Scaling Graph Learning for the Enterprise Production-Ready Graph Learning and Inference (Ahmed Menshawy, Sameh Mohamed etc.) (z-library.sk, 1lib.sk, z-lib.sk)

📄 File Format: PDF

💾 File Size: 4.0 MB

📖 Read Online ⬇️ Download

Registered users can read the full content for free

Register as a Gaohf Library member to read the complete e-book online for free and enjoy a better reading experience.

📄 Page 1

Ahmed Menshawy, Sameh Mohamed & Maraim Rizk Masoud Scaling Graph Learning for the Enterprise Production-Ready Graph Learning and Inference

📄 Page 2

9 7 8 1 0 9 8 1 4 6 0 6 1 5 7 9 9 9 ISBN: 978-1-098-14606-1 US $79.99 CAN $99.99 DATA Tackle the core challenges related to enterprise-ready graph representation and learning. With this hands-on guide, applied data scientists, machine learning engineers, and practitioners will learn how to build an E2E graph learning pipeline. You’ll explore core challenges at each pipeline stage, from data acquisition and representation to real-time inference and feedback loop retraining. Drawing on their experience building scalable and production-ready graph learning pipelines, the authors take you through the process of building robust graph learning systems in a world of dynamic and evolving graphs. • Understand the importance of graph learning for boosting enterprise-grade applications • Navigate the challenges surrounding the development and deployment of enterprise-ready graph learning and inference pipelines • Use traditional and advanced graph learning techniques to tackle graph use cases • Use and contribute to PyGraf, an open source graph learning library, to help embed best practices while building graph applications • Design and implement a graph learning algorithm using publicly available and syntactic data • Apply privacy-preserving techniques to the graph learning process Ahmed Menshawy is the vice president of AI engineering at Mastercard. Sameh Mohamed is a senior applied scientist at Microsoft and an expert in machine learning and health informatics. Maraim Rizk Masoud is a lead machine learning engineer. Scaling Graph Learning for the Enterprise “Scaling Graph Learning for the Enterprise is a comprehensive, hands-on guide to building scalable and enterprise-ready graph learning pipelines. This book equips you with all the tools and knowledge to bring cutting-edge graph capabilities into your enterprise environments.” Lipi Patnaik, senior software developer, Zeta “In an era of boundless connectivity, this book walks you through the phases of building scalable graph representation learning to unlock the structure and meaning concealed within complex networks.” Dr. Emir Muñoz, senior manager of AI/ML, Genesys Cloud Services

📄 Page 3

Praise for Scaling Graph Learning for the Enterprise Scaling Graph Learning for the Enterprise is a comprehensive, hands-on guide to building scalable and enterprise-ready graph learning pipelines. From dynamic graph representation to real-time inference with federated graph learning, it equips you with all the tools and knowledge to bring cutting-edge graph capabilities into your enterprise environments. —Lipi Patnaik, senior software developer, Zeta In an era of boundless connectivity, this book walks you through the phases of building scalable graph representation learning to unlock the structure and meaning concealed within complex networks. —Dr Emir Muñoz, senior manager AI/ML, Genesys Cloud Services I am incredibly impressed by the sheer volume of topics this book covers, taking you all the way from foundational concepts to advanced, enterprise- scale strategies. What truly makes it exceptional is how every complex idea is grounded in a clear, practical example. It’s that rare resource that serves as both a comprehensive reference and an easy-to-follow, hands-on guide. —Mahmoud Fahmy Mohammed, lead AI engineer I thoroughly enjoyed this book. It provided a clear and well-structured path to understanding graph learning and its value in the corporate world. It’s presented in such an intuitive way that I’d recommend it to anyone being introduced to the subject for the first time. —Mohamed Elemam, senior cloud developer, HPE

📄 Page 4

(This page has no text content)

📄 Page 5

Ahmed Menshawy, Sameh Mohamed, and Maraim Rizk Masoud Scaling Graph Learning for the Enterprise Production-Ready Graph Learning and Inference

📄 Page 6

978-1-098-14606-1 [LSI] Scaling Graph Learning for the Enterprise by Ahmed Menshawy, Sameh Mohamed, and Maraim Rizk Masoud Copyright © 2025 Ahmed Menshawy, Sameh Mohamed, and Maraim Masoud. All rights reserved. Published by O’Reilly Media, Inc., 141 Stony Circle, Suite 195, Santa Rosa, CA 95401. O’Reilly books may be purchased for educational, business, or sales promotional use. Online editions are also available for most titles (http://oreilly.com). For more information, contact our corporate/institutional sales department: 800-998-9938 or corporate@oreilly.com. Acquisitions Editor: Nicole Butterfield Development Editor: Michele Cronin Production Editor: Elizabeth Faerm Copyeditor: Charles Roumeliotis Proofreader: Helena Stirling Indexer: WordCo Indexing Services, Inc. Cover Designer: Karen Montgomery Cover Illustrator: Karen Montgomery Interior Designer: David Futato Interior Illustrator: Kate Dullea August 2025: First Edition Revision History for the First Edition 2025-08-06: First Release See http://oreilly.com/catalog/errata.csp?isbn=9781098146061 for release details. The O’Reilly logo is a registered trademark of O’Reilly Media, Inc. Scaling Graph Learning for the Enterprise, the cover image, and related trade dress are trademarks of O’Reilly Media, Inc. The views expressed in this work are those of the authors and do not represent the publisher’s views. While the publisher and the authors have used good faith efforts to ensure that the information and instructions contained in this work are accurate, the publisher and the authors disclaim all responsibility for errors or omissions, including without limitation responsibility for damages resulting from the use of or reliance on this work. Use of the information and instructions contained in this work is at your own risk. If any code samples or other technology this work contains or describes is subject to open source licenses or the intellectual property rights of others, it is your responsibility to ensure that your use thereof complies with such licenses and/or rights.

📄 Page 7

Table of Contents Preface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii 1. Introduction to Graphs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 The Power of Enterprise Graph Learning and Inference at Scale 1 A Bird’s-Eye View: Navigating the Book’s Chapters 6 Graphs and Graph Learning 7 What Is a Graph? 7 Graph Data Representation 9 Graph Learning 11 Scalable Graph Learning: Addressing the Requirements 13 Advantages of Scalable Graph Learning in Enterprise 14 Large-Scale Graphs in Real-World Enterprises: Use Cases 15 Travel-Time Predictions on Google Maps 16 Drug Development: Halicin 16 Fraud Detection 17 The Evolution of Graphs and Graph Learning: From Early Beginnings to Modern Applications 19 Era 1: The Foundation of Graph Theory and Algorithms (1736-1970) 20 Era 2: More Advancement in Graph Algorithms and Technologies (1970-1999) 20 Era 3: Emergence of Graph Databases and Graph Query Languages (2000-2006) 21 Era 4: Graph Analytics and Traditional Machine Learning (2007-2011) 21 Era 5: Rise of Graph Neural Networks (2012-2018) 21 Era 6: Scalability, Robustness, and Enterprise Applications (2019-Present) 22 Challenges of Enterprise-Ready Graph Learning Systems 22 Data Harmonization Challenges 22 Computationally Intensive Workloads 23 v

📄 Page 8

Dynamic Evolving Graphs 23 Active Monitoring and Drift Detection 24 Real-Time Inference 24 Summary 25 2. The Graph Machine Learning Pipeline. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 The Graph Data Pipeline 28 Definition of Graph Data and Graph Data Levels 28 Graph Data Sourcing and Understanding 34 Graph Data Preparation 36 The Graph Training and Inference Pipelines 41 GML Training Pipeline Overview 41 GML Inference Pipeline Overview 43 Summary 46 3. Traditional Machine Learning for Graphs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 Approaches to Graph Machine Learning 48 Traditional Graph-Based Machine Learning 48 Nontraditional Graph-Based Machine Learning 48 Representing Graphs for Traditional ML 49 Graph Representation 49 Representing Amazon Copurchasing Networks as Graphs 50 Navigating Graph Tasks in the Amazon Copurchasing Dataset 53 Graph Feature Engineering 56 Importance and Challenges 57 Types of Graph Features 58 Hands-on: Extracting Features for the Amazon Copurchasing Graph 61 Graph Features in ML Modeling 62 Task and Techniques Overview 63 Predicting High-Rated Products with a Prediction Model 64 Feature Learning with Node Embeddings 67 Random Walk Algorithm 68 Amazon Copurchasing Dataset and Node Embeddings 69 Summary 73 4. PyGraf: End-to-End Graph Learning and Serving. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 Graph Libraries Overview 75 Challenges of Open Source Graph Libraries: PyGraf Opportunities 76 PyGraf: A Solution for Streamlined Graph Learning and Serving 77 Introduction to PyGraf 77 PyGraf Key Features 77 PyGraf Purposes: Empowering Dynamic Environments 78 vi | Table of Contents

📄 Page 9

Architecture and Core Capabilities 79 Core Components Layer 81 Adaptation and Integration Layer 81 Best Practices Layer 82 In-Depth Exploration of Core Library Components 82 Data Component 82 Training Component 88 Serving Component 91 Privacy Preserving Component 92 End-to-End Example Using PyGraf: Amazon Copurchasing Dataset 93 Preprocessing and Transformation 94 Model Training 95 Evaluation and Model Selection 96 Deployment and Monitoring 96 Summary 97 5. Graph Neural Networks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 Introduction to Graph Neural Networks 100 The Significance of GNNs in Graph Learning 100 Overview of GNN Applications 101 Foundations of GNNs 102 Graph Convolutional Networks 104 Transition from Traditional Graph Learning to GCNs 107 How GCNs Simplify Feature Engineering 108 Hands-on E2E Example Using PyG and the PyGraf Interface 108 PyG: Karate Club Example 109 Cora Node Classification 120 PyGraf API Interface 125 Limitations of GCN-Based Architecture 128 Summary 129 6. Advanced Techniques in Graph Learning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 Different Types of Graphs 131 Homogeneous Graphs 132 Heterogeneous Graphs 132 Temporal Graphs 133 Graph Embedding Models 134 How Do Knowledge Graph Embeddings Work? 134 Training Knowledge Graph Embedding Models 135 Embedding Interaction Methods 136 Training Objectives 139 Strengths of Graph Embedding Models 139 Table of Contents | vii

📄 Page 10

Example: Learning on Freebase Dataset 140 Attention on Graphs 145 Feature Vectors 145 Attention Mechanism 146 Multihead Attention 147 Example: Citation Network 147 CiteSeer Use Case Example 150 Data Preprocessing and Visualization 150 Model Training Using PyG 153 Model Testing 155 Embeddings Visualization 155 Summary 156 7. Scalable Graph Neural Networks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 Challenges in Scaling Graph Learning Models 160 Computational Challenges 160 Memory Constraints 160 Data Complexity and Size 161 Mini-Batching in Graph Neural Networks 162 Definition and Importance of Mini-Batching 162 Techniques for Effective Mini-Batching 162 Memory-Efficient Training Techniques 164 Gradient Checkpointing 164 Subgraph Sampling Methods 165 Layer-Wise Relevance Propagation 167 Distributed Data and Compute Strategies 168 Distributed Execution Strategies 168 Graph Partitioning Strategies 171 Distributed Graph Learning Tools 173 Distributed Training with PyG 176 Prepare and Partition the Graph Data 176 Manual Execution of Distributed Training 177 Key Components of Distributed Training with PyG 178 Advanced Architectures for Scaling Graph Neural Networks 179 Sparsity Exploitation 180 Approximation Techniques 180 Use of External Memory 180 Summary 181 8. Enterprise Applications of Graphs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183 Customer and Market Insights 184 Customer Segmentation 185 viii | Table of Contents

📄 Page 11

Social Network Analysis 186 Recommendation Engines 188 Operations and Supply Chain Management 192 Network Optimization in Supply Chains 192 Inventory Management 194 Risk Management and Supply Chain Resilience 194 Real-World Applications 195 Security and Risk Management 197 Threat Detection and Analysis 197 Identity and Access Management 198 Fraud Detection 198 Cybersecurity Incident Response 199 Risk Management and Compliance 199 Healthcare and Life Sciences 200 Patient Journey Mapping and Personalization 201 Drug Discovery and Development 201 Clinical Trial Optimization 202 Genomics and Personalized Medicine 203 Popular Life Science Products 204 Summary 206 9. Privacy-Preserving Graph Learning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 The Importance of Privacy in Graph Learning 209 Enterprise Examples and Applications 210 Overview of Privacy-Preserving Techniques 210 Privacy Threats in Graph Learning 212 Case Studies of Privacy Breaches 214 Privacy-Preserving Techniques for Graph Data 214 Graph Data Anonymization 215 k-Anonymity and t-Closeness in Graphs 215 Graph Modification Techniques 215 Edge Differential Privacy 217 Node Differential Privacy 217 Synthetic Graph Data Generation 218 Privacy-Preserving Graph Computation 219 SMPC for Graphs 219 Homomorphic Encryption in Graph Learning 221 Differentially Private Learning Algorithms 222 Federated Graph Learning 224 Principles and Architecture 225 Federated Graph Neural Networks 228 Training and Communication Protocols 229 Table of Contents | ix

📄 Page 12

Privacy Enhancements in Federated Graph Learning 230 Applications of Privacy-Preserving Graph Learning 233 Privacy-Preserving Techniques in Social Network Analysis 233 Use Cases 235 Case Study: Applying FedGraphNN to a Recommender System Using Epinions Data 235 Dataset 235 Importance of Applying Federated Learning 236 Data Processing 236 Federated Training Logic 238 Running the Federated Pipeline 241 Summary 243 10. Graph Inference and Deployment Strategies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245 Deployment Strategies 246 Canary Deployments: Testing the Waters 247 Blue–Green Deployments: Smooth Transitions 249 Shadow Deployments: Flying Under the Radar 250 Edge Deployments: Taking It to the Edge 251 A/B Testing Deployments: Controlled Experiments 252 Progressive Rollouts: Scaling Up Gradually 253 Containerized Deployments 253 Hardware Considerations 254 Inference Runtimes 255 Inference Frameworks and Libraries 255 Precomputations and Caching 257 Online Versus Offline Inference 258 Model Optimization Techniques 259 Quantization 260 Pruning 264 Knowledge Distillation 267 Scaling Inference for Large Graphs 269 Distributed Inference Systems 270 Graph Partitioning for Scalable Inference 271 Incremental Inference 274 Summary 276 11. Monitoring and Feedback Loops. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279 Challenges in Monitoring Graph Models 280 Graph-Specific Metrics 282 Dynamic Graphs 284 Scale and Complexity 285 x | Table of Contents

📄 Page 13

System and Deployment Issues 286 Designing a Monitoring Framework 287 Dealing with Ever-Changing Connections (Graph Topology) 287 Checking the “Meaning” (Evaluating Embeddings) 288 The Balancing Act (System Performance Versus Graph Tasks) 288 Data Collection Mechanisms 289 Monitoring Deployed Graph Models 290 Thresholds for Alerts 297 Visualization Tools 298 Feedback Loops in Graph Systems 299 User Feedback: Learning from Interaction 300 System Feedback: Optimizing Model Performance 301 Data Feedback: Adapting to a Changing World 302 Synergy Between Feedback Loops 303 Closed-Loop Systems: Automating Adaptation 303 Adaptive Retraining Pipelines 305 Notification Systems 307 Summary 308 12. Future Trends: Graph Learning and LLMs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311 Introduction to Graph-Enhanced LLMs 311 LLMs and Their Transformative Impact 313 The Need for External Memory: The RAG Setup 314 Benefits of Graph-Enhanced RAG Setup 315 Retrieval Augmentation with Graph Integration 316 Challenges in Traditional Retrieval-Augmented Generation 316 Graph Integration 317 GraphRAG Methodology and Pipeline 319 Indexing Process 319 Querying Process 322 Enhanced Capabilities and Whole-Dataset Reasoning 323 Implementation Considerations 324 Baseline RAG Versus GraphRAG 326 RAG with Knowledge Graphs for Customer Service Question Answering 329 Summary 332 Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335 Table of Contents | xi

📄 Page 14

(This page has no text content)

📄 Page 15

Preface Welcome to Scaling Graph Learning for the Enterprise. We wrote this book with a single aim: to give professionals a clear, direct path from a first idea to a working system. In many technical books, the gap between theory and practice can feel like a canyon; in ours, it is meant to be a footbridge. We’ve tried to balance every page with just enough background information to make sense of the subject, along with concrete guidance that you can apply in the same afternoon. What Is Graph Learning for the Enterprise? During the last few years, developments in the field of graph technology and machine learning have been astonishing. With the growing recognition of the interconnected‐ ness of data, graph methods have moved from academic niches to a critical tool for understanding complex systems. From detecting sophisticated fraud rings to optimiz‐ ing supply chains and personalizing recommendations, graph learning is becoming vital across every industry. Despite this surge in interest, many practitioners find themselves navigating a land‐ scape with powerful tools but limited guidance on how to systematically apply graph learning in real-world enterprise settings. While model architectures and theoretical concepts have received significant attention, the practical aspects of building, deploy‐ ing, and maintaining robust graph-based systems often receive less focus. This book aims to bridge that gap. We believe that applying graph learning effec‐ tively in an enterprise context requires more than just understanding algorithms; it demands a structured approach to problem-solving, data integration, model deploy‐ ment, and ongoing maintenance. We will show you how to build a robust graph learning system that delivers actionable insights and is designed for the challenges of production environments. xiii

📄 Page 16

Why Does Graph Learning Matter for the Enterprise? In the rush for the most performant machine learning solutions, we’ve observed a few things that have received less attention, particularly in the realm of graph data. Data scientists and machine learning engineers often lack good sources of information for concepts and tools to accelerate, reuse, manage, and deploy graph-based develop‐ ments. What’s needed is a practical framework for applying graph learning. From our personal experience, most data science projects aiming to deploy models into production don’t have the luxury of a large team, especially when dealing with the unique complexities of graph data. This can make it incredibly difficult to build an entire graph learning pipeline in-house from scratch. Without a structured approach, graph projects often turn into one-off efforts where scaling becomes a significant challenge, performance degrades over time, or the model isn’t widely used across the enterprise. An automated, reproducible, and scalable pipeline for graph learning is therefore crucial, as it dramatically reduces the effort required to scale the deployment and maintenance of these powerful models. Our intention with this book is to contribute to the standardization of graph learning projects by walking readers through an entire graph learning pipeline, end to end. Clear writing alone isn’t enough if the reader is left guessing about context. For that reason, each technique is anchored to a story drawn from the banking, retail, cybersecurity, or healthcare domains, areas where a mistake is costly and timelines are tight. Just as important, you’ll read where early versions failed and what was changed to keep the project alive. These small detours are included so you can identify similar issues in your own work and hopefully avoid them later. We also know that most enterprise teams work under strict rules concerning data privacy, fairness, and uptime. Whenever those rules collide with a modeling choice, we call it out and show how to innovate within those boundaries to achieve your goals compliantly. The same applies to hardware limits and budget caps. The goal is to save you from discovering a deal-breaking constraint after weeks of effort. Open source tools now make it possible to build and run large graph models with a few commands, and the surge of interest in large language models lets us combine free-text knowledge with graph structure in ways that were impossible just a year ago. The book reflects that momentum but keeps the focus on what you can ship today. We describe emerging ideas, yet always bring the discussion back to code that will run on a modest cluster without an army of researchers behind it. xiv | Preface

📄 Page 17

Who Is This Book For? The primary audience for this book includes data scientists and machine learning engineers who want to go beyond training a one-off graph model and successfully productize their data science projects. You should be comfortable with basic machine learning concepts and familiar with at least one machine learning framework (e.g., PyTorch, TensorFlow, Keras). A secondary audience for this book includes managers of data science projects, software developers, and DevOps engineers who want to enable their organization to accelerate their data science projects with graph technologies. If you are interested in better understanding automated graph learning life cycles and how they can benefit your organization, this book will introduce a toolchain to achieve exactly that. Overview of the Chapters In each chapter, we will introduce specific steps for building effective graph learning systems and demonstrate how these work with practical examples. Chapter 1, “Introduction to Graphs”, provides an overview of graph structures, their applications, and why they are powerful for enterprise problems. Chapter 2, “The Graph Machine Learning Pipeline”, outlines the end-to-end process for building, deploying, and maintaining graph-based machine learning systems. Chapter 3, “Traditional Machine Learning for Graphs”, explores how classical machine learning techniques can be adapted and applied to graph data. Chapter 4, “PyGraf: End-to-End Graph Learning and Serving”, introduces PyGraf, a practical framework for building and serving graph learning models in a production environment. Chapter 5, “Graph Neural Networks”, dives into the foundational concepts and archi‐ tectures of graph neural networks (GNNs). Chapter 6, “Advanced Techniques in Graph Learning”, covers more sophisticated graph learning methods and their applications. Chapter 7, “Scalable Graph Neural Networks”, addresses the challenges of scaling GNNs to large, enterprise-sized datasets and discusses solutions. Chapter 8, “Enterprise Applications of Graphs”, showcases real-world use cases of graph learning across various industries, drawing from our experience. Chapter 9, “Privacy-Preserving Graph Learning”, explores techniques for building graph models while adhering to strict data privacy regulations. Preface | xv

📄 Page 18

Chapter 10, “Graph Inference and Deployment Strategies”, focuses on deploying graph learning models efficiently for real-time and batch inference. Chapter 11, “Monitoring and Feedback Loops”, discusses how to monitor the perfor‐ mance of deployed graph models and establish feedback mechanisms for continuous improvement. Chapter 12, “Future Trends: Graph Learning and LLMs”, provides an outlook on emerging technologies, particularly the intersection of graph learning and large lan‐ guage models (LLMs), and their potential impact. As you turn the pages, try to look at each new dataset through a simple lens: how do the records affect one another, and what can those links tell us that single rows cannot? That habit is often the spark that changes a minor improvement into a breakthrough. Thank you for giving this book a place on your desk. May it guide you to cleaner designs, stronger models, faster deployment, and more accurate insights. Open your favorite editor, start a notebook, and let’s map some relationships. Happy reading, and happy graphing! Conventions Used in This Book The following typographical conventions are used in this book: Italic Indicates new terms, URLs, email addresses, filenames, and file extensions. Constant width Used for program listings, as well as within paragraphs to refer to program elements such as variable or function names, databases, data types, environment variables, statements, and keywords. Constant width bold Shows commands or other text that should be typed literally by the user. Constant width italic Shows text that should be replaced with user-supplied values or by values deter‐ mined by context. This element signifies a general note. xvi | Preface

📄 Page 19

Using Code Examples Supplemental material (code examples, exercises, etc.) is available for download at https://github.com/gl4ebook/py-graf. If you have a technical question or a problem using the code examples, please send email to support@oreilly.com. This book is here to help you get your job done. In general, if example code is offered with this book, you may use it in your programs and documentation. You do not need to contact us for permission unless you’re reproducing a significant portion of the code. For example, writing a program that uses several chunks of code from this book does not require permission. Selling or distributing examples from O’Reilly books does require permission. Answering a question by citing this book and quoting example code does not require permission. Incorporating a significant amount of example code from this book into your product’s documentation does require permission. We appreciate, but generally do not require, attribution. An attribution usually includes the title, author, publisher, and ISBN. For example: “Scaling Graph Learning for the Enterprise by Ahmed Menshawy, Sameh Mohamed, and Maraim Rizk Masoud (O’Reilly). Copyright 2025 Ahmed Menshawy, Sameh Mohamed, and Maraim Masoud, 978-1-098-14606-1.” If you feel your use of code examples falls outside fair use or the permission given above, feel free to contact us at permissions@oreilly.com. O’Reilly Online Learning For more than 40 years, O’Reilly Media has provided technol‐ ogy and business training, knowledge, and insight to help companies succeed. Our unique network of experts and innovators share their knowledge and expertise through books, articles, and our online learning platform. O’Reilly’s online learning platform gives you on-demand access to live training courses, in-depth learning paths, interactive coding environments, and a vast collection of text and video from O’Reilly and 200+ other publishers. For more information, visit https://oreilly.com. Preface | xvii

📄 Page 20

How to Contact Us Please address comments and questions concerning this book to the publisher: O’Reilly Media, Inc. 141 Stony Circle, Suite 195 Santa Rosa, CA 95401 800-889-8969 (in the United States or Canada) 707-827-7019 (international or local) 707-829-0104 (fax) support@oreilly.com https://oreilly.com/about/contact.html We have a web page for this book, where we list errata, examples, and any additional information. You can access this page at https://oreil.ly/scaling-graph-learning-for-the- enterprise. For news and information about our books and courses, visit https://oreilly.com. Find us on LinkedIn: https://linkedin.com/company/oreilly-media. Watch us on YouTube: https://youtube.com/oreillymedia. Acknowledgements A project like this stands on many shoulders, and we are profoundly grateful to everyone who contributed to its creation. We extend our sincere thanks to the team at O’Reilly for their guidance and support throughout this process: Nicole Butterfield, Michele Cronin, Elizabeth Faerm, and Charles Roumeliotis. Our technical reviewers provided invaluable feedback that sharpened our exam‐ ples and clarified our explanations. We are especially grateful to Emir Muñoz, Mahmoud Fahmy, and Lipi Deepaakshi Patnaik for their diligent review and insight‐ ful suggestions. xviii | Preface