Skip to content
Back to search results

Anti-kT Jet Algorithm

Introduced by
Matteo Cacciari and Gavin P. Salam
Year
2008
Type
Particle clustering algorithm
Family
Jet clustering
Font size

What does it do?

The anti-kT jet algorithm, introduced by Cacciari and Salam in 2008, is a sequential recombination jet clustering algorithm widely used in high-energy physics to group final-state particles into jets.

Unlike earlier algorithms, anti-kT produces jets with regular cone-like geometries in transverse momentum-rapidity space, making it ideal for precision measurements and new physics searches.

The intuition behind the method

The anti-kT algorithm builds jets by repeatedly combining pairs of particles or existing jets based on a distance metric. The key insight is that it combines soft (low-energy) particles with nearby hard (high-energy) particles preferentially, creating cone-shaped jets around hard cores. This 'anti-fragmentation' property contrasts with earlier kT-algorithms, which combine soft particles with each other first. The result is cleaner jet isolation and reduced sensitivity to underlying event contamination, crucial for accurate physics measurements.

History

The anti-kT algorithm was proposed in 2008 by Cacciari and Salam as an improvement over the original kT-algorithm and cone-based clustering methods. It was immediately adopted by ATLAS and CMS at the LHC as the standard jet reconstruction algorithm due to its superior properties for precision measurements. The FastJet software library, released in 2008, provides efficient implementations of anti-kT and other algorithms. Since then, variants and extensions like variable-R anti-kT and substructure-aware clustering have been developed.

When to use it

Use anti-kT for any jet reconstruction in collider experiments, particularly when precision is needed or when distinguishing signal jets from background. It is the standard algorithm at the Large Hadron Collider. The algorithm is also suited for identifying boosted objects (high-momentum W bosons, top quarks, Higgs bosons) and studying jet substructure with grooming techniques. Avoid if you need variable-R clustering or adaptive cone sizes.

Limitations and considerations

  • Fixed jet radius parameter may be suboptimal for all physics processes; jets can include overlapping contributions from unrelated hard processes
  • Dependence on underlying event modeling at high pile-up (many collisions per event), requiring careful calibration
  • Grooming or filtering techniques are needed to study substructure within jets; the algorithm alone does not separate substructure
  • Definition of jet boundary can be ambiguous in regions with gradual transitions between soft and hard activity

Source conditions do not replace independent scientific review or validation by an analysis engine.

Strengths

  • Produces regular cone-shaped jets in rapidity-azimuth space, simplifying theoretical calculations and experimental corrections
  • Preferentially combines soft particles with hard cores, resulting in jets with well-defined boundaries and low contamination from underlying event
  • Infrared and collinear safe, enabling reliable calculations at all orders of perturbative QCD
  • Computationally efficient with linear scaling in the number of particles when using FastJet
  • Widely used by LHC experiments, ensuring robust validation and well-understood systematic uncertainties

Applications

  • Standard jet reconstruction at the Large Hadron Collider for all physics analyses (ATLAS, CMS, LHCb)
  • Boosted Higgs boson reconstruction using jet substructure and anti-kT grooming techniques
  • Top quark pair production studies where jets from top decays are reconstructed and characterized
  • New physics searches including large-radius jets for resonance decays into multiple partons
  • B-physics analyses at LHCb using anti-kT clustering for suppressing pile-up from proton-proton collisions

How it works

How it works

1

Read in all final-state particles from the collision event, including muons, electrons, and neutral particles, with their transverse momenta and rapidity coordinates.

2

Compute the distance metric between each pair of particles and between each particle and the beam axis, using the anti-kT distance measure.

3

Identify the minimum distance pair and combine them into a new cluster if the distance is smaller than the beam distance; otherwise, remove the particle as a complete jet.

4

Repeat the distance calculation and combination procedure with the new cluster until all particles are either merged or promoted to jets.

5

Extract final jet four-momenta by summing the constituents and classify jets by their hardness, isolation, and substructure for physics analysis.

Common pitfalls

  • Neglecting to properly calibrate jet energy scale for experimental resolution and detector effects
  • Using fixed jet radius without considering the physics process; boosted objects may require smaller radii
  • Forgetting that jets are experimental constructs; comparing only at the jet level without accounting for migration and smearing
  • Mixing results from different jet algorithms or radii in the same analysis without accounting for differences in jet composition

Frequently asked questions

What is the difference between anti-kT and the original kT algorithm?

The kT algorithm combines soft particles with each other first, leading to irregular jet boundaries and sensitivity to underlying event. Anti-kT reverses this by combining soft particles with nearby hard ones, producing cone-shaped jets with well-defined boundaries.

How do I choose the jet radius parameter?

Common choices are R=0.4 for standard physics (already resolved), R=0.8 for boosted objects, and R=1.0 for high-mass resonances. Smaller radii reduce pile-up contamination but increase sensitivity to soft radiation; larger radii capture more physics but can merge distinct processes.

What is infrared-collinear safety?

An algorithm is infrared-collinear safe if the jet content is insensitive to soft radiation emission or collinear splitting of quarks/gluons. This ensures reliable perturbative QCD calculations at all orders without large uncertainties from missing higher-order terms.

Can I use anti-kT on experimental data with detector effects?

Yes, but you must calibrate the jet energy scale to account for detector resolution and efficiency. Unfolding techniques correct for migration between true and reconstructed jets, enabling precision measurements.

Related methods

Sources

  1. Cacciari, M., Salam, G. P., & Sapeta, S. (2008). On the characterisation of the underlying event. Journal of High Energy Physics, 2008(04), 063.

    Open publication reference
    Copied.
  2. Ellis, S. D., Vermilion, C. K., & Walsh, J. R. (2010). Recombination algorithms for jet substructure. Physical Review D, 81(9), 094023.

    Open publication reference
    Copied.
  3. Cacciari, M., & Salam, G. P. (2008). FastJet user manual. The European Physical Journal C, 72(3), 1896.

    Open publication reference
    Copied.