
SIGMA: Self-Guided Integrated Gradient Method for Attribution
A feature attribution method that follows a models decision boundary to explain deep neural network predictions.

Hi, I'm
PhD Researcher | Explainable AI
PhD researcher in the Quantum Optics & Computational Imaging lab at Heriot-Watt University, affiliated with the national quantum hubs in sensing and imaging and position, navigation and timing. Specialising in explainable and interpretable AI for high-stakes applications including healthcare, security, and autonomous vehicles.
A selection of my recent work

A feature attribution method that follows a models decision boundary to explain deep neural network predictions.
Self-Guided Integrated Gradient Method for Attribution
S.Henry, A. Ruget, S. Scholes, J. Leach
CVPR 2026 — Findings Track, Denver, CO, June 2026
Preprint · Code
A Self-Guided Approach to Integrated Gradients
S.Henry
Turing Connections Day, Alan Turing Institute, Edinburgh, March 2026
Attribution Analysis for Interpretable Machine Learning
S. Henry,
AI for Science Conference, The Royal Society, London, March 2026
Heriot-Watt University
Rank Prize
Heriot-Watt University
Heriot-Watt University
I'm always interested in hearing about new projects and opportunities. Feel free to reach out!