Machine learning, neural interfaces and software development
Latest project · Brain–computer interfaces
Decoding imagined movement from recorded brain signals
IntentLab predicts imagined left or right fist movement from recorded EEG, with an interactive demonstration that shows how model confidence affects cursor control.
A group project combining neural networks with neighbourhood graph features to predict California house prices, reporting a 35% reduction in test error against its baseline.
Look through the code for House prices & neighbourhoods (opens in a new tab)A macOS automation agent combining screen reading, action planning and SQLite memory to carry out tasks described in everyday language.
Look through the code for Desktop automation agent (opens in a new tab)Controlled trials investigating how population size and mutation rate affect convergence, performance and variation across evolutionary algorithms.
Look through the code for Evolutionary optimisation (opens in a new tab)A web crawler, a SQLite link graph and an iterative PageRank implementation, with a D3 visualiser to explore how pages connect.
Look through the code for PageRank visualiser (opens in a new tab)A genetic algorithm that evolves random strings towards a target Shakespearean sonnet through selection, mutation and crossover.
Look through the code for Evolving Shakespeare (opens in a new tab)A Next.js portfolio featuring an interactive pixel-art simulation, accessible navigation and a central file for editing the main website copy.
Look through the code for Portfolio website (opens in a new tab)