A warm welcome to our new PhD students, Andjela Todorovic and Vincent Zhang, who join CRUMBS this October!
Can Reasoning Unveil Misaligned BehaviourS?
CRUMBS is a research group in the Department of Computing at Imperial College London. Our goal is to develop methods to make AI transparent and safe, with a focus on explainability across learning and symbolic systems.
Latest news
All news →Our paper “Efficient bias mitigation in T2I diffusion models using Concept Graphs” has been accepted at NeurIPS 2026! Congratulations to Mansi, first author of the paper.
Our paper “Enhancing the Robustness of Counterfactual Explanations via Adversarial Training” has been accepted at SAIV 2026!
Francesco has been selected for the IJCAI-ECAI’26 Early Career Spotlight!
