Qutrits for physics at the LHC

Basic Information

• Paper ID: 2510.14001
• Title: Qutrits for physics at the LHC
• Authors: Miranda Carou Laiño, Veronika Chobanova, Miriam Lucio Martínez
• Classification: quant-ph (Quantum Physics)
• Publication Date: October 17, 2025
• Paper Link: https://arxiv.org/abs/2510.14001

Abstract

This paper explores the application of quantum machine learning models based on three-level quantum systems (qutrits) for anomaly detection in high-energy physics data, with particular focus on Large Hadron Collider (LHC) experiments. The study proposes a qutrit-based quantum model and benchmarks it against qubit-based approaches, evaluating performance in terms of accuracy, scalability, and computational efficiency. The research aims to determine whether qutrit architectures can provide advantages in addressing the computational and analytical demands of future collider experiments.

Research Background and Motivation

Problem Definition

1. Data Processing Challenges: Next-generation particle colliders (such as the High-Luminosity LHC) will generate unprecedented data volumes, requiring three times the real-time data processing compared to LHC Run 3
2. Anomaly Detection Requirements: Identifying anomalous events that cannot be explained by the Standard Model of particle physics and discovering exotic physical phenomena
3. Computational Resource Constraints: Traditional machine learning methods face limitations when processing exabyte-scale datasets

Research Motivation

• Qubit Limitations: Conventional binary quantum bits have limited expressive power and resource efficiency
• Qutrit Advantages: Three-level quantum systems provide larger Hilbert space with enhanced expressive capability and more compact variational circuits
• Encoding Efficiency: Qutrits can store log₂3 ≈ 1.56 bits of information, a significant improvement over the 1 bit capacity of qubits

Core Contributions

1. First Application of Qutrits to High-Energy Physics: Developed qutrit-based quantum autoencoders (QAE) for LHC data anomaly detection
2. Majorana Representation Implementation: Successfully implemented qutrit geometric representation on the unit sphere using Majorana notation
3. Novel Quantum Gate Design: Designed rotation gates and SWAP gates for qutrits based on Gell-Mann matrices
4. Innovative Encoding Scheme: Proposed "One Particle - One Qutrit" encoding scheme capable of encoding more particle parameters
5. Performance Benchmarking: Conducted comprehensive performance comparison analysis with qubit methods

Methodology Details

Task Definition

Input: Kinematic parameters of particle collision data (transverse momentum pT, azimuthal angle φ, pseudorapidity η, etc.) Output: Anomaly detection results (distinguishing signal from background through quantum fidelity distribution) Objective: Identify anomalies of physics signals such as W→qq̄, H→bb̄, t→bqq̄ relative to QCD background

Model Architecture

1. Qutrit Fundamental Theory

• State Representation: Qutrit states spanned by three orthogonal basis states {|0⟩, |1⟩, |2⟩}
• Gell-Mann Matrices: Eight Gell-Mann matrices λᵢ as generators of the SU(3) group
• Density Matrix: ρ(n⃗) = 1/3(I₃ₓ₃ + √3 n⃗·λ⃗)

2. Majorana Representation

A pure qutrit state is completely characterized by two points P₁(θ₁,φ₁) and P₂(θ₂,φ₂) on the unit sphere:

|ψg⟩ = Γ√2cos(θ₁/2)cos(θ₂/2), e^(iφ₁)sin(θ₁/2)cos(θ₂/2) + e^(iφ₂)cos(θ₁/2)sin(θ₂/2), √2e^(i(φ₁+φ₂))sin(θ₁/2)sin(θ₂/2)ᵀ

where the normalization factor is Γ = √23 + cosθ₁cosθ₂ + sinθ₁sinθ₂cos(φ₁-φ₂)^(-1/2)

3. Quantum Gate Design

• Rotation Gates: Based on SO(3) generators Σᵢ, Rᵢ(ξ) = e^(iξΣᵢ) = I + (cosξ-1)Σᵢ² + isinξΣᵢ
• Chrestenson Gate: Qutrit version of the Hadamard gate
• Generalized SWAP Gate: 9×9 matrix implementing state exchange for qutrit pairs

4. Quantum Autoencoder Structure

• Encoder: Uses variational quantum circuit U(Θ) to compress input data into latent representation
• Decoder: Reconstructs input through the Hermitian conjugate of the encoding operator
• Loss Function: Defined as negative fidelity between reference state and garbage state

Technical Innovations

1. Multi-Parameter Encoding: Compared to qubit's 2-parameter (η,φ) encoding, qutrits can encode 4 parameters including jet mass, energy, longitudinal and transverse impact parameters
2. Geometric Intuitiveness: Majorana representation provides clear geometric distinction between pure and mixed states
3. Circuit Depth Optimization: Higher information density reduces required circuit depth and decoherence accumulation

Experimental Setup

Dataset

1. Training Data: Real data collected by the CMS detector in 2016, primarily QCD jets (contamination rate <1%)
2. Test Data: JetClass simulation dataset containing 125 million jets, divided into 10 categories
3. Signal Types: Jets produced from W boson, Higgs boson, and top quark decays

Evaluation Metrics

• AUC Score: Used to evaluate the model's ability to distinguish signal from background
• Quantum Fidelity Distribution: Analyzes model compression and reconstruction capability
• Anomaly Detection Performance: Distinguishes anomalous events through fidelity threshold

Comparison Methods

• Reference Model: Qubit-based quantum autoencoder (from literature 15)
• Reproduced Model: Qubit version implemented by the authors as baseline

Implementation Details

• Simulation Platform: PennyLane quantum machine learning library
• Circuit Limitations: Due to PennyLane constraints, maximum support for 8 qutrits
• Optimizer: Automatic differentiation and differentiable workflow

Experimental Results

Main Results

Key Findings

1. Performance Equivalence: Qutrit models achieved AUC performance comparable to qubit models
2. Signal Discrimination Capability: Qutrit models demonstrated better signal discrimination ability in fidelity distributions
3. Physical Consistency: All models correctly identified t→bqq̄ as the most anomalous signal, consistent with physical expectations
4. Encoding Advantages: Qutrits can encode more information at the same circuit scale

Ablation Studies

Tested encoding effectiveness with different parameter combinations:

• Optimal Combination A: φ₁=mass-related parameter (ς), φ₂=energy-related parameter (ε)
• Combination B: φ₁=longitudinal impact parameter (ρ₀), φ₂=transverse impact parameter (ρz)

Model Validation

Verified through analytical calculations:

1. Correctness of Majorana encoding
2. Mathematical implementation of rotation gates
3. Overall robustness of quantum circuits

Related Work

Quantum Machine Learning in High-Energy Physics

• Traditional QML primarily based on qubit systems
• Existing work focused on classification and regression tasks
• Anomaly detection applications relatively limited

Qudit Quantum Computing

• Theoretical advantages widely recognized
• Hardware implementation still in early stages
• Shows potential in quantum error correction

Innovation of This Work

• First application of qutrit systems to high-energy physics data analysis
• Provides complete framework from theory to implementation
• Establishes direct performance comparison between qutrit and qubit systems

Conclusions and Discussion

Main Conclusions

1. Feasibility Proof: Qutrit systems are feasible for anomaly detection in high-energy physics
2. Performance Equivalence: Under current hardware constraints, qutrits achieve performance comparable to qubits
3. Potential Demonstration: Higher expressive power and better signal discrimination indicate future advantages

Limitations

1. Hardware Constraints: Current quantum simulators have limited qutrit support (maximum 8 qutrits)
2. Circuit Scale: Limited by PennyLane framework, unable to test large-scale circuits
3. Dataset Scale: Test scale relatively small compared to complete LHC data

Future Directions

1. Hardware Development: Awaiting more mature qutrit quantum hardware
2. Algorithm Optimization: Exploring qutrit mixed states and higher-dimensional qudit systems
3. Application Extension: Testing data from other LHC experiments (ATLAS, LHCb)
4. New Physics Discovery: Applications in searching for physics beyond the Standard Model

In-Depth Evaluation

Strengths

1. Strong Innovation: First introduction of qutrit systems to high-energy physics data analysis
2. Theoretical Completeness: Provides complete framework from mathematical foundations to practical implementation
3. Sufficient Verification: Dual verification through analytical calculation and numerical simulation
4. Good Practicality: Uses real LHC data for testing with actual application value

Weaknesses

1. Hardware Dependency: Limited by current quantum simulator capabilities, unable to fully demonstrate qutrit potential
2. Scale Limitations: Relatively small circuit scale, may not reflect real performance of large-scale applications
3. Limited Comparison: Lacks direct comparison with classical machine learning methods
4. Theoretical Analysis: Could provide deeper theoretical analysis of qutrit advantages

Impact

1. Academic Value: Opens new application direction for qutrit quantum computing in high-energy physics
2. Practical Prospects: Provides new insights for future High-Luminosity LHC data processing
3. Technology Advancement: May promote development needs for qutrit quantum hardware

Applicable Scenarios

1. Anomaly Detection: Applicable to various high-energy physics experiments requiring anomalous pattern identification
2. Data Compression: Broad application potential in quantum data encoding and compression
3. Pattern Recognition: Extensible to other scientific computing domains requiring complex pattern recognition

References

This paper cites 31 important references covering cutting-edge research in quantum computing, high-energy physics, machine learning and other fields, providing a solid theoretical foundation for interdisciplinary research.

Overall Assessment: This is a groundbreaking interdisciplinary research paper that successfully introduces qutrit quantum computing technology to high-energy physics data analysis. Although constrained by current hardware limitations, the research demonstrates potential advantages of qutrit systems and establishes an important foundation for future quantum-enhanced high-energy physics data analysis.