Jet Classification Dataset with 100 Million Training Samples

JetClass contains 100 million jets for training, categorized into 10 different types. The dataset is designed for deep learning applications in jet tagging, distinguishing between background jets from light quarks or gluons and signal jets from particles like top quarks, W, or Z bosons.

Use Cases

• Train deep learning models for jet tagging using the 10 different jet types as classification targets.
• Develop algorithms to distinguish signal jets from background jets based on their initiating particles.
• Benchmark performance of neural networks on the large-scale classification of 100 million training jets.

Strengths

• Contains 100 million jets for training, providing a large-scale resource.
• Includes 10 distinct types of jets for classification tasks.
• Specifically designed to advance deep learning for jet tagging in high-energy physics.

Limitations

• The specific features (columns) describing each jet are not detailed in the provided input.
• The dataset's temporal and spatial coverage for the jet events is not specified.
• The distribution of samples across the 10 jet types is unknown, which could indicate class imbalance.

Provenance

Source

Author 'jet-universe' on Hugging Face.

Collection Method

Likely generated from simulated particle collision events, as referenced by its purpose for LHC research.

Freshness

Last updated on 2022-05-27.