View Proposal

---

Proposer

Michael Lones

Title

Can large language models replace traditional classifiers?

Goal

Description

Large language models (LLMs) are increasingly being used as a new form of machine learning system: instead of training a specialised model, users simply provide data and ask the LLM to make predictions, classifications, or decisions directly. This approach is attractive because it requires little or no additional training and can potentially leverage the broad knowledge already encoded within modern LLMs. However, it remains unclear how effective this strategy is in practice, particularly when compared with conventional machine learning techniques. This project will investigate the ability of LLMs to perform data classification tasks using established benchmark datasets. One avenue of research is to evaluate their performance on tabular data, where features have already been extracted and traditional methods such as decision trees, support vector machines, and neural networks have long been the standard approach. The project will explore whether LLMs can match or exceed the accuracy of these models and identify situations in which they may offer advantages. A second focus could be understanding the limitations of LLMs as classifiers. Unlike conventional machine learning models, LLMs are designed for language generation and may exhibit behaviours such as hallucination, inconsistency, or sensitivity to prompt wording. The project could investigate how these characteristics affect classification performance and reliability, and whether prompting strategies can be used to improve results. This project offers the opportunity to explore an emerging area at the intersection of machine learning and generative AI, addressing a fundamental question: when should we use large language models as decision-makers, and when are traditional machine learning approaches still the better choice?

Resources

Background

Url

Difficulty Level

Variable

Ethical Approval

None

Number Of Students

2

Supervisor

Michael Lones

Keywords

machine learning, llm

Degrees

Bachelor of Science in Computer Science
Bachelor of Science in Computer Systems
Master of Engineering in Software Engineering
Master of Science in Artificial Intelligence
Master of Science in Computer Science for Cyber Security
Master of Science in Computing (2 Years)
Master of Science in Data Science
Bachelor of Engineering in Robotics
Bachelor of Science in Computer Science (Cyber Security)
MSc Applied Cyber Security