Speaker
Description
Detecting subtle new physics signals, such as those predicted by the Standard Model Effective Field Theory (SMEFT) with small Wilson coefficients, is inherently challenging when individual event-level kinematic differences are marginal. Since all collision events are governed by the same underlying physics parameters, we investigate the predictive power of permutation-invariant neural network architectures for set-level classification, where each input consists of a set of multiple collision events. This approach processes sets of multiple collision events collectively to distinguish between physics hypotheses (e.g., SM vs. SMEFT).
Preliminary results on simulated datasets show a dramatic improvement over standard event-level baselines, which typically perform near random chance for small Wilson coefficients. This gain arises from the model's ability to aggregate weak, but statistically consistent, kinematic alterations across multiple events within a set. While rigorous calibration and systematic uncertainty studies are ongoing, our findings suggests that set-based classification offers a powerful and scalable method for enhancing sensitivity in precision measurements and searches for new physics characterized by subtle kinematic shifts at the LHC.
| Would you like to be considered for an oral presentation? | Yes |
|---|
