LHC Jet Data for High-Level Feature Machine Learning

High-level features from simulated LHC jet events, likely used for training machine learning models in particle physics. The dataset originates from the hls4ml project, which focuses on deploying neural networks on FPGAs for fast inference. Specific details on the number of events, features, and creation date are not provided in the metadata.

Use Cases

• Classifying jet types (e.g., quark vs. gluon) using high-level kinematic features like jet mass and transverse momentum.
• Training regression models to predict jet energy or angular resolution from reconstructed high-level features.
• Benchmarking neural network architectures for FPGA deployment via the hls4ml framework on physics-specific tabular data.
• Studying feature importance for jet substructure analysis using engineered high-level observables.

Strengths

• Data is derived from simulated LHC collisions, providing a controlled environment for algorithm development.
• Designed for the hls4ml framework, indicating relevance for high-performance, low-latency inference applications.

Limitations

• The exact number of jet events and the specific set of high-level features are unknown.
• Contains only simulated data, which may not capture all complexities of real detector effects and noise.

Provenance

Source

hls4ml project, likely using simulation frameworks like GEANT4 or Delphes.

Collection Method

Generated from Monte Carlo simulations of proton-proton collisions at the LHC.