Large Language Models, Knowledge Graphs and Search Engines: A Crossroads for Answering Users' Questions

TL;DR

The paper investigates how Large Language Models, Knowledge Graphs, and Search Engines can be combined to better serve diverse user information needs. It develops a user-centric taxonomy of information requirements, compares the strengths and weaknesses of SEs, KGs, and LLMs, and argues for complementary, cross-technology solutions. The authors outline concrete research directions, including augmenting LLMs with curated KG knowledge, enhancing SEs with LLM capabilities, and using LLMs to generate or refine KG content, all within an evolving four-phase framework of augmentation, ensemble, federation, and amalgamation. The work highlights the importance of user perspective, proposes a roadmap for future work, and emphasizes practical considerations such as freshness, provenance, efficiency, and long-tail query handling to guide the design of integrated information systems.

Abstract

Much has been discussed about how Large Language Models, Knowledge Graphs and Search Engines can be combined in a synergistic manner. A dimension largely absent from current academic discourse is the user perspective. In particular, there remain many open questions regarding how best to address the diverse information needs of users, incorporating varying facets and levels of difficulty. This paper introduces a taxonomy of user information needs, which guides us to study the pros, cons and possible synergies of Large Language Models, Knowledge Graphs and Search Engines. From this study, we derive a roadmap for future research.

Figures (4)

• Figure 1: Replies from three LLMs to a user question
• Figure 2: LLM reply to a long-tail question
• Figure 3: Potential combinations of SEs, KGs and LLMs
• Figure 4: Knowledge graph describing Manuel Blum