Having reached the halfway point of my research, I am very proud of how far I have come. It has been a steep learning curve especially during the first couple of weeks as I was doing a lot of background reading which, as a STEM student, is not my usual modus operandi! Am I reading useful literature? When do I stop reading? How far down a reference rabbit hole is too far?
Nevertheless, I have gained insight into educational psychology tools and tricks that I will definitely implement into my regular studies. I have also completed a prompt engineering course by DeepLearning.AI to understand how to optimise GPT technology. I have thoroughly enjoyed exploring the capabilities of ChatGPT 4o and other GPTs, such as ‘physics’, ‘maths mastermind’ and ‘Mermaid Charts’, as well as pushing the boundaries of the technology by creating and refining my own GPT – the Physualiser. I have trained it on exemplar data which I honed myself (in-distribution data) and am now testing it using inputs that are similar to the training data but, crucially, not part of it (out-distribution data).
Now, however, I am almost done with the data collection phase of the project and will be moving onto the analysis and representation. I am looking forward to collating and comparing my results over the coming weeks.