Slides Rethink Machine Learning For Regulated Industries

22 February 2023
© 2023, Amazon Web Services, Inc. or its affiliates. All rights reserved.
Rethinking machine learning for
regulated industries
Juan Carlos Bedoya

Solutions Architect
Amazon Web Services
Agenda
- A modern machine learning platform
- Maturity model
- Making datasets available across the organization
- Automated governance and compliance
- MLOps
- Federated ML
Modern Machine Learning Platforms
Increased Anticipate Improved customer Learning from

Better decisions
efficiency events satisfaction others

Better decisions
Transparency Inclusion Accountability Security Scale

Better decisions
Data

Better decisions
Platform
Bias Federated
Guardrails Automation Explainability Auditing
detection learning
Data
MLOps Maturity Model
MLOps Requirements
• ML pipelines
• Data warehouses
• Data pipelines
• CI/CD
Tech • Model development environment
• Model serving infrastructure
• Training
• Deployment
• Monitoring
Processes • Logging
• Performance
• Governance
• Roles required (Data

scientist, ML engineer,
SysOps, etc)
People • Capability
• Collaboration
• Culture
MLOps Maturity Model
Making Datasets and Features
Available Across The Organization
Data Lake
Data Lake
Data Engineers
& Owners
Data Lake
Data Lake
AWS Lake Formation
Data Engineers
& Owners
Data Lake
Data Lake
AWS Lake Formation
Data Engineers
& Owners
Amazon S3 AWS Glue
Data Lake
Data Lake
AWS Lake Formation
Data Engineers
& Owners
Amazon SageMaker Amazon S3 AWS Glue

Feature Store
Amazon SageMaker Feature Store
Data Amazon
SageMaker Feature Models
sources Amazon
Pipelines
SageMaker
Vehicle damage Feature Store
images
Vehicle total loss
Claims,
customers, ...
Text Customer churn

interactions
Telemetry
Next best action

Helps with auditability and
traceability requirements to
manage risk and compliance
Automated Governance
and Compliance
AWS Service Catalogue
AWS Shared Services account AWS Dev account
Data Engineers/
Infrastructure
Engineers
AWS Service Catalogue
AWS Shared Services account AWS Dev account
Data Engineers/
Infrastructure
Engineers • A studio environment
• Studio user profiles
• Model deployment pipelines
• Training pipelines
• Inference pipelines
• A system for monitoring
• and alerting
MLOps Tools and
Resources
Amazon SageMaker Model Registry
SageMaker
Model Registry
Model 1 v1 Model 1 v2 Model 1 v3
Model 2 v1 Model 2 v2
Configuration
management & model
metadata for models
Model3 v1 Model 3 v2 Model 3 v3
Amazon SageMaker Pipelines
Raw Data
SageMaker Pipelines 1
Data Processing
Data Automated
model build
Pipeline Train workflows Training
triggers Data
Evaluate
Register
Model development phase Model Training
Source &
version
control Data Automated
model Model
Train re-training
workflows
Evaluate
Register Model Evaluation
Operational phase
Model build
Accuracy Yes
Register Model
> 90%
Amazon SageMaker Projects
Build, train, deploy
Build, train
deploy
SageMaker Pipelines SageMaker 3

1 2
Model Registry
Data Automated
Pipeline
model build
Train workflows
triggers
Evaluate Model
Register deployment
Approval pipeline
Source Model development phase
workflows
&
version Data Automated
control model Configuration
Train re-training management & model Operational phase
workflows metadata for models
Evaluate
Register
Operational phase
Model build Model deploy and manage
Federated Learning
What is Federated Learning (FL)?
FL – a distributed ML approach that trains ML models on distributed datasets
Goal – improve the accuracy of ML models while preserving privacy
On premises
Dataset Local training Local model
FL server
Edge Aggregate local models
into a new global model
Dataset Local training Local model
Global model
Cloud
Dataset Local training Local model FL exchanges ML models between the

central FL server and clients without
FL clients
moving training data to a central repository
Federated Learning (FL)
1. Server starts with some global
model
2. Server distributes global model to
each endpoint
3. Each endpoint incrementally trains
the global model based on their
local training data
4. Endpoints upload locally trained
models
5. Server aggregates the local models
to produce a new global model
6. Repeat until some termination
condition is met
7. Distribute the final model to all
parties
We are here to help you transform your ML platform
AWS SAs, AWS Professional Services, Amazon SageMaker workshop

AWS ML Labs, AWS Prototyping Get hands-on with self-paced
Skills and experience from AWS experts learning workshops
to supplement your team
AWS Support AWS Community – repost.aws

Go beyond reactive break-fix with Learn from other users and get quick
proactive support programs responses to ‘how-to’ questions
AWS Training and Certification AWS Partners

Advance your team’s skills and build your fluency A global network of partners with
by learning from AWS experts deep AWS expertise
*Subject to qualification criteria

AWS Training and Certification
Access the AI & ML learning plan courses built by AWS experts on AWS Skill Builder
• Get started with digital self-paced, on-demand training and

ramp-up guides to help you grow your technical skills
• Learn how to apply machine learning, artificial intelligence, and

deep learning to unlock new insights and value in your role
• Take the steps today, towards validating your expertise with an

AWS Certified Machine Learning – Specialty Certification https://bit.ly/3FnxDH7
Learn your way explore.skillbuilder.aws »
Thank you for attending AWS Innovate – Data & AI/ML Edition
We hope you found it interesting! A kind reminder to complete the survey.
Let us know what you thought of today’s event and how we can improve the event
experience for you in the future.
aws-apj-marketing@amazon.com
twitter.com/AWSCloud
facebook.com/AmazonWebServices
youtube.com/user/AmazonWebServices
slideshare.net/AmazonWebServices
twitch.tv/aws
Thank you!

Slides Rethink Machine Learning For Regulated Industries

Uploaded by

Copyright:

Available Formats

You might also like

Slides Rethink Machine Learning For Regulated Industries

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Slides Rethink Machine Learning For Regulated Industries

Uploaded by

Copyright:

Available Formats

22 February 2023

Juan Carlos Bedoya

- Making datasets available across the organization

- Automated governance and compliance

Increased Anticipate Improved customer Learning from

Increased Anticipate Improved customer Learning from

Transparency Inclusion Accountability Security Scale

Increased Anticipate Improved customer Learning from

Transparency Inclusion Accountability Security Scale

Increased Anticipate Improved customer Learning from

Transparency Inclusion Accountability Security Scale

• Roles required (Data

AWS Lake Formation

AWS Lake Formation

Amazon S3 AWS Glue

AWS Lake Formation

Amazon SageMaker Amazon S3 AWS Glue

Text Customer churn

Next best action

AWS Shared Services account AWS Dev account

AWS Shared Services account AWS Dev account

Model 1 v1 Model 1 v2 Model 1 v3

Model3 v1 Model 3 v2 Model 3 v3

SageMaker Pipelines SageMaker 3

Dataset Local training Local model

Dataset Local training Local model FL exchanges ML models between the

AWS SAs, AWS Professional Services, Amazon SageMaker workshop

AWS Support AWS Community – repost.aws

AWS Training and Certification AWS Partners

*Subject to qualification criteria

• Get started with digital self-paced, on-demand training and

• Learn how to apply machine learning, artificial intelligence, and

• Take the steps today, towards validating your expertise with an

Learn your way explore.skillbuilder.aws »

You might also like