A Deep Q-Network using matrix-element rewards reconstructs parton assignments in collider events, enabling theory-based tagging and anomaly detection without labels.
Finding physics signals with event deconstruction
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
We develop a matrix element based reconstruction method called event deconstruction. The method uses information from the hard matrix element and a parton shower to assign probabilities to whether a final state was initiated by a signal or background process. We apply this method in the signal process of a Z' decaying to boosted top quarks in an all hadronic final state and discuss envisioned improvements of the method. We find that event deconstruction can considerably improve on existing reconstruction techniques.
citation-role summary
background 1
citation-polarity summary
fields
hep-ph 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Theory-informed neural networks for particle physics
A Deep Q-Network using matrix-element rewards reconstructs parton assignments in collider events, enabling theory-based tagging and anomaly detection without labels.