After the discovery of the Higgs boson in 2012, finding evidence for physics beyond the Standard Model has been the primary objective of the experiments at the Large Hadron Collider. However, searches for new physics typically suffer from a trade-off between generality (“model independence”) and sensitivity. In this talk, I will present a new CMS result that utilizes novel analysis strategies based on unsupervised and weakly supervised machine learning to minimize this trade-off, i.e., to maximize the search sensitivity while at the same time casting a wide net to catch exotic signatures. Specifically, the search is focusing on anomalous hadronic activity clustered into a dijet system, which is one of the main channels where new physics might appear. This result opens up a new direction for highly sensitive yet generic searches for new physics and bolsters the development of new algorithmic techniques.

After stations at KIT, CERN, and MIT, Benedikt Maier is currently a Research Fellow at Imperial College London. He focuses on searches for new physics, specifically coming from a dark sector, and on the development and application of novel machine learning.