Details

Subject(s)

• Public administration—Technological innovations [Browse]
• Artificial intelligence [Browse]

Related name

• Godavarthi, Srinath [Browse]
• Ding, Sherry [Browse]

Summary note

This is your guide book to Generative AI (GenAI) and its application in addressing real-world challenges within the public sector. The book addresses a range of topics from GenAI concepts and strategy to public sector use cases, architecture patterns, and implementation best practices. With a general background in technology and the public sector, you will be able to understand the concepts in this book. The book will help you develop a deeper understanding of GenAI and learn how GenAI differs from traditional AI. You will explore best practices such as prompt engineering, and fine-tuning, and architectural patterns such as Retrieval Augmented Generation (RAG). And you will discover specific nuances, considerations, and strategies for implementation in a public sector organization. You will understand how to apply these concepts in a public sector setting and address industry-specific challenges and problems by studying a variety of use cases included in the book in the areas of content generation, chatbots, summarization, and program management.

Source of description

Description based on publisher supplied metadata and other sources.

Contents

• Intro
• Table of Contents
• About the Authors
• About the Technical Reviewer
• Acknowledgments
• Preface
• Foreword
• Chapter 1: Introduction to Generative AI
• 1.1 What Is AI?
• 1.2 AI and Estimating Home Prices
• 1.2.1 Machine Learning
• 1.2.2 How Machine Learning Algorithms Work
• 1.2.3 Artificial Neural Networks and Deep Learning
• 1.3 Different Types of Learning Tasks
• 1.3.1 Supervised Learning
• 1.3.2 Unsupervised Learning
• 1.3.3 Reinforcement Learning
• 1.3.4 Self-Supervised Learning
• 1.4 What Is GenAI and Where Does It Fit?
• 1.5 Examples of Foundation Models
• 1.6 How an LLM Works
• 1.6.1 Pre-training
• 1.6.2 Inference
• 1.7 How Image Generation FMs Work
• 1.7.1 Pre-training
• 1.7.2 Inference
• 1.8 Key Takeaways
• 1.9 Conclusion
• Chapter 2: GenAI and the Public Sector
• 2.1 Characteristics of a PSO
• 2.1.1 Mission
• 2.1.2 Budget Size
• 2.1.3 Type and Number of Customers Served
• 2.1.4 Employees
• 2.2 Public Sector Challenges: A Closer Look
• 2.2.1 Customer and Employee Experience
• 2.2.2 Data-Driven Decision-Making and Operational Costs
• 2.2.3 IT Systems and Technical Debt
• 2.2.4 Cybersecurity Concerns
• 2.2.5 Other Challenges
• 2.3 How Can GenAI Help PSOs?
• 2.3.1 Content Generation
• 2.3.2 Conversational Agents or Chatbots
• 2.3.3 Content Summarization
• 2.3.4 Business Intelligence, Analytics, and Reporting
• 2.4 Conclusion
• Chapter 3: GenAI Strategy: A Blueprint for  Successful Adoption
• 3.1 GenAI Strategy and Blueprint
• 3.1.1 Align GenAI with Mission Objectives
• 3.1.2 Establish PSO-Wide Policies, Acquisition, and Operating Guidelines
• 3.1.3 Establish AI/GenAI Center of Excellence (ACOE)
• 3.1.4 Identify and Prioritize GenAI Use Cases
• 3.1.5 Establish Tactical Road Map for Production Rollout and Operations.
• 3.1.6 Ensure Compliance with Federal, State, and Local GenAI and AI Guidance and Regulations
• 3.2 GenAI Implementation
• 3.2.1 Business Problem Definition and Planning Stage
• 3.2.2 Data Collection and Processing
• 3.2.3 FM Evaluation and Selection
• 3.2.4 FM Training and Fine-Tuning
• 3.2.5 Application and Orchestration Layer Development
• 3.2.6 Testing, Validation, Monitoring, and Auditing
• 3.2.7 Production Deployment
• 3.2.8 Continuous Monitoring, Auditing, and  Fine-Tuning
• 3.3 GenAI Risks and Challenges Within the  Public Sector
• 3.3.1 Data Bias
• 3.3.2 Data Privacy and Security
• 3.3.3 Content Safety: Misinformation and Disinformation
• 3.3.4 Lack of Transparency and Explainability
• 3.3.5 Social and Economic Impact
• 3.3.6 Model Bias and Discrimination
• 3.3.7 Regulatory Compliance, Legal, Copyright, and Liability
• 3.3.8 Challenges with People, Process, Technology, and Data
• People
• Process
• Technology
• Data
• 3.4 High-Level GenAI Implementation Framework
• 3.4.1 Implementation from a People Perspective
• Establish the Strategic Vision and Policies for GenAI
• Establish a Path for Gradual Culture Evolution Toward GenAI
• Establish AI Centers of Excellence (ACOEs)
• Establish a Culture of Innovation
• Identify the Right Workforce
• Enable the Workforce to Adopt GenAI
• 3.4.2 Implementation from a Process Perspective
• Problem Definition and Benefit
• Risk Assessment and Categorization
• Procurement Guidelines
• Responsible AI
• Process Automation
• 3.4.3 Implementation from a Technology Perspective
• Building an FM or Selecting a Pre-trained FM
• Procuring a Pre-trained FM
• Infrastructure
• Tools for Building an Application
• Out-of-the-Box Applications
• 3.4.4 Implementation from a Data Perspective
• 3.5 Conclusion
• Chapter 4: Building a Generative AI Application.
• 4.1 Anatomy of a Prompt
• 4.2 Prompt Engineering
• 4.2.1 Zero-Shot Learning
• 4.2.2 Few-Shot Learning
• 4.2.3 Chain-of-Thought Prompting
• 4.2.4 Prompt Chaining
• 4.2.5 Prompt Templates
• 4.2.6 ReAct (Reasoning + Acting)
• 4.2.7 Agents
• 4.3 Best Practices for Constructing Prompts
• 4.3.1 Clarity and Specificity
• 4.3.2 Structured and Logical Format
• 4.3.3 Understanding FM Capabilities and Limitations
• 4.3.4 Iterative Approach
• 4.3.5 Use of Examples
• 4.3.6 Avoiding Biases and Assumptions
• 4.3.7 Creative Use of Prompts
• 4.3.8 Ethical Considerations
• 4.4 Model Parameters and Configurations
• 4.4.1 Temperature: The Creativity Knob
• 4.4.2 Top-k Sampling: Focusing the Output
• 4.4.3 Top-p Sampling: Balancing Control and Exploration
• 4.5 Model Evaluation
• 4.6 Techniques to Handle Domain-Specific Data
• 4.6.1 Retrieval-Augmented Generation (RAG)
• 4.6.2 Evaluating RAG Responses
• 4.6.3 FM Fine-Tuning
• 4.6.4 RAG vs. Fine-Tuning
• Focus
• Data Usage
• Resource Intensity
• Applications
• Flexibility
• Explainability
• 4.7 GenAI Application Architecture
• 4.8 Conclusion
• Chapter 5: Content Generation
• 5.1 General Areas of GenAI Content Generation
• 5.1.1 Document Generation
• 5.1.2 Image Generation
• 5.1.3 Code Generation
• 5.2 Document Generation
• 5.2.1 High-Level Architecture
• 5.2.2 Key Considerations
• Front-End Application UI
• Handling Large and Complex Document Generation
• Multiple Models
• Human Review
• Data Quality
• Cost
• 5.2.3 Document Generation Use Case 1: Contracts and Procurement
• Problem
• Solution
• Outcome and Benefits
• Additional Considerations
• 5.2.4 Document Generation Use Case 2: Public Communication, Alerts, and Notices
• 5.3 Image Generation
• 5.3.1 High-Level Architecture.
• 5.3.2 Key Considerations
• 5.3.3 Image Generation Use Case 1: Synthetic Image Generation for Medical Research
• 5.3.4 Image Generation Use Case 2: Education
• 5.4 Code Generation
• 5.4.1 Code Generation Evaluation Framework
• 5.5 Conclusion
• Chapter 6: GenAI-Powered Chatbots
• 6.1 Differences Between Traditional and GenAI-Based Techniques
• 6.2 Broad Areas of GenAI Chatbot Usage and Challenges
• 6.3 High-Level Architecture
• 6.4 Key Considerations
• 6.4.1 Data Privacy and Content Filtering
• 6.4.2 Document Chunking
• 6.4.3 Handling Complex Documents
• 6.4.4 Graph Databases
• 6.4.5 Memory Management
• 6.4.6 Context Switching
• 6.4.7 Fine-Tuning
• 6.4.8 Integration with the UI
• 6.4.9 Triggering Other Actions
• 6.4.10 Performance Considerations
• 6.4.11 Cost Considerations
• 6.5 Example Use Cases
• 6.5.1 Use Case 1: HR Self-Service
• 6.5.2 Use Case 2: Internal Knowledge Search
• 6.5.3 Use Case 3: Constituent Chatbot
• Outcomes and Benefits
• 6.6 Conclusion
• Chapter 7: Summarization
• 7.1 High-Level Architecture
• 7.2 Key Considerations
• 7.2.1 Batch or Real-Time Processing
• 7.2.2 Context Window Size and Length of Document
• Strategies for Summarizing Long Documents
• 7.2.3 Cost Control
• 7.3 Example Use Cases
• 7.3.1 Summarization Use Case 1: Policies and Guidelines
• 7.3.2 Summarization Use Case 2: Research Paper Abstracts
• Outcome and Benefits.
• 7.3.3 Summarization Use Case 3: Comment Summarization
• 7.4 Conclusion
• Chapter 8: Program Management, Business Intelligence, and Reporting
• 8.1 Broad Areas Where GenAI Can Assist with Reporting, Business Intelligence, and Analytics
• 8.1.1 Report Generation
• 8.1.2 Business Intelligence
• 8.1.3 Analysis of Large Datasets
• 8.1.4 Data Visualization and Storytelling
• 8.1.5 Predictive Analytics
• 8.1.6 Interactive Data Exploration and "What-If" Analysis
• 8.1.7 Document Analysis and Insights
• 8.2 Business Intelligence, Data Analytics, and Reporting High-Level Architecture
• 8.2.1 Workflow Steps
• 8.3 Key Considerations
• 8.3.1 Front-End Application and UI Design
• 8.3.2 Prompt Design and Prompt Template Library
• 8.3.3 Query Handling, Integration with Multiple Metadata, and Data Sources
• 8.3.4 Cost
• 8.4 Data Analytics and Reporting: Use Cases
• 8.4.1 Use Case 1: Compliance and Reporting
• 8.4.2 Use Case 2: Improve Customer and Employee Experience
• 8.4.3 Use Case 3: Program and Project Management
• 8.4.4 Use Case 4: Cyber Security, Threat Intelligence, and Analytics
• 8.5 Conclusion
• Chapter 9: Implementation, Operations, and Maintenance
• 9.1 Anatomy of a GenAI Application
• 9.2 Foundation Model Operations (FMOps)
• 9.2.1 Manage and Govern
• 9.2.2 Evaluate
• 9.2.3 Experiment
• 9.2.4 Test
• 9.2.5 Deploy
• 9.2.6 Infer and Monitor
• 9.2.7 Secure
• 9.3 FMOps Benefits
• 9.4 Performance, Reliability, and Scalability of GenAI Applications
• 9.4.1 Application
• 9.4.2 Orchestration
• 9.4.3 Model.
• 9.4.4 Data.

Show 233 more Contents items

ISBN

9798868804731 ((electronic bk.))

OCLC

1450302566

Statement on language in description

Princeton University Library aims to describe library materials in a manner that is respectful to the individuals and communities who create, use, and are represented in the collections we manage. Read more...