A holistic and real-world approach to operationalizing artificial intelligence in your company
In Operating AI, Director of Technology and Architecture at Ericsson AB, Ulrika Jägare, delivers an eye-opening new discussion of how to introduce your organization to artificial intelligence by balancing data engineering, model development, and AI operations. You'll learn the importance of embracing an AI operational mindset to successfully operate AI and lead AI initiatives through the entire lifecycle, including key areas such as; data mesh, data fabric, aspects of security, data privacy, data rights and IPR related to data and AI models.
In the book, you’ll also discover:
- How to reduce the risk of entering bias in our artificial intelligence solutions and how to approach explainable AI (XAI)
- The importance of efficient and reproduceable data pipelines, including how to manage your company's data
- An operational perspective on the development of AI models using the MLOps (Machine Learning Operations) approach, including how to deploy, run and monitor models and ML pipelines in production using CI/CD/CT techniques, that generates value in the real world
- Key competences and toolsets in AI development, deployment and operations
- What to consider when operating different types of AI business models
With a strong emphasis on deployment and operations of trustworthy and reliable AI solutions that operate well in the real world―and not just the lab―Operating AI is a must-read for business leaders looking for ways to operationalize an AI business model that actually makes money, from the concept phase to running in a live production environment.
Author(s): Ulrika Jägare
Publisher: Wiley
Year: 2022
Language: English
Pages: 273
City: Hoboken
Cover
Title Page
Copyright Page
About the Author
About the Technical Editor
Acknowledgments
Contents at a Glance
Contents
Foreword
Introduction
What Does This Book Cover?
How to Contact the Publisher
How to Contact the Author
Chapter 1 Balancing the AI Investment
Defining AI and Related Concepts
Operational Readiness and Why It Matters
Applying an Operational Mind-set from the Start
The Operational Challenge
Strategy, People, and Technology Considerations
Strategic Success Factors in Operating AI
People and Mind-sets
The Technology Perspective
Chapter 2 Data Engineering Focused on AI
Know Your Data
Know the Data Structure
Know the Data Records
Know the Business Data Oddities
Know the Data Origin
Know the Data Collection Scope
The Data Pipeline
Types of Data Pipeline Solutions
Data Quality in Data Pipelines
The Data Quality Approach in AI/ML
Scaling Data for AI
Key Capabilities for Scaling Data
Introducing a Data Mesh
When You Have No Data
The Role of a Data Fabric
Why a Data Fabric Matters in AI/ML
Key Competences and Skillsets in Data Engineering
Chapter 3 Embracing MLOps
MLOps as a Concept
From ML Models to ML Pipelines
The ML Pipeline
Adopt a Continuous Learning Approach
The Maturity of Your AI/ML Capability
Level 0—Model Focus and No MLOps
Level 1—Pipelines Rather than Models
Level 2—Leveraging Continuous Learning
The Model Training Environment
Enabling ML Experimentation
Using a Simulator for Model Training
Environmental Impact of Training AI Models
Considering the AI/ML Functional Technology Stack
Key Competences and Toolsets in MLOps
Clarifying Similarities and Differences
MLOps Toolsets
Chapter 4 Deployment with AI Operations in Mind
Model Serving in Practice
Feature Stores
Deploying, Serving, and Inferencing Models at Scale
The ML Inference Pipeline
Model Serving Architecture Components
Considerations Regarding Toolsets for Model Serving
The Industrialization of AI
The Importance of a Cultural Shift
Chapter 5 Operating AI Is Different from Operating Software
Model Monitoring
Ensuring Efficient ML Model Monitoring
Model Scoring in Production
Retraining in Production Using Continuous Training
Data Aspects Related to Model Retraining
Understanding Different Retraining Techniques
Deployment after Retraining
Disadvantages of Retraining Models Frequently
Diagnosing and Managing Model Performance Issues in Operations
Issues with Data Processing
Issues with Data Schema Change
Data Loss at the Source
Models Are Broken Upstream
Monitoring Data Quality and Integrity
Monitoring the Model Calls
Monitoring the Data Schema
Detecting Any Missing Data
Validating the Feature Values
Monitor the Feature Processing
Model Monitoring for Stakeholders
Ensuring Stakeholder Collaboration for Model Success
Toolsets for Model Monitoring in Production
Chapter 6 AI Is All About Trust
Anonymizing Data
Data Anonymization Techniques
Pros and Cons of Data Anonymization
Explainable AI
Complex AI Models Are Harder to Understand
What Is Interpretability?
The Need for Interpretability in Different Phases
Reducing Bias in Practice
Rights to the Data and AI Models
Data Ownership
Who Owns What in a Trained AI Model?
Balancing the IP Approach for AI Models
The Role of AI Model Training
Addressing IP Ownership in AI Results
Legal Aspects of AI Techniques
Operational Governance of Data and AI
Chapter 7 Achieving Business Value from AI
The Challenge of Leveraging Value from AI
Productivity
Reliability
Risk
People
Top Management and AI Business Realization
Measuring AI Business Value
Measuring AI Value in Nonrevenue Terms
Operating Different AI Business Models
Operating Artificial Intelligence as a Service
Operating Embedded AI Solutions
Operating a Hybrid AI Business Model
Index
EULA
