The AI Engineer's Guide to Surviving the EU AI Act Larysa Visengeriyeva

Dostępność produktu

Produkt nie został jeszcze oceniony pod kątem ułatwień dostępu lub nie podano żadnych informacji o ułatwieniach dostępu lub są one niewystarczające. Prawdopodobnie Wydawca/Dostawca jeszcze nie umożliwił dokonania walidacji produktu lub nie przekazał odpowiednich informacji na temat jego dostępności.

Spis treści książki

• Preface
  • Who Should Read This Book
  • Navigating This Book
  • Conventions Used in This Book
  • OReilly Online Learning
  • How to Contact Us
  • Acknowledgments
• 1. Understanding the AI Regulations
  • The Motivation for the EU AI Act: Trustworthy AI
    • Human Agency and Oversight
    • Technical Robustness and Safety
    • Privacy and Data Governance
    • Transparency
    • Diversity, Non-Discrimination, and Fairness
    • Societal and Environmental Well-Being
    • Accountability
  • The EU AI Act in a Nutshell
    • Definitions
    • Key Players from Creation to Market Operation
    • Classification of AI Systems by Risk Levels
    • Enforcement and Implementation
    • The Full Picture of Compliance
    • Penalties for EU AI Act Violation
  • Existing AI Regulations and Standards
  • Conclusion
  • Further Reading
• 2. AI Engineering: A Proactive Compliance Catalyst
  • Structuring the AI System Development Process with CRISP-ML(Q)
    • Risk Mitigation, Quality Assurance, and Alignment with the EU AI Act
    • The Six Phases of CRISP-ML(Q)
      • Business and data understanding
      • Data preparation
      • Model engineering
      • Model evaluation
      • Model deployment
      • Model monitoring and maintenance
  • Understanding MLOps
  • Defining Technical Components with the MLOps Stack Canvas
    • Value Proposition
    • Data and Code Management
      • Data Sources and Data Versioning
      • Data Analysis and Experiment Management
      • Feature Store and Workflows
      • Foundations (Reflecting DevOps)
    • Model Management
      • CI/CT/CD: ML Pipeline Orchestration
      • Model Registry and Model Versioning
      • Model Deployment
      • Prediction Serving
      • Model, Data, and Application Monitoring
    • Metadata Management
    • Additional Considerations of MLOps
  • CRISP-ML(Q) and MLOps
  • Conclusion
• 3. Data and AI Governance and AI Engineering
  • The Importance of Data and AI Governance in the EU AI Act Era
  • Overview of Data Governance
    • Data Governance Defined
    • The Data Engineering Lifecycle and Data Governance
      • Data generation
      • Ingestion
      • Transformation
      • Storage
      • Serving
    • Integrating Data Governance into MLOps
      • Value Proposition
      • Data Sources and Data Versioning
      • Data Analysis and Experiment Management
      • Feature Store and Workflows
      • Foundations (Reflecting DevOps)
      • CI/CT/CD: ML Pipeline Orchestration
      • Model Registry and Model Versioning
      • Model Deployment
      • Prediction Serving
      • Model, Data, and Application Monitoring
      • Metadata Management
  • Overview of AI Governance
    • AI Governance Defined
    • The AI System Lifecycle and AI Governance
  • Emerging Trends in Data and AI Governance
  • Conclusion
• 4. AI System Assessment and Tailoring AI Engineering for Different Risk Levels
  • AI Compliance, Governance, and Risk Management
  • Creating an AI System Inventory
  • Applicability of the EU AI Act
  • Unacceptable RiskProhibited AI Practices
    • Prohibited AI Use Cases
    • Deep Dive: Social Scoring
  • Determining EU AI Act Obligations
    • Framework for Classification of AI Systems by Risk Levels
      • High risk
        • EU AI Act high-risk system questionnaire
        • Interpretation guide
      • Limited risk (transparency obligations)
        • EU AI Act limited-risk system questionnaire
        • Interpretation guide
      • Low risk
    • Deployer or Provider
  • Integrating EU AI Act Engineering Throughout the AI Development Lifecycle
  • Emerging Roles in Organizations for EU AI Act Compliance
  • Conclusion
• 5. AI Engineering for High-Risk AI Systems
  • AI Engineering for the EU AI Act
    • Goals
    • Alignment with CRISP-ML(Q) Phases
  • AI Engineering Practices for Achieving Compliance
    • Article 9: Risk Management System
      • 1. Business and data understanding
        • Failure Mode and Effects Analysis
        • Testability and value alignment
      • 2. Data preparation
      • 3. Modeling
      • 4. Evaluation
        • Risk identification
        • Testability
        • Value alignment
      • 5. Deployment
        • Risk identification
        • Testability
        • Value alignment
      • 6. Monitoring and maintenance
      • Summary
      • Checklist for compliance
      • Further reading
    • Article 10: Data and Data Governance
      • Key tools
      • Checklist for compliance
    • Article 11: Technical Documentation and Article 12: Record-Keeping
      • Technical documentation requirements
      • Managing documentation debt
      • Existing frameworks for documenting data and AI systems
      • Article 12: Recordkeeping requirements for high-risk AI systems
      • Interdependence of Articles 11 and 12
      • Data and AI system metadata
      • Checklist for compliance
      • Further reading
    • Article 13: Transparency and Provision of Information to Deployers and Article 14: Human Oversight
      • Article 13 quality attributes
      • Article 14 quality attributes
      • Checklist for compliance
    • Article 15: Accuracy, Robustness, and Cybersecurity
      • Checklist for compliance
      • Further information
  • Conclusion
• 6. AI Engineering for Limited-Risk AI Systems
  • Compliance Assessment Versus Transparency Obligation
  • Understanding Transparency Obligations
  • Aligning AI Engineering with SMACTR and CRISP-ML(Q) for Transparency
    • The Five Stages of the SMACTR Framework
    • How the SMACTR Framework Aligns with the EU AI Act
    • Business and Data Understanding Phase
      • CRISP-ML(Q) activities
      • SMACTR integration
      • Key artifacts
    • Data Preparation Phase
      • CRISP-ML(Q) activities
      • SMACTR integration
      • Key artifacts
    • Modeling Phase
      • CRISP-ML(Q) activities
      • SMACTR integration
      • Key artifacts
      • Technical implementation guide
    • Evaluation Phase
      • CRISP-ML(Q) activities
      • SMACTR integration
      • Key artifacts
      • Available tools and technologies
    • Deployment Phase
      • CRISP-ML(Q) activities
      • SMACTR integration
      • Key artifacts
      • Available tools and technologies
    • Monitoring and Maintenance Phase
      • CRISP-ML(Q) activities
      • SMACTR integration
      • Key artifacts
      • Available tools and technologies
  • Technology Trend: AI Governance Platforms
    • Leading Platforms
    • Further Reading
  • Conclusion
• 7. Toward Trustworthy General-Purpose AI and Generative AI
  • The EU AI Act and Generative AI
    • GPAI Systems and Transparency Obligations
    • Regulating General-Purpose AI
      • Definition and scope
      • Systemic risk criteria
      • Obligations for providers of GPAI models
      • Additional obligations for providers of GPAI models with systemic risk 
      • Obligations for deployers of GPAI models
  • Predictive ML Versus GPAI
  • GenAIOpsOperationalizing EU AI Act Compliance for GPAI
    • Key Components
    • How GenAIOps Extends MLOps
    • GenAIOps Tools, Workflows, and Frameworks to Support Transparency
  • Aligning AI Engineering with SMACTR and CRISP-ML(Q) for Transparency
    • Business and Data Understanding Phase
      • Implementation
      • SMACTR integration
    • Data Preparation Phase
      • Implementation
      • SMACTR integration
    • Modeling Phase
      • Implementation
      • SMACTR integration
    • Evaluation Phase
      • Implementation
      • SMACTR integration
    • Deployment Phase
      • Implementation
      • SMACTR integration
    • Model Monitoring and Maintenance Phase
      • Implementation
      • SMACTR integration
  • Conclusion
  • Final Words and Future of AI Policymaking
• A. Designing AI-Powered Applications
  • Backward Thinking for Designing AI Systems
  • The Machine Learning Canvas
    • Value Proposition
    • Monitoring
    • The Prediction Phase
      • Prediction Task
      • Decisions
      • Making Predictions
      • Impact Simulation
    • The Training Phase
      • Data Sources
      • Data Collection
      • Building Models
      • Features
• B. Data Products and Data Contracts
  • Data Products
  • Data Contracts
• C. The Integration of AI Governance and MLOps
  • Value Proposition
  • Data Sources and Data Versioning
  • Data Analysis and Experiment Management
  • Feature Store and Workflows
  • CI/CT/CD: ML Pipeline Orchestration
  • Model Registry and Model Versioning
  • Model Deployment
  • Prediction Serving
  • Model, Data, and Application Monitoring
  • Metadata Management
  • Integrating Data and AI Governance into MLOps
• D. Emerging Roles in Organizations for EU AI Act Compliance
  • Emerging Roles
  • EU AI Compliance for ML Teams
    • Team Topologies for ML Teams
    • Aligning Team Topologies with Ethics, Compliance, and Governance Roles
      • Enabling team: AI ethics and governance
      • Complicated subsystem team: Data and AI security and auditing
• Index

pokaż cały spis treści

ukryj recenzje