Training teams of AI to play "Capture the Flag" through the use of Deep Q-Learning with a hybrid CNN and FCN pipeline for local perception and decisions.

Finding out if collective intelligence can emerge at a faster rate through the use of a direct, inter-agent communication system.

• Python
• PyTorch
• Deep Q-Learning

Working with a large dataset to create a model that is capable of forecasting the yield of different crops in multiple countries across the globe by using the data from previous years.

• Python
• PyTorch
• Predictive Networks

Studying the effects of worm viruses on connected peer-to-peer networks with user action accounted for. This is mimicing the WannaCry virus that affected the NHS in 2017.

• Python
• Network Analysis
• Simulated Attacks

Creating language models to process and identify both single words, and spans of text that show signs of propaganda. Comparing simple and complex models.

• Python
• Large Language Models
• Text Processing

Showing how self-orgainsed criticality is present in extinction dynamics on ecosystems and their foodchain dependencies.

• Python
• Network Dynamics
• Biological Modelling

Exploring how the size of an environment can encourage or limit the extent that training AI agents for a simple game of "Hunters vs Runners" can be successful.

• Python
• Game AI
• Deep Q-Learning

Experimenting with the effects of climate fluctuation and median temperature on both endothermic and ectothermic orgainisms. SHowing how climate change may affect these different species and the survival of their populations.

• Python
• Biological Modelling
• ABM