Source-linked AI summary
Dijet resonance search with weak supervision using $\sqrt{s}=13$ TeV $pp$ collisions in the ATLAS detector
ATLAS Collaboration
TL;DR
The paper addresses the limited coverage of generic searches for dijet resonances, especially when all particles in A→BC may be beyond the Standard Model. It uses weakly supervised classifiers trained directly on collision data and finds sensitivity across the three-dimensional mass space, with limits weaker for some low-signal-to-background scenarios.
Problem
Many Standard Model and beyond-Standard-Model dijet topologies are not covered by dedicated searches, motivating a generic search that avoids a large statistical trials factor.
Method
Weakly supervised neural networks distinguish each dijet-mass signal region from neighboring sidebands using only the two leading-jet masses, with cross-validation preventing events from training the network applied to them.
Results
For m_B=m_C=400 GeV, the excluded W′ production cross section is about 1 fb, while sensitivity varies across m_A, m_B, and m_C.
Takeaways & Limitations
The search is more sensitive than the inclusive dijet search for jets from Lorentz-boosted heavy-particle decays and extends all-hadronic diboson coverage away from Standard Model boson masses.
Takeaways & Limitations
The neural network cannot effectively identify some models, such as (m_A,m_B,m_C)=(5000,80,80) GeV, producing weaker limits than previous searches.
Abstract
from arXiv · showhide
This Letter describes a search for narrowly resonant new physics using a machine-learning anomaly detection procedure that does not rely on a signal simulations for developing the analysis selection. Weakly supervised learning is used to train classifiers directly on data to enhance potential signals. The targeted topology is dijet events and the features used for machine learning are the masses of the two jets. The resulting analysis is essentially a three-dimensional search $A\rightarrow BC$, for $m_A\sim\mathcal{O}(\text{TeV})$, $m_B,m_C\sim\mathcal{O}(100\text{ GeV})$ and $B,C$ are reconstructed as large-radius jets, without paying a penalty associated with a large trials factor in the scan of the masses of the two jets. The full Run 2 $\sqrt{s}=13$ TeV $pp$ collision data set of 139 fb$^{-1}$ recorded by the ATLAS detector at the Large Hadron Collider is used for the search. There is no significant evidence of a localized excess in the dijet invariant mass spectrum between 1.8 and 8.2 TeV. Cross-section limits for narrow-width $A$, $B$, and $C$ particles vary with $m_A$, $m_B$, and $m_C$. For example, when $m_A=3$ TeV and $m_B\gtrsim 200$ GeV, a production cross section between 1 and 5 fb is excluded at 95% confidence level, depending on $m_C$. For certain masses, these limits are up to 10 times more sensitive than those obtained by the inclusive dijet search. These results are complementary to the dedicated searches for the case that $B$ and $C$ are Standard Model bosons.