Kieran Holland
University of the Pacific, CA·Physics
The strong nuclear force binds quarks and gluons into protons, neutrons and atomic nuclei, and its mathematical description is quantum chromodynamics (QCD). Unlike many physical theories, QCD usually cannot be solved with pencil and paper; predictions come from large-scale computer simulations in which space and time are replaced by a grid, or lattice, and quantities are computed with Monte Carlo methods that sample random configurations of the fields.
- Fields
Physics
Computer Science
- Topics
lattice QCD
quantum chromodynamics
Monte Carlo simulation
- Grants
- 1 known grant, through Aug 2027
What they work on
- Use machine learning algorithms and software platforms to parametrize a highly improved lattice action for QCD, the theory of quarks and gluons, so that accurate predictions can be obtained on coarse space-time grids.
- Test the machine-learning-designed action in Monte Carlo lattice simulations and compare its accuracy and computational cost against existing lattice actions.
- Apply the approach first to pure SU(3) Yang-Mills theory, which describes gluon dynamics without quarks, as a cheaper testbed and as a route to predicting the QCD coupling strength at the electroweak energy scale.
More about the research
The accuracy of such a simulation depends on how fine the grid is, much as image quality depends on pixel size — but computing cost climbs steeply as the grid gets finer, quickly outrunning even the largest supercomputers. This project pursues the alternative: designing a smarter description of the theory on a coarse grid, known as an improved lattice action, so that correct physical predictions can be obtained without brute-force resolution.
Why it matters
Lattice QCD simulations consume large shares of national supercomputing resources, so actions that deliver accurate results on coarse grids would let many research groups obtain more precise predictions about quarks, gluons and nuclei at lower cost. The work also tests how far machine learning can be pushed as a design tool in theoretical physics, and would give undergraduates training in scientific computing applicable to industry or graduate study.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Andrea Shindler
University of California-Berkeley, CA·Physics
Calculate the electric dipole moments of protons and neutrons induced by theories beyond the Standard Model, to probe the source of the universe's matter-antimatter asymmetry.
Huey-Wen Lin
Michigan State University, MI·Physics
Compute parton distribution functions of nucleons and mesons with lattice QCD at high precision, varying lattice spacing and volume and working at the physical pion mass to control errors from the simulation.
Nikolaos Kidonakis
Kennesaw State University, GA·Physics
Calculate soft-gluon corrections, from low-energy gluon radiation, to top-antitop pair, single-top, and top-pair-plus-Z or Higgs production at NNLO and N3LO, including differential distributions for LHC and future colliders.
The school
University of the Pacific: 1 labs on PI Pages.
Reach out
A start for your email. Fill in the brackets.