دانلود کتاب گراف دانش و LLM ها در عمل

When I was nearing completion of my previous book, Graph-Powered Machine Learning, I reached out to my acquisitions editor, Mike Stephens, with a proposal for a natural continuation. That earlier work introduced knowledge graphs and demonstrated how they could be built using natural language processing, but many readers pointed out that graph neural networks were a significant missing piece. My proposed book would fill that gap while extending the knowledge graph story further, including detailed analysis and building techniques. Mike accepted the proposal, and I embarked on a new adventure with the working title Knowledge Graphs Applied. Recognizing the scope of the challenge, I invited three colleagues from GraphAware—Fabio, Giuseppe, and Vlastimil—to join the effort, confident that their combined expertise would be invaluable. I naively thought that if one author could write a book in four years, four authors could complete a book in just a year. That assumption proved as flawed as expecting nine women to deliver a baby in one month. Reality had other plans. Over the past years, significant changes swept through the technology landscape. Large language models (LLMs) and generative AI disrupted the field entirely, and knowledge graph practitioners suddenly found themselves with unprecedented opportunities to use this established technology in revolutionary ways. We initially planned to build on existing natural language processing (NLP) tools like BERT, but these were rapidly being superseded by LLM capabilities that opened new possibilities for building, querying, and analyzing knowledge graphs. This was precisely where many practitioners, ourselves included, were struggling. Rather than resist this transformation, we decided, together with Mike and Dustin Archibald (our development editor), to embrace it. We adjusted our title to Knowledge Graphs and LLMs in Action and substantially revised the content to position LLMs as an integral component of our ultimate goal: intelligent advisor systems that empower humans in performing complex decision-making tasks. This pivot required extensive refactoring and a fundamental shift in our approach, but the result exceeded our expectations. The book you are reading has evolved into a manifesto for the power of hybrid systems. It demonstrates how combining these technologies—knowledge graphs, which are well established, and LLMs, which are newly emerged— creates a flywheel effect that delivers remarkable long-term results. Knowledge graph practitioners will discover how to use LLM capabilities for greater impact, and LLM practitioners will learn techniques that address some of the major limitations of language models.