Yapay Zekâ (AI)
Machine Learning / Deep Learning
Cloud
Computing
Big Data
Powerful
Algorithms
Agent Applications Services Infrastructure
Microsoft AI Portfolio
Cortana Office 365
Dynamics 365
SwiftKey
Pix
Customer Service
and Support
Skype
Calendar.help
Cortana Intelligence
Cognitive Services
Bot Framework
Cortana Devices SDK
Cognitive Toolkit
Azure Machine
Learning
Azure Notebooks
Azure N Series
FPGA
People
Agent Applications Services Infrastructure
Microsoft AI Portfolio
Cortana Office 365
Dynamics 365
SwiftKey
Pix
Customer Service
and Support
Skype
Calendar.help
Cortana Intelligence
Cognitive Services
Bot Framework
Cortana Devices SDK
Cognitive Toolkit
Azure Machine
Learning
Azure Notebooks
Azure N Series
FPGA
People
H Y P E R S C A L E ,
E N T E R P R I S E - G R A D E
I N F R A S T R U C T U R E
D E V E L O P E R T O O L S &
S E R V I C E S
O P E N P L A T F O R M F O R
D A T A S C I E N C E
Hardware
Storage management
Software
T H E M L & A I P L A T F O R M
AI Applications (1st & 3rd party) Cognitive Services Bot Framework
Spark AI Batch Training DS VM SQL Server ACS
BLOB Cosmos DB SQL DB/DW ADLS
CPUs FPGA GPUs IoT
Azure Machine Learning
Model deployment & management
Machine Learning toolkits
Experimentation management,
data prep, & collaboration
CNTK
Tensorflow
ML Server
Scikit-Learn
Other Libs.
PROSE
Docker
Cloud – Spark, SQL, other engines
ML Server – Spark, SQL, VMs
Edge
Machine Learning & AI Portfolio
When to use what?
What engine(s) do you want to use?
Deployment target
Which experience do you want?
Build your own or consume pre-trained models?
Microsoft ML &
AI products
Build your own
Azure Machine
Learning
Code first
(On-prem)
ML Server
On-prem
Hadoop
SQL Server
(cloud)
AML (Preview)
SQL Server Spark Hadoop Azure Batch DSVM Azure Container
Service
Visual tooling
(cloud)
AML Studio
Consume
Cognitive
services, bots
VISUAL DRAG-AND-DROP CODE-FIRST
Azure Machine Learning Studio & Workbench
Azure Notebooks
Use your favorite IDE
Leverage all types of data
Use what you want
U S E T H E M O S T P O P U L A R I N N O V A T I O N S
U S E A N Y T O O L
U S E A N Y F R A M E W O R K O R L I B R A R Y
AML Workbench
Sample, understand, and
prep data rapidly
Support for Spark + Python
+ R (roadmap)
Execute jobs locally, on
remote VMs, Spark clusters,
SQL on-premises
Git-backed tracking of
code, config, parameters,
data, run history
• Column statistics : Numeric
• Histogram
• Value Counts
• Box Plot
• Scatter Plot
• Time Series
• Map
Inspectors
Spark
SQL Server
Virtual machines
GPUs
Container services
Notebooks
Azure Machine Learning Workbench
Visual Studio Code Tools for AI
Visual Studio Tools for AI
PyCharm
SQL Server
Machine Learning Server
O N - P R E M I S E S
E D G E C O M P U T I N G
Azure IoT Edge
Experimentation and
Model Management
A Z U R E M A C H I N E L E A R N I N G S E R V I C E S T R A I N & D E P LO Y O P T I O N S
A Z U R E
Local machine
Scale up to DSVM
Scale out with Spark on HDInsight
Azure Batch AI (Coming Soon)
ML Server (Coming Soon)
A ZURE ML
EXPERIMENTATION
Command line tools
IDEs
Notebooks in Workbench
VS Code Tools for AI
VS Tools for AI
DOCKER
Single node deployment
(cloud/on-prem)
Azure Container Service
Azure IoT Edge
Microsoft ML Server
Spark clusters
SQL Server (Coming Soon)
A ZURE ML
MODEL MANAGEMENT
http://aka.ms/dsvm
Data Science
Virtual Machines
(DSVM)
Data Science
Virtual Machines
(DSVM) DSVM – Windows Server 2016
DSVM – Linux – Ubuntu
Deep Learning Virtual Machines
Yes
Similar
image
Query
image
 Researchers took a traditional machine learning approach
• Example: HoG Detectors
- Histogram of oriented
gradients (HoG) features
- Sliding window detector
- SVM Classifier
- Very fast OpenCV
implementation (<100ms)
Deep Neural Network for Computer Vision
cat? YES
dog? NO
car? NO
Convolutional Layers
Fully
Connected
Layers
Complex Objects
& Scenes
(people, animals,
cars, beach
scene, etc.)
Image
Low-Level Features
(lines, edges,
color fields, etc.)
High-Level Features
(corners, contours,
simple shapes)
Object Parts
(wheels, faces,
windows, etc.)
Cognitive Toolkit
Unlock deeper learning
A free, easy-to-use, open-source toolkit that
trains deep learning algorithms to learn like the
human brain.
Microsoft Cognitive Toolkit
Clothing texture dataset:
Can we apply transfer learning to accurately classify clothing texture?
LeopardStriped Dotted
Pre-Built CNN from General Task on Millions of Images
Output
Layer
Stripped
Outputs of penultimate layer of
ImageNet Trained CNN provide excellent
general purpose image features
cat? YES
dog? NO
car? NO
Classi
fier
e.g.
SVM
dotted?
Pre-Built CNN from General Task on Millions of Images
Output
Layer
Stripped
Using a pre-trained DNN, an accurate
model can be achieved with thousands (or
less) of labeled examples instead of millions
cat? YES
dog? NO
car? NO
dotted?
Train one or more
layers in new network
Ibrahim KIVANÇ
http://www.ibrahimkivanc.com
@ikivanc
ikivanc@microsoft.com

Microsoft & Machine Learning / Artificial Intelligence

  • 1.
  • 2.
    Machine Learning /Deep Learning
  • 3.
  • 4.
    Agent Applications ServicesInfrastructure Microsoft AI Portfolio Cortana Office 365 Dynamics 365 SwiftKey Pix Customer Service and Support Skype Calendar.help Cortana Intelligence Cognitive Services Bot Framework Cortana Devices SDK Cognitive Toolkit Azure Machine Learning Azure Notebooks Azure N Series FPGA People
  • 6.
    Agent Applications ServicesInfrastructure Microsoft AI Portfolio Cortana Office 365 Dynamics 365 SwiftKey Pix Customer Service and Support Skype Calendar.help Cortana Intelligence Cognitive Services Bot Framework Cortana Devices SDK Cognitive Toolkit Azure Machine Learning Azure Notebooks Azure N Series FPGA People
  • 7.
    H Y PE R S C A L E , E N T E R P R I S E - G R A D E I N F R A S T R U C T U R E D E V E L O P E R T O O L S & S E R V I C E S O P E N P L A T F O R M F O R D A T A S C I E N C E Hardware Storage management Software T H E M L & A I P L A T F O R M AI Applications (1st & 3rd party) Cognitive Services Bot Framework Spark AI Batch Training DS VM SQL Server ACS BLOB Cosmos DB SQL DB/DW ADLS CPUs FPGA GPUs IoT Azure Machine Learning Model deployment & management Machine Learning toolkits Experimentation management, data prep, & collaboration CNTK Tensorflow ML Server Scikit-Learn Other Libs. PROSE Docker Cloud – Spark, SQL, other engines ML Server – Spark, SQL, VMs Edge
  • 9.
    Machine Learning &AI Portfolio When to use what? What engine(s) do you want to use? Deployment target Which experience do you want? Build your own or consume pre-trained models? Microsoft ML & AI products Build your own Azure Machine Learning Code first (On-prem) ML Server On-prem Hadoop SQL Server (cloud) AML (Preview) SQL Server Spark Hadoop Azure Batch DSVM Azure Container Service Visual tooling (cloud) AML Studio Consume Cognitive services, bots
  • 10.
    VISUAL DRAG-AND-DROP CODE-FIRST AzureMachine Learning Studio & Workbench
  • 11.
  • 12.
    Use your favoriteIDE Leverage all types of data Use what you want U S E T H E M O S T P O P U L A R I N N O V A T I O N S U S E A N Y T O O L U S E A N Y F R A M E W O R K O R L I B R A R Y
  • 15.
    AML Workbench Sample, understand,and prep data rapidly Support for Spark + Python + R (roadmap) Execute jobs locally, on remote VMs, Spark clusters, SQL on-premises Git-backed tracking of code, config, parameters, data, run history
  • 16.
    • Column statistics: Numeric • Histogram • Value Counts • Box Plot • Scatter Plot • Time Series • Map Inspectors
  • 18.
    Spark SQL Server Virtual machines GPUs Containerservices Notebooks Azure Machine Learning Workbench Visual Studio Code Tools for AI Visual Studio Tools for AI PyCharm SQL Server Machine Learning Server O N - P R E M I S E S E D G E C O M P U T I N G Azure IoT Edge Experimentation and Model Management A Z U R E M A C H I N E L E A R N I N G S E R V I C E S T R A I N & D E P LO Y O P T I O N S A Z U R E
  • 19.
    Local machine Scale upto DSVM Scale out with Spark on HDInsight Azure Batch AI (Coming Soon) ML Server (Coming Soon) A ZURE ML EXPERIMENTATION Command line tools IDEs Notebooks in Workbench VS Code Tools for AI VS Tools for AI
  • 20.
    DOCKER Single node deployment (cloud/on-prem) AzureContainer Service Azure IoT Edge Microsoft ML Server Spark clusters SQL Server (Coming Soon) A ZURE ML MODEL MANAGEMENT
  • 21.
  • 22.
    Data Science Virtual Machines (DSVM)DSVM – Windows Server 2016 DSVM – Linux – Ubuntu Deep Learning Virtual Machines
  • 24.
  • 25.
     Researchers tooka traditional machine learning approach • Example: HoG Detectors - Histogram of oriented gradients (HoG) features - Sliding window detector - SVM Classifier - Very fast OpenCV implementation (<100ms)
  • 27.
    Deep Neural Networkfor Computer Vision cat? YES dog? NO car? NO Convolutional Layers Fully Connected Layers Complex Objects & Scenes (people, animals, cars, beach scene, etc.) Image Low-Level Features (lines, edges, color fields, etc.) High-Level Features (corners, contours, simple shapes) Object Parts (wheels, faces, windows, etc.)
  • 31.
    Cognitive Toolkit Unlock deeperlearning A free, easy-to-use, open-source toolkit that trains deep learning algorithms to learn like the human brain. Microsoft Cognitive Toolkit
  • 32.
    Clothing texture dataset: Canwe apply transfer learning to accurately classify clothing texture? LeopardStriped Dotted
  • 33.
    Pre-Built CNN fromGeneral Task on Millions of Images Output Layer Stripped Outputs of penultimate layer of ImageNet Trained CNN provide excellent general purpose image features cat? YES dog? NO car? NO Classi fier e.g. SVM dotted?
  • 34.
    Pre-Built CNN fromGeneral Task on Millions of Images Output Layer Stripped Using a pre-trained DNN, an accurate model can be achieved with thousands (or less) of labeled examples instead of millions cat? YES dog? NO car? NO dotted? Train one or more layers in new network
  • 36.