1 Knowledge graphs and LLMs: a killer combination

 

This chapter covers

  • Introduction to Knowledge Graphs (KGs)
  • Introduction to Large Language Models (LLMs)
  • How to combine KGs and LLMs to get the best of two
  • How to build data-driven applications using KGs and LLMs

Artificial intelligence (AI) has transformed how we work and live, with Generative AI technologies revolutionizing business after business. For example, ChatGPT, launched by OpenAI on November 30, 2022, reached 1 million users in days and over 100 million in months, marking the beginning of a new era where simple chat interfaces allow users to interact using natural language without programming knowledge.

Despite this success, these technologies fall short in domains like biomedicine, law enforcement and others where specific domain knowledge, high accuracy, and explainability are mandatory. Limitations include black-box models with limited explainability, infrequent updates due to lengthy training cycles, generic training unsuitable for specialized domains, hallucination, and lack of context and relations. This is where knowledge graphs come into play, providing the contextual information, like experiences, environmental characteristics, cultural aspects, and social norms, needed to build the "third wave of AI" [1] for mission-critical applications.

1.1 Knowledge graphs

1.2 Large Language Models

1.3 KGs and LLMs: stronger together

1.4 The paradigm shift in data-driven applications

1.4.1 The four pillars of knowledge graphs (KGs)

1.5 Building data-driven applications using KGs and LLMs

1.5.1 Example use case: Drug discovery and development

1.5.2 Example Use Case: Conversational AI for customers support

1.5.3 What should I ask myself?

1.6 Knowledge graph technologies

1.6.1 Taxonomies and ontologies

1.7 How do we teach knowledge graphs and LLMs?

1.8 Summary

1.9 References