Centre for Computing Science Education

SPEAKER: Zoltan Sojtory
TITLE: Large Language Model-Aided Pair Programming for Algorithm Tracing
DESCRIPTION:
The widespread popularity of generative AI models has inspired the development of numerous large- language-model (LLM) based tools for educational purposes. We explore LLM-aided pair programming specifically for algorithm tracing, in an attempt to combine the benefits of these two traditional educational techniques. Our main objective is to address challenges inherent to pair programming while utilising the flexibility of LLM tools.

SPEAKER: Manuel Simonetta
TITLE: Exploring the Use of Large Language Models in Thematic Analysis
DESCRIPTION:
This research investigates the use of Large Language Models (LLMs) such as OpenAI's ChatGPT and Google Bard, into thematic analysis (TA), a qualitative data analysis method. By utilising advanced prompt engineering techniques and iterative refinements, we explore how different prompting strategies influence the quality and depth of TA insights produced by LLMs. The research method is experimental, using an iterative prompt development approach and sample data. Output quality is assessed in three ways, by triangulation, using quality criteria and comparing LLM outputs to human-conducted thematic analysis. We contribute an experimental framework for LLM-conducted thematic analysis and envision that LLMs can be effectively used as co-analysts alongside humans. The paper contributes to addressing tensions surrounding technology integration in social science research and paves the way for further exploration of ethical and effective use of LLMs in research.

First published: 25 April 2024