Professor Desmond Higham is an applied mathematician and Professor of Numerical Analysis. His main area of research is stochastic computation, with applications in artificial intelligence, data science, network science and computational biology. Des is interested in algorithm development and evaluation, as well as AI regulation.His research uses theory and algorithms to expose, and where possible fix, vulnerabilities in AI systems in high-stakes applications. Current AI projects Awards and fellowshipsFellow of the Royal Society of EdinburghRoyal Society/Wolfson Research Merit Award holderSIAM Fellow and editor-in-chief of SIAM ReviewFellow of Edinburgh's Generative AI Laboratory (GAIL)Get in touchPlease visit Des' research website for contact details: Des Higham | School of Mathematics Recent publications using AI techniques (open access links)A Survey of Inevitability Results in AI Instability, (to appear in Phil. Trans. Royal Society A, 2026)Deceptive Diffusion: Generating Synthetic Adversarial Examples | Proceedings Scale Space and Variational Methods in Computer Vision: 10th International Conference, 2025Stealth Edits for Provably Fixing or Attacking Large Language Models | Proceedings NeurIPS 2024The feasibility and inevitability of stealth attacks | IMA Journal of Applied Mathematics, 2024How Adversarial Attacks Can Disrupt Seemingly Stable Accurate Classifiers | Neural Networks, 2024 This article was published on Thursday 4 September 2025
