Designed to address the 'black box' problem, this volume unpacks attention visualization, post-hoc explanation pipelines, and auditable algorithmic frameworks necessary for deploying AI systems in highly regulated legal, financial, and medical industries.
Prof. Thorne is the lead researcher at the Trustworthy Machine Intelligence Lab and advises international bodies on algorithmic fairness and
transparency policies.
A mathematical and structural look inside deep neural networks, detailing methodologies like SHAP, LIME, and layer-wise relevance propagation to make complex models transparent.