How Developers UseAI in Azure Build Modern Applications with AI Understand & interpret meaning of data including text, voice, images Build Conversational Applications Engage with users in natural ways Optimize Business Processes with AI Reason, learn & form conclusions from data Azure Cognitive Services Azure Bot Services Azure Machine Learning
  • 9.
    Powerful prebuilt AImodels exposed as API services Simple REST APIs with .NET, Java, Python, Node SDKs Easily customize for highest accuracy Train in the cloud and deploy anywhere Vision Speech Language Conversation Bing Search Knowledge
  • 11.
    1. Object Detection& Recognition for thousands of objects 2. Video Indexer (Preview) 3. Speech Recognition with customization 4. Speech Synthesis with customizable voice 5. Speech to Speech Translation 6. Text analytics with entity detection 7. Language Understanding (LUIS) with new integrated offer 8. QnA Maker is Generally Available 9. Bing Visual Search with smart identification 10.Bing Search SDK is Generally Available Vision Speech Language Conversation Bing Search Knowledge
  • 12.
    Vision Speech Language Conversation Bing Search Knowledge Azure Searchwith Cognitive Services Built-in enrichment of content using AI models Enable immersive search experiences over any data
  • 13.
    Brian Smith Principal EngineeringManager @BrianSmithMSFT
  • 14.
    November 22, 1963 JohnF. Kennedy (JFK)
  • 15.
    HOW DO WEDO THIS? Ingest Enrich Explore
  • 16.
    Ingest Enrich Explore DataCognitive Skills Azure Search AN AI–FIRST APPROACH TO CONTENT UNDERSTANDING Cognitive Search
  • 18.
    JFK FILES COGNITIVE SEARCH ARCHITECTURE Explore Azure Storage AzureFunctions -Cryptonyms -Redactions Cognitive Skills -OCR + Handwriting -Computer Vision -Entities Azure ML Search Index Azure Search The JFK Files Architecture Content Extraction
  • 19.
  • 20.
    AZURE SEARCH PORTALJFK FILES ON GITHUB https://aka.ms/jfk-fileshttps://portal.azure.com Get started today
  • 21.
    Azure Search withbuilt-in AI COGNITIVE SEARCH BREAKOUT SESSION Learn more at session 3201 Learn more at //build
  • 23.
    Conversational AI American ExpressBriefing – Jan 17, 2018 Chris Mullins Group Program Manager, Azure Bot Service
  • 24.
    190,000 total botsregistered 33000 active bots in service >3200 bots created per week
  • 25.
    What is aBot? { Your Code } Intelligence Cognitive Services Tools Bot Framework HTTP REST Endpoint Services Bot Service Cortana OthersFacebookSkype Conversational and Business logic Development tools
  • 26.
    Knowledge Cognitive Search CustomSpeech Understanding Custom Voice Personality Azure Bot Service Conversational AI
  • 27.
    Noelle LaCharite Principal ProgramManager @NoelleLaCharite Wilson Lee Senior Software Engineer @SimplyWilson
  • 38.
  • 40.
    Building your ownAI models for Transforming Data into Intelligence Prepare Data Build & Train Deploy
  • 41.
    Quickly launch andscale Spark on demand Rich interactive workspace and notebooks Seamless integration with all Azure data services Azure Databricks Step 1: Prepare Data Apache® Spark™ based analytics platform optimized for Azure
  • 42.
    Broad frameworks andtools support:  TensorFlow, Cognitive Toolkit, Caffe2, Keras, MxNET, PyTorch Scale training from 1  100,000’s of servers Azure ML Python SDK Azure ML packages:  Computer Vision, Text, Forecasting Azure Machine Learning Step 2: Build and Train
  • 43.
    Docker Containers Azure Kubernetes Service(AKS) Azure Batch Azure IoT EdgeAzure Machine Learning Step 3: Deploy Any other container host…
  • 44.
    Integrated with AzureMachine Learning Create new deep learning projects easily Monitor model training progress & GPU utilization Visualize your model performance with integrated open tools like TensorBoard Get started quickly with the Samples Gallery
  • 45.
    The Machine Learningframework made by and for .NET developers Proven & Extensible Open Source Supported on Windows, Linux, and macOS Developer Focused Join at github.com/dotnet/machinelearning ML.NET Preview Cross-platform Open Source Machine learning framework for .NET Extensively used across Microsoft: Windows, Bing, Azure High productivity throughout the entire ML workflow Extensible to other frameworks (TensorFlow, CNTK…) Learn more: BRK 3203 on Tuesday, May 8th at 1:15pm.
  • 46.
    ML.NET Usage atMicrosoft + more! Windows 10 Power Point Excel Bing
  • 47.
    Forecasting Text Analytics Supported onWindows, Linux, and macOS Computer Vision Azure ML Packages Preview Python packages for Computer Vision, Forecasting and Text Analytics Easily build and deploy models on Azure Machine Learning Provides high level APIs for data preparation, augmentation, training, evaluating and deployment. Model experimentation and comparison, run history, model management and deployment through Azure ML Join at BRK 3226 What's new with Azure Machine Learning (breakout session)
  • 49.
  • 51.
    The Machine Learningframework made by and for .NET developers Supported on Windows, Linux, and macOS Azure ML and Project Brainwave Real-time AI at cloud scale with industry-leading performance and lowest cost Models are very easy to create and deploy as webservices Record-setting DNN performance with accelerated ResNet50 Record speed: ResNet 50 on FPGA in <1.8 ms per image Lowest cost: Only 21 cents per million images during preview More accelerated models coming soon
  • 53.
    FPGAs for ultra-fastinferencing DATA PREP BUILD TRAIN DEPLOY Stored on Azure Premium Storage Azure Machine Learning Jabil Classification Model ResNet-50 Circuit board images Jupyter Notebook Azure Machine Learning Ultra-fast Inferencing using FPGAs
  • 54.
    Ted Way, PhD SeniorProgram Manager @tedwinway
  • 58.
    Planet scale Real-TimeInferencing for GeoAI svFreely Available Imagery Labeled Training Data Inferred Land Cover Map
  • 59.
    415K images/sec FPGA Real-Time Low-Latency Inferencingwith FPGAs Setup: 800 FPGAs on Azure 195 Million Images; 20TB Real-time inferencing 1 image at a time Results: 415K inferences/second @ 1.8ms latency 10.6 minutes total Order of magnitude better Price/Perf over CPU & GPU (V100 with TensorRT
  • 60.
  • 62.
    What we demonstrated BuildModern Applications with AI 1. Content based indexing, search and analysis Build Conversational Applications 2. Engaging with users in natural ways through chats, speech, voice and translation Optimize Business Processes with AI 3. Building custom AI models for classification 4. Real-time AI with FPGAs and edge support Cognitive Search (Azure Search + Azure Cognitive Services) Azure Bot Services Azure Machine Learning
  • 63.
  • 65.
    Why you shoulddevelop your next AI app on Azure 1. Broadest set of Pre-Built AI capabilities 2. Customizable and flexible AI services 3. Most advanced conversational AI 4. Most differentiated support for AI @ the Edge + unique hardware for AI 5. The strongest enterprise cloud for Data + AI