Speed up your Machine Learning workflows with build-in algorithms

© 2018, Amazon Web Services, Inc. or Its Affiliates. All rights reserved.© 2018, Amazon Web Services, Inc. or Its Affiliates. All rights reserved.
Julien Simon
Principal AI/ML Evangelist, Amazon Web Services
Speed up your Machine Learning
workflows with built-in algorithms
@julsimon

One-click training for
ML, DL, and custom
algorithms
Easier training with
hyperparameter
optimization
Highly-optimized
machine learning
algorithms
Deployment
without engineering
effort
Fully-managed
hosting at scale
Build
Pre-built notebook
instances
Deploy
Train
Amazon SageMaker

© 2018, Amazon Web Services, Inc. or Its Affiliates. All rights reserved.
Training
code
• Matrix Factorization
• Regression
• Principal Component Analysis
• K-Means Clustering
• Gradient Boosted Trees
• And More!
Amazon provided Algorithms
Bring Your Own Container
Amazon SageMaker: model options
Bring Your Own Script
IM Estimators in
Apache Spark

Streaming datasets, for
cheaper training
Train faster, in a single
pass
Greater reliability on
extremely large
datasets
Choice of several ML
algorithms
Amazon SageMaker: 10x better algorithms

Infinitely scalable algorithms

Streaming
GPU State

Streaming
Data Size
Memory
Data Size
Time/Cost

Distributed
GPU State
GPU State
GPU State

Shared State
GPU
GPU
GPU Local
State
Shared
State
Local
State
Local
State

Cost vs. Time
$$$$
$$$
$$
$
Minutes Hours Days Weeks Months
Best Alternative
Amazon SageMaker

Linear Learner
Regression (mean squared error)
SageMaker Other
1.02 1.06
1.09 1.02
0.332 0.183
0.086 0.129
83.3 84.5
Classification (F1 Score)
SageMaker Other
0.980 0.981
0.870 0.930
0.997 0.997
0.978 0.964
0.914 0.859
0.470 0.472
0.903 0.908
0.508 0.508
30 GB datasets for web-spam and web-url classification
0
0.2
0.4
0.6
0.8
1
1.2
0 5 10 15 20 25 30
CostinDollars
Billable time in Minutes
sagemaker-url sagemaker-spam other-url other-spam

Factorization Machines
Log_loss F1 Score Seconds
SageMaker 0.494 0.277 820
Other (10 Iter) 0.516 0.190 650
Other (20 Iter) 0.507 0.254 1300
Other (50 Iter) 0.481 0.313 3250
Click Prediction 1 TB advertising dataset,
m4.4xlarge machines, perfect scaling.
$-
$20.00
$40.00
$60.00
$80.00
$100.00
$120.00
$140.00
$160.00
$180.00
$200.00
1 2 3 4 5 6 7 8CostinDollars
Billable Time in Hours
10
machines
20
machines
30
machines
4050

Demo: building a movie recommender with
Factorization Machines
h t t p s : / / m e d i u m . c o m / @ j u l s i m o n / b u i l d i n g - a - m o v i e - r e c o m m e n d e r - w i t h - f a c t o r i z a t i o n -
m a c h i n e s - o n - a m a z o n - s a g e m a k e r - c e d b f c 8 c 9 3 d 8

0
1
2
3
4
5
6
7
8
10 100 500
BillableTimeinMinutes Number of Clusters
sagemaker other
K-Means Clustering
k SageMaker Other
Text
1.2GB
10 1.18E3 1.18E3
100 1.00E3 9.77E2
500 9.18.E2 9.03E2
Images
9GB
10 3.29E2 3.28E2
100 2.72E2 2.71E2
500 2.17E2 Failed
Videos
27GB
10 2.19E2 2.18E2
100 2.03E2 2.02E2
500 1.86E2 1.85E2
Advertising
127GB
10 1.72E7 Failed
100 1.30E7 Failed
500 1.03E7 Failed
Synthetic
1100GB
10 3.81E7 Failed
100 3.51E7 Failed
500 2.81E7 Failed
Running Time vs. Number of Clusters
~10x Faster!

Principal Component Analysis (PCA)
More than 10x faster
at a fraction the cost!
0.00
20.00
40.00
60.00
80.00
100.00
120.00
8 10 20
Mb/Sec/Machine
Number of Machines
other sagemaker-deterministic sagemaker-randomized
Cost vs. Time Throughput and Scalability
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
0 10 20 30 40 50
CostinDollars
Billable time in Minutes
other sagemaker-deterministic sagemaker-randomized

Neural Topic Modeling
Perplexity vs. Number of Topic
Encoder: feedforward net
Input term counts vector
Document
Posterior
Sampled Document
Representation
Decoder:
Softmax
Output term counts vector
0
2000
4000
6000
8000
10000
12000
0 50 100 150 200
Perplexity
Number of Topics
NTM Other

DeepAR: Time Series Forecasting
Mean absolute
percentage error
P90 Loss
DeepAR R DeepAR R
traffic
Hourly occupancy rate of 963
Bay Area freeways
0.14 0.27 0.13 0.24
electricity
Electricity use of 370 homes
over time
0.07 0.11 0.08 0.09
pageviews
Page view hits of
websites
10k 0.32 0.32 0.44 0.31
180k 0.32 0.34 0.29 NA
One hour on p2.xlarge, $1
Input
Network

DeepAR
https://arxiv.org/abs/1704.04110

Demo: predicting world temperature
with DeepAR
h t t p s : / / m e d i u m . c o m / @ j u l s i m o n / p r e d i c t i n g - w o r l d - t e m p e r a t u r e - w i t h - t i m e - s e r i e s -
a n d - d e e p a r - o n - a m a z o n - s a g e m a k e r - e 3 7 1 c f 9 4 d d b 5

More built-in algorithms

Spectral LDA
Training Time vs. Number of Topics
0
50
100
150
200
250
0 20 40 60 80 100TrainingTimeinMinutes
Number of Topics
lda-data-a lda-data-b other-data-a other-data-b

Boosted Decision Trees
Throughput vs. Number of Machines
XGBoost is one of the most
commonly used classifiers.
0
200
400
600
800
1000
1200
1400
0 10 20 30 40 50 60 70
ThroughputinMB/Sec
Number of Machines (C4.8xLarge)
https://github.com/dmlc/xgboost

Sequence to Sequence
English-German Translation
0
5
10
15
20
25
0 5 10 15 20 25
BLEUScore
Billable Time in Hours
P2.16x P2.8x P2.x
Best known result!
• Based on Sockeye
and Apache MXNet.
• Multi-GPU.
• Can be used for Neural
Machine Translation.
• Supports both RNN/CNN as
encoder/decoder

https://arxiv.org/abs/1712.05690
https://github.com/awslabs/sockeye
Sockeye

Image Classification
• ResNet implementation
with Apache MXNet.
• More networks to come.
• Transfer learning: begin
with a model already
trained on ImageNet!
0
0.5
1
1.5
2
2.5
3
3.5
0 1 2 3 4 5
Speedup
Number of Machines (P2)
Linear Speedup with Horizontal Scaling

Demo: fine-tuning an image classification
model
h t t p s : / / m e d i u m . c o m / @ j u l s i m o n / i m a g e - c l a s s i f i c a t i o n - o n - a m a z o n - s a g e m a k e r -
9 b 6 6 1 9 3 c 8 b 5 4

Latest addition: Blazing Text
https://dl.acm.org/citation.cfm?id=3146354

Resources
https://aws.amazon.com/machine-learning
https://aws.amazon.com/blogs/ai
https://aws.amazon.com/sagemaker (free tier available)
https://github.com/awslabs/amazon-sagemaker-examples
An overview of Amazon SageMaker https://www.youtube.com/watch?v=ym7NEYEx9x4
https://medium.com/@julsimon

Thank you!
Julien Simon
Principal AI/ML Evangelist, Amazon Web Services
@julsimon

Speed up your Machine Learning workflows with build-in algorithms

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Similar to Speed up your Machine Learning workflows with build-in algorithms

Similar to Speed up your Machine Learning workflows with build-in algorithms (20)

More from Julien SIMON

More from Julien SIMON (20)

Recently uploaded

Recently uploaded (20)

Speed up your Machine Learning workflows with build-in algorithms