Machine learning–driven anomaly detection has rapidly emerged as a powerful tool to enhance the traditional search strategies used in ATLAS and CMS. By harnessing recent progress in unsupervised learning, these techniques boost sensitivity to potential new physics in a model-independent way - broadening coverage across diverse final states while requiring fewer resources. In this seminar, I will outline several of the common strategies for anomaly detection before highlighting the areas that we are working on at IFAE.
Giada Caneva, Elia Bertoldo, Francesco Sciotti