Combining Large Language Models with Domain Knowledge
In the rapidly evolving world of artificial intelligence (AI), large language models (LLMs) have emerged as pivotal in reshaping how people interact with digital information. However, as industries are increasingly specialised, there is an emerging demand for LLMs to possess not just extensive humanlike outputs for general information, but also precise, trustworthy domain-specific details. LLMs often find their limits in sectors known for their profound and intricate expertise.
Huang, Jessica, 2024
Art der Arbeit Master Thesis
Auftraggebende
Betreuende Dozierende Hinkelmann, Knut
Views: 11 - Downloads: 2
One promising solution to address these limitations is to enhance LLMs with domain-specific knowledge. This involves cross-disciplinary efforts, merging natural language processing (NLP) with expert insights.
In this master thesis proposal, my primary goal in the literature review is to investigate the inherent deficits and limitations and respectively their current solutions, as well as applications of LLMs in information retrieval.
I have deployed design science research combined with case study as my research strategy. This research demonstrated this approach by developing a knowledge base for the "Master of Science in Business Information Systems" programme at the FHNW University of Applied Sciences and Arts Northwestern Switzerland (FHNW). This knowledge base was integrated with a language model, specifically tailored to the organisation by using retrieval augmented generation (RAG)-based technology with programme-related documents. The aim was to create a practical, testable solution that efficiently assists students.
Studiengang: Business Information Systems (Master)
Keywords Large language model, natural language processing, domain knowledge
Vertraulichkeit: öffentlich