Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Quantum artificial intelligence for pattern recognition at high-energy colliders: Tales of Three "Quantum's"

Hideki Okawa

TL;DR

The paper surveys the current landscape of quantum computing for pattern recognition at high-energy colliders, focusing on three technologies: quantum circuits, quantum annealing, and quantum-inspired algorithms. It explains how track and jet reconstruction tasks can be cast as Ising/QUBO problems and solved with approaches such as QAOA/VQE, D-Wave annealing, simulated annealing, and simulated bifurcation, among others. Key findings indicate that quantum-inspired methods deliver notable near-term speedups and that early quantum approaches can match classical baselines on simplified problems, but hardware limitations (qubit counts, connectivity, noise) hinder large-scale gains. The authors argue for a balanced, near-term-to-long-term strategy where quantum-inspired techniques provide practical gains now, while circuit- and annealing-based methods may realize substantial advantages as hardware scales and architectures (e.g., QuAM) mature, especially when integrated with HPC ecosystems.

Abstract

Quantum computing applications are an emerging field in high-energy physics. Its ambitious fusion with artificial intelligence is expected to deliver significant efficiency gains over existing methods and/or enable computation from a fundamentally different perspective. High-energy physics is a big data science that utilizes large-scale facilities, detectors, high-performance computing, and its worldwide networks. The experimental workflow consumes a significant amount of computing resources, and its annual cost will continue to grow exponentially at future colliders. In particular, pattern recognition is one of the most crucial and computationally intensive tasks. Three types of quantum computing technologies, i.e., quantum gates, quantum annealing, and quantum-inspired, are all actively investigated for high-energy physics applications, and each has its pros and cons. This article reviews the current status of quantum computing applications for pattern recognition at high-energy colliders.

Quantum artificial intelligence for pattern recognition at high-energy colliders: Tales of Three "Quantum's"

TL;DR

The paper surveys the current landscape of quantum computing for pattern recognition at high-energy colliders, focusing on three technologies: quantum circuits, quantum annealing, and quantum-inspired algorithms. It explains how track and jet reconstruction tasks can be cast as Ising/QUBO problems and solved with approaches such as QAOA/VQE, D-Wave annealing, simulated annealing, and simulated bifurcation, among others. Key findings indicate that quantum-inspired methods deliver notable near-term speedups and that early quantum approaches can match classical baselines on simplified problems, but hardware limitations (qubit counts, connectivity, noise) hinder large-scale gains. The authors argue for a balanced, near-term-to-long-term strategy where quantum-inspired techniques provide practical gains now, while circuit- and annealing-based methods may realize substantial advantages as hardware scales and architectures (e.g., QuAM) mature, especially when integrated with HPC ecosystems.

Abstract

Quantum computing applications are an emerging field in high-energy physics. Its ambitious fusion with artificial intelligence is expected to deliver significant efficiency gains over existing methods and/or enable computation from a fundamentally different perspective. High-energy physics is a big data science that utilizes large-scale facilities, detectors, high-performance computing, and its worldwide networks. The experimental workflow consumes a significant amount of computing resources, and its annual cost will continue to grow exponentially at future colliders. In particular, pattern recognition is one of the most crucial and computationally intensive tasks. Three types of quantum computing technologies, i.e., quantum gates, quantum annealing, and quantum-inspired, are all actively investigated for high-energy physics applications, and each has its pros and cons. This article reviews the current status of quantum computing applications for pattern recognition at high-energy colliders.

Paper Structure

This paper contains 20 sections, 23 equations, 12 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Quantum Computing
  3. Quantum circuits
  4. Quantum annealing
  5. Quantum-inspired
  6. Ising/QUBO Problems
  7. Solving Ising problems with quantum annealing
  8. Solving Ising problems with quantum circuits
  9. Solving Ising problems with quantum-inspired algorithms
  10. Simulated annealing (SA)
  11. Simulated bifurcation (SB)
  12. Track reconstruction
  13. Global track finding as an Ising problem
  14. Global track finding with neural networks or kernels
  15. Iterative quantum tracking
  16. ...and 5 more sections

Figures (12)

  • Figure 1: Three types of quantum computing technologies: (a) quantum circuits, (b) quantum annealing (From Ref. dwave-doc, https://creativecommons.org/licenses/by-nc-sa/4.0/), and (c) quantum-inspired (courtesy: Xianzhe Tao).
  • Figure 2: Display of reconstructed tracks in an event from the TrackML dataset Amrouche:2019wmxAmrouche:2021nbs, which is simulated under the HL-LHC conditions. The green (red) lines represent tracks that are correctly (incorrectly) reconstructed, whereas the blue lines are unreconstructed tracks. Reproduced from Ref. qaiatrack. https://creativecommons.org/licenses/by/4.0/.
  • Figure 3: Criteria for assigning QUBO quadratic coefficients depending on the configurations of triplet pairs. Reproduced from Ref. Bapst:2019llh. https://creativecommons.org/licenses/by/4.0/.
  • Figure 4: Schematic overview of tracking workflow using the QUBO formulation (b). Reproduced from Ref. Bapst:2019llh. https://creativecommons.org/licenses/by/4.0/.
  • Figure 5: Schematic diagram illustrating the sub-QUBO procedure. Reproduced from Ref. Crippa:2023ieq. https://creativecommons.org/licenses/by/4.0/.
  • ...and 7 more figures