Calendar of Events

Speaker: Benjamin Nachman (Lawrence Berkeley National Laboratory)

Despite extensive theoretical motivation for physics beyond the Standard Model (BSM) of particle physics, searches at the Large Hadron Collider (LHC) and elsewhere have found no convincing evidence for BSM physics. Therefore, it is essential to broaden the sensitivity of the search program to include unexpected scenarios. Machine Learning (ML) gives us hyper-variate vision to identify BSM signatures in ways that were not previously possible. However, traditional ML techniques require us to provide labeled training data in order to teach them how to look for something “new”. I will introduce an innovative and exciting research direction in semi-supervised learning (“weak supervision”) where ML techniques can be applied to unlabeled data in order to teach us how to look for something genuinely new (i.e. unexpected). To illustrate the power of these methods, I will use examples from the LHC, though many of the techniques likely have broader applicability.