Semantic Data 2024: Taxonomy, Ontology, and Knowledge Graphs

A case study of developing an information discovery application that integrated a knowledge graph with a large language model to ensure that all that the organisation had learned could be accessed.

The goals:

• To demonstrate how well-established knowledge management techniques can enhance the effectiveness and user-friendliness of contemporary Retrieval-Augmented Generation (RAG) information retrieval systems.
• To show how stringent production requirement typical of regulated enterprises, delivers relevance, compliance and security.
• To discuss how innovative techniques can overcome traditional obstacles in information retrieval.

The problems:

• Users frequently submit (natural language) queries that are poorly constructed, vague, ambiguous, lacking clear and adequate context and/or contain narrow domain-specific terminology.

The solutions:

• Address each problem – one by one.

The outcomes:

• A measurable improvement in the discovery of relevant information.
• Increased ease of use.
• Improved findability, understanding and trust in the process, leading to increased reuse of assets, greater repurpose capabilities and better performance in the daily work of knowledge workers and many other specialists.