Parton-to-Jet Mapping proton-proton dataset at 14 TeV

This dataset contains simulated proton-proton collision events at $\sqrt s = 14$ TeV simulated with Pythia8. It is intended originally for quantum machine learning tasks that aim to learn mappings between parton-level data (before QCD showering and hadronisation) and jet-level observables (as seen in the detector). Contents - Parton-Level Data: Includes 4-vector components ($E$, $p_x$, $p_y$, $p_z$), particle ID, and charge for two initial partons per event. These represent theoretical pre-hadronisation inputs. - Jet-Level Data: Includes reconstructed jet 4-vectors (up to 5 jets per event, zero-padded), derived using the Anti-$k_t$ algorithm with $R=0.4$. Simulation Details The dataset is generated using the `Pythia8` Monte Carlo event generator, simulating the process $qg \rightarrow Zq$ with $Z \rightarrow q\bar{q}$ The invariant mass and transverse momentum of the hard process are required to exceed 50 GeV. The final-state hadrons are clustered into jets after hadronisation using the Anti-$k_t$ algorithm with $R=0.4$, while multiple parton interactions are disabled to maintain clarity of the underlying parton-to-jet relationship. This dataset is especially suited for benchmarking classical and quantum machine learning models on a physically rich, structured regression task. It enables exploration of how well models can reconstruct observable final states from incomplete and high-level theoretical inputs. See Aqora LHC challenge.