University Projects

My dissertation for my final year is around using AI models for the novel task of analysing disinformaiton, not just detecting it. The main agent is written in Typescript with LangGraph, with supporting work in Python for data analysis.

The project covers a range of topics from NLP, to AI training and software engineering

Abstract

Disinformation on the internet has become a significant and growing challenge, driven by the sheer volume and speed of content generation across digital platforms.

While prior work has thoroughly investigated tasks such as automatic disinformation detection, especially using large language models, comparatively little attention has been paid to applying these techniques to generative analyses of claims. In particular, the retrieval of structured contextual background information that explains why disinformation spreads remains underexplored.

This paper introduces a dataset of trigger events, defined as actions or statements that influence the creation, spread or believability of disinformation claims. The dataset is created using large language models as few-shot retrievers and subsequently validated through an ensemble classifier to ensure consistency.

We find that we can successfully retrieve a large dataset of trigger events with high accuracy, and that the dataset proves useful in applications that can be useful in the real-world scenarios of debunking and prebunking.

The dataset is available on HuggingFace

As part of the dataset creation, it was important to prove usefulness. One possible use was a graph-based visualtion. Trigger events and disinformation claims are plotted using a force graph, and their links are visualised. This provides a never-before-seen view of the causes and effects of disinformation propogation.

LINK

As part of the dataset creation, it was important to prove usefulness. One other possible use was a prediction engine for possible upcoming disinformaiton claims, based on existing trends following real-world events.

For Mobile Development, as part of a large team, we created a management app for a landlord agency. I worked on many of the key aspects such as dynamic maps, syncing, popups and more. Linked below is the submission walkthrough video.

LINK

For 3D computer graphics, I created a complex 3D scene using the Java wrapper for OpenGL, JOGL. The project demonstrates multiple configurable lights (including spotlights), character hieracies and animation, billboards, a complex skybox and much more. Linked below is a walkthrough video of the project.

LINK

For Bioinspired computing I completed a team project in which we researched, implemented and analysed a robust wildfire simulation tool, based on Cellular Automata. We were able to model material types, density, temperature, spotting and more using an extensible CA setup.

University Team Project in which we must create and present a web application in Ruby for the discovery of on-line courses. This is written using ruby and sinatra, using erb for the creation of dynamic webpages and sqlite as a backend. This project demonstrates my ability to work collaborativley as well as my commuication skills during conversations with both the client and larger presentations.

As part of a machines and intelligence module my team worked on creating a model based reflex agenct capable of navigating in a real world environment using Python and the University of Sheffield’s MiRo robots. This project cumulated in a final report showing the development process and data analytics showing a consistently fast solve time for any situations it was given.