Rate My Professor James Smith

Smith's academic interests encompass interactive artificial intelligence, machine learning, evolutionary computation, machine vision, and statistical disclosure control, particularly privacy leakage from AI models and preservation of privacy when using sensitive data like health records for public good research. He directs several funded initiatives, including UKRI-supported GRAIMATTER, SACRO, SDC-Reboot, and TREvolution projects within the DARE UK programme; an Innovate UK collaboration with Lyons Davidson Ltd. on interactive machine learning for predicting insurance claim settlement values; and a project with Ribbon Communications Ltd. on detecting security threats in large-scale mobile networks. Key publications include the second edition of the textbook Evolutionary Computation, co-authored with Gusz Eiben in 2015, and the article 'From evolutionary computation to the evolution of things' in Nature (2015). He currently supervises PhD students on convolutional neural architectures for air quality modelling, interactive machine learning for network security, and data privacy in federated learning. His work bridges theoretical AI advancements with practical applications in industry and government.