build machine models with amazon sagemaker...ml frameworks & infrastructure. the amazon ml...
TRANSCRIPT
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Build Machine Models with Amazon SageMaker
Julien SimonGlobal Evangelist, AI & Machine Learning@julsimon
M L F R A M E W O R K S &
I N F R A S T R U C T U R E
The Amazon ML Stack: Broadest & Deepest Set of Capabilities
A I S E R V I C E SR E K O G N I T I O N
I M A G EP 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
C O M P R E H E N DM E D I C A L
L E XR E K O G N I T I O NV I D E O
Vis ion Speech Chatbots
A M A Z O N S A G E M A K E R
B U I L D T R A I N
F O R E C A S TT E X T R A C T P E R S O N A L I Z E
D E P L O YPre-bui l t algorithms & notebooks
Data label ing (G R O U N D T R U T H )
One-cl ick model training & tuning
Optimization ( N E O )
One-cl ick deployment & host ingM L S E R V I C E S
F r a m e w o r k s I n t e r f a c e s I n f r a s t r u c t u r e
E C 2 P 3 & P 3 d n
E C 2 C 5 F P G A s G R E E N G R A S S E L A S T I CI N F E R E N C E
Models without training data (REINFORCEMENT LEARNING)Algorithms & models ( A W S M A R K E T P L A C E )
Language Forecast ing Recommendat ions
NEW NEWNEW
NEW
NEWNEWNEW
NEW
NEW
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Amazon SageMaker: Build, Train, and Deploy ML Models at Scale
Collect and prepare training data
Choose and optimize yourML algorithm
Train andTune ML Models
Set up andmanage
environmentsfor training
Deploy modelsin production
Scale and managethe productionenvironment
123
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Machine learning cycle
Business Problem
ML problem framing Data collection
Data integration
Data preparation and cleaning
Data visualization and analysis
Feature engineering
Model training and parameter tuning
Model evaluation
Monitoring and debugging
Model deployment
Predictions
Are business
goals met?
YESNO
Dat
a au
gmen
tatio
n
Feat
ure
augm
enta
tion
Re-training
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Successful models require high-quality data
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Successful models require high-quality data
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Amazon SageMaker Ground Truthhttps://aws.amazon.com/blogs/aws/amazon-sagemaker-ground-truth-build-highly-accurate-datasets-and-reduce-labeling-costs-by-up-to-70
Easily integrate human labelers
Get accurateresults
K E Y F E A T U R E S
Automatic labeling via machine learning
Ready-made and custom workflows for image bounding box,
segmentation, and text
Labelmanagement
Quickly label training data
Private and public human workforce
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Manage data on AWS
Business Problem
ML problem framing Data collection
Data integration
Data preparation and cleaning
Data visualization and analysis
Feature engineering
Model training and parameter tuning
Model evaluation
Monitoring and debugging
Model deployment
Predictions
Are business
goals met?
YESNO
Dat
a au
gmen
tatio
n
Feat
ure
augm
enta
tion
Re-training
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Build and train models using SageMaker
Business Problem
ML problem framing Data collection
Data integration
Data preparation and cleaning
Data visualization and analysis
Feature engineering
Model training and parameter tuning
Model evaluation
Monitoring and debugging
Model deployment
Predictions
Are business
goals met?
YESNO
Dat
a au
gmen
tatio
n
Feat
ure
augm
enta
tion
Re-training
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Deploy models using SageMaker
Business Problem
ML problem framing Data collection
Data integration
Data preparation and cleaning
Data visualization and analysis
Feature engineering
Model training and parameter tuning
Model evaluation
Monitoring and debugging
Model deployment
Predictions
Are business
goals met?
YESNO
Dat
a au
gmen
tatio
n
Feat
ure
augm
enta
tion
Re-training
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Amazon SageMaker
Collect and prepare training data
Choose and optimize yourML algorithm
Train andTune ML Models
Set up andmanage
environmentsfor training
Deploy modelsin production
Scale and managethe productionenvironment
123
Model compilationElastic inferenceInference pipelines
P3DN, C5NTensorFlow on 256 GPUsResume HPO tuning job
New built-in algorithmsscikit-learn environment
Model marketplaceSearch
Git integrationElastic inference
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
AWS Machine Learning Marketplace
S U M M I T © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
The Amazon SageMaker API• Python SDK orchestrating all Amazon SageMaker activity
• High-level objects for algorithm selection, training, deploying, automatic model tuning, etc.
• Spark SDK (Python & Scala)
• AWS CLI: ‘aws sagemaker’
• AWS SDK: boto3, etc.
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Training code
Factorization MachinesLinear LearnerPrincipal Component AnalysisK-Means ClusteringXGBoostAnd more
Built-in Algorithms Bring Your Own ContainerBring Your Own Script
Model options
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
IMAGE RECOGNITION | for the good of | PRODUCT SEARCH
Sébastien [email protected] IT Projects & CCOE ManagerTarkett
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
→€2.8B Net sales(2018 figures)
→13,000employees
→Present in more than 100countries
→1.3M square meters of flooring sold each day
A worldwide leader in flooring & sports surfaces
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Could we recommend products to each user?
VINYL & LINOLEUM CARPET WOOD & LAMINATE ACCESSORIES & RUBBER
SPORTS SURFACES
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
«Building a bot»
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Why image search?Why does it make sense in the context of flooring?
Field research showed that architects and designers were already leveraging image search engines
Because getting resultsfrom an image search
engine leads to inspiration
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Project Aladdin
Deep Learning
GPU instances
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Why we used Amazon SageMaker
• Go quicker from idea to production
• Distributed training out of the box
• One line of code to deploy models
• Almost the same cost as Amazon EC2 ($300/month)
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Demo: https://professionnels.tarkett.fr
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Next steps
Explore new possibilities opened by image search
• Find the best substitution when a product is out of stock • Understand « in-situ » the real user demand
• Incorporate external factors (seasons, fashion, styles)• Position our products against the competition• Generate new designs
Merci!Sébastien [email protected] IT Projects & CCOE ManagerTarkett
S U M M I T © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Built-in algorithmsorange: supervised, yellow: unsupervised
Linear Learner: regression, classification Image Classification: Deep Learning (ResNet)
Factorization Machines: regression, classification, recommendation
Object Detection (SSD): Deep Learning (VGG or ResNet)
K-Nearest Neighbors: non-parametric regression and classification
Neural Topic Model: topic modeling
XGBoost: regression, classification, rankinghttps://github.com/dmlc/xgboost
Latent Dirichlet Allocation: topic modeling (mostly)
K-Means: clustering Blazing Text: GPU-based Word2Vec, and text classification
Principal Component Analysis: dimensionality reduction
Sequence to Sequence: machine translation, speech to text and more
Random Cut Forest: anomaly detection DeepAR: time-series forecasting (RNN)
Object2Vec: general-purpose embedding IP Insights: usage patterns for IP addresses
Semantic Segmentation: Deep Learning
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Demo: Text Classification with BlazingText
https://github.com/awslabs/amazon-sagemaker-examples/tree/master/introduction_to_amazon_algorithms/blazingtext_text_classification_dbpedia
https://dl.acm.org/citation.cfm?id=3146354
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Demo: Image classification with Caltech-256
https://gitlab.com/juliensimon/dlnotebooks/sagemaker/
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T
Amazon SageMaker
Collect and prepare training data
Choose and optimize yourML algorithm
Train andTune ML Models
Set up andmanage
environmentsfor training
Deploy modelsin production
Scale and managethe productionenvironment
123
Build Train Deploy
© 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.S U M M I T© 2018, Amazon Web Services, Inc. or Its Affiliates. All rights reserved.
Getting startedhttp://aws.amazon.com/free
https://ml.awshttps://aws.amazon.com/sagemaker
https://github.com/aws/sagemaker-python-sdkhttps://github.com/aws/sagemaker-sparkhttps://github.com/awslabs/amazon-sagemaker-exampleshttps://gitlab.com/juliensimon/ent321
https://medium.com/@julsimonhttps://gitlab.com/juliensimon/dlnotebooks
Thank you!
S U M M I T © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Julien SimonGlobal Evangelist, AI and Machine Learning
@julsimonhttps://medium.com/julsimon
S U M M I T © 2019, Amazon Web Services, Inc. or its affiliates. All rights reserved.