© 2017, Amazon Web Services, Inc. or its Affiliates. All rights reserved.

Kumar Venkateswar, Product Manager, Amazon AI

6/12/2017

Amazon Machine Learning

Agenda

Smart applications by example

Developing with Amazon ML

Demo

How Amazon ML fits into other AWS AI services

Q&A

Machine learning and smart applications

Machine learning is the technology that

automatically finds patterns in your data and

uses them to make predictions for new data

points as they become available

Your data + machine learning = smart applications

Challenge: Insurance companies spend billions of

dollars on roof losses annually

Solution: BuildFax provides roof age and condition

estimates with the help of Amazon ML.

BuildFax

Amazon Machine Learning democratizes the process of building predictive models. It's easy and fast to

use, and has machine-learning best practices encapsulated in the product, which lets us deliver results

significantly faster than in the past.

Joe Emison

Founder & Chief Technology Officer

Challenge: Restaurant owners want to use data to

better serve their customers

Solution: Upserve provides predictions on the

number of customers and items ordered with the

help of Amazon ML.

Upserve

Using Amazon Machine Learning, we can predict the total number of customers who will walk through a

restaurant’s doors in a night. As a result, restaurateurs can better prep and plan their staffing for that night.

Bright Fulton

Director of Infrastructure Engineering

And a few more examples…

Fraud detection

(binary)

Detecting fraudulent transactions, filtering spam emails,

flagging suspicious reviews, …

Personalization

(categorical)

Recommending content, predictive content loading,

improving user experience, …

Targeted marketing

(binary/categorical)

Matching customers and offers, choosing marketing

campaigns, cross-selling and up-selling, …

Content classification

(categorical)

Categorizing documents, matching hiring managers and

resumes, …

Churn prediction

(binary)

Finding customers who are likely to stop using the

service, free-tier upgrade targeting, …

Customer support

(categorical)

Predictive routing of customer emails, social media

listening, …

But why not use business rules instead?

Smart applications by counter-example

SELECT c.ID

FROM customers c

LEFT JOIN orders o

ON c.ID = o.customer

LEFT JOIN products p

ON p.ID = o.product

GROUP BY c.ID

HAVING o.category = ‘toys’

AND ((p.description LIKE ‘%copter%’

AND o.date > GETDATE() - )

OR (COUNT(*) > 2

AND SUM(o.price) > 150

AND o.date > GETDATE() – 40)

)

Use machine learning

technology to learn

your business rules

from data, instead of

endlessly iterating

business rules

Trainmodel

Evaluate andoptimize

Retrieve predictions

1 2 3

Building smart applications with Amazon ML

Trainmodel

Evaluate andoptimize

Retrieve predictions

1 2 3

Building smart applications with Amazon ML

- Create a Datasource object pointing to your data

- Explore and understand your data

- Transform data and train your model

Create a Datasource object

>>> import boto3

>>> ml = boto3.client('machinelearning')

>>> ds = ml.create_data_source_from_s3(

DataSourceId = 'my_datasource',

DataSpec= {

'DataLocationS3':'s3://bucket/input/',

'DataSchemaLocationS3':'s3://bucket/input/.schema'},

ComputeStatistics = False)

Explore and understand your data

Train your model

>>> import boto3

>>> ml = boto3.client('machinelearning')

>>> model = ml.create_ml_model(

MLModelId=’my_model',

MLModelType='REGRESSION',

TrainingDataSourceId='my_datasource')

Trainmodel

Evaluate andoptimize

Retrieve predictions

1 2 3

Building smart applications with Amazon ML

- Understand model quality

- Adjust model interpretation

Explore model quality

Fine-tune model interpretation

Fine-tune model interpretation

Trainmodel

Evaluate andoptimize

Retrieve predictions

1 2 3

Building smart applications with Amazon ML

- Batch predictions

- Real-time predictions

Batch predictions

Asynchronous, large-volume prediction generation

Request through service console or API

Best for applications that deal with batches of data records

>>> import boto3

>>> ml = boto3.client('machinelearning')

>>> model = ml.create_batch_prediction(

BatchPredictionId = 'my_batch_prediction’

BatchPredictionDataSourceId = ’my_datasource’

MLModelId = ’my_model',

OutputUri = 's3://examplebucket/output/’)

Real-time predictions

Synchronous, low-latency, high-throughput prediction generation

Request through service API or server or mobile SDKs

Best for interaction applications that deal with individual data

records

>>> import boto3

>>> ml = boto3.client('machinelearning')

>>> ml.predict(

MLModelId=’my_model',

PredictEndpoint=’example_endpoint’,

Record={’key1':’value1’, ’key2':’value2’})

{

'Prediction': {

'predictedValue': 13.284348,

'details': {

'Algorithm': 'SGD',

'PredictiveModelType': 'REGRESSION’

}

}

}

© 2016, Amazon Web Services, Inc. or its Affiliates. All rights reserved

Demo

We are here to make machine learning widespread…

by sharing what we use ourselves!

Developer APIs: Rekognition, Polly, Lex, Amazon Kinesis Analytics

Developer platform components: Amazon Machine Learning

Data Scientist platform components: Apache Spark MLlib,

Apache MXNet, Deep Learning AMI

Lower level platform components: Elastic MapReduce, EC2

(CPU, GPU, and FPGA instances)

More to come, of course!

Amazon Confidential

© 2016, Amazon Web Services, Inc. or its Affiliates. All rights reserved

Q&A

Amazon Confidential

What else can we do?