Learning Objectives:
- Get an introduction to Amazon SageMaker
- Learn how to integrate Amazon SageMaker and other AWS Services within an Enterprise environment
- View a walkthrough of the machine learning lifecycle to cover best practices in the ML process
6. ML @ AWS
OUR MISSION
Put machine learning in the
hands of every developer and
data scientist
7. FRAMEWORKS
KERAS
PLATFORMS
APPLICATION SERVICES
R E K O G N I T I O N R E K O G N I T I O N
V I D E O
P O L L Y T R A N S C R I B E T R A N S L A T E C O M P R E H E N D L E X
Amazon SageMaker Amazon Mechanical Turk
The AWS Machine Learning Stack
8. K E R A S
F R A M E W O R K S A N D I N T E R F A C E S
F r a m e w o r k
s
I n t e r f a c e
s
Complete Control over Frameworks & Infrastructure
For the data scientist, AI researcher, or advanced ML practitioner
9. K E R A S
F R A M E W O R K S A N D I N T E R F A C E S
F r a m e w o r k
s
I n t e r f a c e
s
NVIDIA
Tesla V100 GPUs
(14x faster than P2)
P3
Machine Learning
AMIs
5,120 Tensor cores
128GB of memory
1 Petaflop of compute
NVLink 2.0
I N F R A S T R U C T U R E
Complete Control over Frameworks & Infrastructure
For the data scientist, AI researcher, or advanced ML practitioner
17. BUI LD T R AI N
Where should you spend your time?
Run scalable training
18. BUI LD T R AI N
Where should you spend your time?
Tune Hyperparameters
T UN E
19. BUI LD T R AI N D EPLOY
Where should you spend your time?
T UN E
Deploy and operate
20. FRAMEWORKS
KERAS
PLATFORMS
APPLICATION SERVICES
R E K O G N I T I O N R E K O G N I T I O N
V I D E O
P O L L Y T R A N S C R I B E T R A N S L A T E C O M P R E H E N D L E X
Amazon SageMaker Amazon Mechanical Turk
The AWS Machine Learning Stack
22. Pre-built
notebooks for
common problems
Built-in, high
performance
algorithms
One-click
training
Hyperparameter
optimization
D EPLOY
BUI LD T R AI N & T UN E
Build, train, tune, and host your own models
Amazon SageMaker
23. Fully managed
hosting with auto-
scaling
One-click
deployment
Pre-built
notebooks for
common problems
Built-in, high
performance
algorithms
One-click
training
BUI LD T R AI N & T UN E D EPLOY
Build, train, tune, and host your own models
Hyperparameter
optimization
Amazon SageMaker
24. Fully managed
hosting with auto-
scaling
One-click
deployment
Pre-built
notebooks for
common problems
Built-in, high
performance
algorithms
One-click
training
BUI LD T R AI N & T UN E D EPLOY
Build, train, tune, and host your own models
End-to-end encryption with KMS
End-to-end VPC support
Compliance and audit capabilities
Metadata and experiment management capabilities
Pay as you go
Amazon SageMaker
Hyperparameter
optimization
26. Enabling experimentation speed
Tr a i n w i t h
l o c a l n o t e b o o k s
Train on notebook
instances
PetaFLOP
training on p3.16xl
Go distributed
with one line of code
Same containers
Amazon SageMaker Local Mode
34. Algorithms for
“infinite scale”
Distributed by
default
Train on a
data stream
Checkpoint
for re-training
Single pass
training
Not memory
bound
Rethinking Algorithms Design for infinite scale & streaming data
Amazon SageMaker
36. Rethinking Algorithms Design for large scale & streaming data
AlgorithmData Model
N EW D AT A
PREDICTION
Amazon SageMaker
37. Rethinking Algorithms Design for large scale & streaming data
Convolutional
neural network
Images Model
NE W DATA
OBJECTS, LOCATIONS, FACES
Amazon SageMaker
38. Rethinking Algorithms Design for large scale & streaming data
K-Means
clustering
Inventory Model
NE W DATA
PREDICTED SALES ACTIVITY
Amazon SageMaker
39. Rethinking Algorithms Design for large scale & streaming data
Logistic
Regression
Purchases Model
NE W DATA
FRAUD RISK
Amazon SageMaker
40. K-Means
XGBoost
Linear Learner
PCA
Seq2Seq
DeepAR
Bring your
own algorithm
Factorization Machines
Image Classification
LDA NTM
BlazingText
Algorithms
as-a-service
+ open source
Amazon
SageMaker
Rethinking Algorithms Design for large scale & streaming data
Amazon SageMaker
52. Data Lake Storage
Amazon S3
Security
Access Control
Encryption
VPC
KMS
Auditing
Compliance
Roles
Fine Grained Access Controls
Compute
Powerful GPU & CPU Instances
AWS Lambda
Analytics
Amazon Athena
Amazon EMR
Amazon Redshift & Redshift Spectrum
Broadest & Easiest to Use ML Platform with the
Most Customers
53. ml.aws
Dan R. Mbanga – dmmbanga@amazon.com
Global Lead BD Mgr, Amazon AI Platforms