Kostant Relation in Filtered Randomized Benchmarking for Passive Bosonic Devices

Basic Information

• Paper ID: 2511.00842
• Title: Kostant relation in filtered randomized benchmarking for passive bosonic devices
• Author: David Amaro-Alcalá (Institute of Physics, Slovak Academy of Sciences)
• Classification: quant-ph (Quantum Physics)
• Publication Date: November 4, 2025
• Paper Link: https://arxiv.org/abs/2511.00842

Abstract

This paper proposes an improved scheme for bosonic randomized benchmarking (BRB) that significantly reduces computational and experimental costs. The core contributions include: (1) introducing immanant-based filtering functions that avoid computing Clebsch-Gordan coefficients; (2) proposing a data collection scheme requiring only a single measurement type; (3) demonstrating that weak coherent states and intensity measurements are sufficient for characterization. These improvements enable simpler experimental platforms to perform quantum device characterization and substantially simplify data analysis procedures.

Research Background and Motivation

Problem Statement

Characterizing passive bosonic devices is an important step in developing continuous-variable quantum computing. While existing bosonic randomized benchmarking schemes inherit advantages of standard randomized benchmarking (such as robustness to state preparation and measurement errors), they suffer from two critical limitations:

1. Computational Complexity: Requires computing matrix permanents, which is computationally difficult (#P-complete problem). Moreover, the required permanents must be individually determined through complex decompositions involving Clebsch-Gordan coefficients.
2. Experimental Difficulty: Requires preparation of Fock states and use of photon-number-resolving detectors, which present challenges for most laboratories.

Importance of the Problem

Continuous-variable quantum computing represents an important paradigm in quantum computation, with passive bosonic devices (such as linear optical interferometers) serving as core components. Accurate characterization of these devices' performance is crucial for:

• Quality control of quantum computers
• Understanding and mitigating noise effects
• Verifying quantum advantage

Limitations of Existing Methods

The main limitations of the original scheme 11 include:

• Requires computing multiple matrix permanents (at least ♯λ - 1 + dλ)
• Must compute SU(m) Clebsch-Gordan coefficients
• Requires complex experimental apparatus (Fock state preparation and photon-number resolution)
• Complex data analysis dependent on specific initial states and measurements

Research Motivation

This paper aims to maintain the advantages of the original scheme while fundamentally simplifying computational and experimental requirements by leveraging the Kostant relation (a mathematical relationship connecting immanants with zero-weight states).

Core Contributions

1. Proposes New Immanant-Based Filtering Functions: Through the Kostant relation, the filtering process is expressed as computation of immanants, completely avoiding Clebsch-Gordan coefficient calculations.
2. Reduces Computational Complexity:
   • Reduces required permanents from ♯λ - 1 + dλ to only 1
   • Required immanants number ♯λ - 1 (typically much smaller than the original scheme)
   • Eliminates the need for Clebsch-Gordan coefficient computation
3. Simplifies Experimental Requirements: Demonstrates that weak coherent states and intensity measurements are sufficient for characterization, eliminating the need for photon-number-resolving detectors or complex Fock state preparation.
4. Provides Theoretical Guarantees: Proves that the new filtering functions still produce single exponential decay, enabling accurate estimation of fidelity metrics F(E).
5. Enhances Robustness: The method exhibits natural robustness to photon loss and gain errors; filtering expressions can be predetermined and do not depend on initial state or measurement choices.

Methodology Details

Task Definition

Input:

• n photons entering an m-port interferometer
• K randomly sampled unitary gate sequences Ug_s of depth g
• Initial state ϱ and measurement E (can be weak coherent states and intensity measurements)

Output:

• Fidelity metric of the noisy channel E: F(E) = d_λ^(-2) Σ_μ d_μ p_μ(E)

Constraints:

• Assumes gate-independent, time-independent Markovian noise model: Ũ = E ∘ U

Model Architecture

1. Mathematical Framework

The system's Hilbert space is H_m^n, representing the state space of n indistinguishable photons in m modes. Unitary operations U ∈ SU(m) act as:

U(|n⟩) = (∏_k 1/√(n_k!) (U(a_k†))^(n_k)) |0⟩

Under vectorization, the action of unitary operations is:

Γ: U ↦ U ⊗ Ū

2. Representation Theory Decomposition

The tensor product of the symmetric irreducible representation λ = (n, 0, ..., 0) and its dual λ* decomposes as:

Γ := λ ⊗ λ* ≅ ⊕_μ μ

The representations μ are computed via the Pieri rule: adding n boxes to different columns of the Young diagram of λ*.

3. Core Innovation: Immanant Filtering Functions

Theorem 2 (Immanant Filtering Functions): Define the filtering function as:

f_(g,s)_Imm,μ(U^(g,s)) := Imm_μ(U^(g,s))

Then:

Φ_g^(f) := E_s [f_μ^(g,s)(U^(g,s)) d_ϱ,E^(g,s)(Ũ^(g,s))] = κ p_μ^(g-1)

where κ is a constant independent of the characterization.

4. Application of the Kostant Relation

Theorem 1 (Kostant Relation):

Σ_(|ζ_κ⟩∈Z_κ) ⟨ζ_κ|Γ(U)|ζ_κ⟩ = Imm_κ(U)

This relation connects immanants with traces over zero-weight states and forms the mathematical foundation of this paper's approach.

Technical Innovations

1. Elimination of Projection Operators

The original scheme's filtering function is:

f_λ,orig := ⟨ϱ|P_λ S^+ Γ(U^g_s)† |E⟩

This requires computing projection operators P_μ onto irreducible representation subspaces, necessitating Clebsch-Gordan coefficients.

The proposed scheme directly uses immanants:

f_Imm,μ := Imm_μ(U^(g,s))

This completely avoids projection operators and Clebsch-Gordan coefficients.

2. Exploitation of Zero-Weight State Structure

Through the Kostant relation, immanants can be expressed as sums over zero-weight states:

Imm_μ(U) = Σ_(|ζ_μ⟩∈Z_μ) ⟨ζ_μ|Γ(U)|ζ_μ⟩

The number of zero-weight states is much smaller than a complete basis, and their structure is simple and easy to compute.

3. Guarantee of Single Exponential Decay

Key proof steps:

• Average over zero-weight states
• Utilize properties of the twirling operation TE
• Each p_μ decays independently due to orthogonality of different irreducible representations

Mathematically:

⟨ζ_μ^(i)|S_Immμ^(i) T[E]^g |ϱ̃⟩ = (Σ_j s_(i,j)⟨ζ_μ^(j)|ϱ̃⟩) p_μ^g

Data Processing Pipeline

1. Data Collection: Construct data matrix D where D_(g,s) = ⟨Ẽ|Γ(Ũ^(g,s))|ϱ̃⟩
2. Immanant Computation: Construct filtering matrix F_μ where (F_μ)_(g,s) = Imm_μ(U^(g,s))
3. Hadamard Product: Compute Φ_g = Σ_s (F_μ ⊙ D)_(g,s)
4. Exponential Fitting: Fit exponential function to {g, Φ_g} to extract p_μ
5. Fidelity Calculation: Use F(E) = d_λ^(-2) Σ_μ d_μ p_μ

Experimental Setup

Theoretical Verification

This paper is primarily theoretical and methodological, verifying method correctness through mathematical proofs and concrete examples.

Concrete Example: Verification of Kostant Relation for SU(3)

System Parameters:

• Unitary group: SU(3)
• Irreducible representation: μ = (2,1,0)
• Zero-weight states:

|ζ_(2,1,0)^(0)⟩ = |2 1 0; 1 1; 1⟩
|ζ_(2,1,0)^(1)⟩ = |2 1 0; 2 0; 1⟩

Verification Process:

1. Compute D-functions using Gelfand-Tsetlin patterns
2. Calculate diagonal matrix elements of zero-weight states (equations A2a and A3)
3. Compute immanant using character tables (equation A4)
4. Verify: Σ_(i=0,1) ⟨ζ^(i)|Γ(U)|ζ^(i)⟩ = Imm_(2,1,0)(U)

Results show complete agreement between the two computational methods, verifying the Kostant relation.

Feasibility of Weak Coherent States and Intensity Measurements

Extended Hilbert Space: Consider (H_e)m^n := ⊕(n≥0) H_m^n, allowing photon number variation.

Key Observation: For weak coherent state input:

ϱ = |0,0⟩ + α|1,0⟩

In the extended space representation theory decomposition:

(0 ⊕ λ) ⊗ (0 ⊕ λ*) = 0 ⊕ λ ⊕ λ ⊕ λ* ⊕ Γ

The target irreducible representation still appears only once, so the filtering process remains valid.

Experimental Advantages:

• Weak coherent states are easier to prepare and have higher frequency
• Intensity measurements are much simpler than photon-number-resolving detectors
• Natural robustness to loss and gain errors

Experimental Results

Computational Complexity Comparison

Table II Summary:

Detailed Analysis:

1. Permanent Computation: Reduced from multiple (at least ♯λ - 1 + d_λ) to only 1
2. Immanant Computation: While requiring ♯λ - 1 immanants:
   • Immanant computation complexity is lower than permanents
   • Can be directly computed using existing software packages (e.g., Wolfram)
   • Structure is simpler (typically single-column or near-single-column Young diagrams)
3. Clebsch-Gordan Coefficients: Completely eliminated, the most significant simplification

Numerical Examples

SU(3) Case:

• For n=2 photons, m=3 modes
• Number of irreducible representations ♯λ = 3
• Dimension d_λ = 3

Original Scheme Requires:

• Permanents: at least 3 - 1 + 3 = 5
• Clebsch-Gordan coefficients: complete computation required

This Work Requires:

• Permanents: 1
• Immanants: 3 - 1 = 2
• Clebsch-Gordan coefficients: 0

Method Correctness Verification

Through the SU(3) example in Appendix A, specifically verifies:

1. Correct identification of zero-weight states
2. Accurate D-function computation
3. Equivalence between immanants and traces of zero-weight states
4. Theoretical guarantee of single exponential decay

Related Work

Randomized Benchmarking

1. Standard RB: Emerson et al. 5, Magesan et al. 6, Knill et al. 7 established the randomized benchmarking framework for finite-dimensional quantum gates
2. Qudit Extensions: Amaro-Alcalá et al. 8, Jafarzadeh et al. 9 extended RB to high-dimensional systems
3. General Framework: Helsen et al. 10 provided a unified theoretical framework

Bosonic System Benchmarking

1. Original Scheme: Arienzo et al. 11 first proposed the bosonic RB scheme
2. Parallel Work: Wilkens et al. 12 studied benchmarking of bosonic and fermionic dynamics
3. This Work's Contribution: Significantly simplified the computational and experimental requirements of 11

Representation Theory Tools

1. Immanant Theory: Classical work by Littlewood 23
2. Kostant Relation: Established by Kostant 15 connecting immanants with zero-weight states
3. Applications: Applications in quantum optics by de Guise et al. 25,26

Noise Handling

1. Gauge Freedom Issues: Proctor et al. 19 identified gauge-freedom problems in RB
2. Randomized Compiling: Wallman et al. 20,21 proved that randomized compiling can convert coherent noise to incoherent noise
3. This Work's Position: Combining RC and RB can effectively address gauge issues

Conclusions and Discussion

Main Conclusions

1. Theoretical Contribution: Through the Kostant relation, establishes new immanant-based filtering functions that theoretically guarantee single exponential decay and accurate fidelity estimation.
2. Computational Simplification:
   • Eliminates Clebsch-Gordan coefficient computation
   • Reduces permanent count from O(♯λ + d_λ) to 1
   • Immanant count is O(♯λ) with simple structure
3. Experimental Simplification:
   • Weak coherent states replace Fock states
   • Intensity measurements replace photon-number resolution
   • Robustness to loss/gain errors
4. Enhanced Practicality:
   • Filtering functions can be predetermined
   • Data analysis pipeline simplified
   • Applicable to more experimental platforms

Limitations

1. Passive Transformation Restriction: Current scheme applies only to passive bosonic transformations (linear optics), excluding active transformations (e.g., squeezing operations).
2. Noise Model Assumptions: Assumes gate-independent, time-independent Markovian noise; actual systems may be more complex.
3. Extension Challenges: Extending to active bosonic transformations faces theoretical challenges as the relevant transformation group is non-compact.
4. Missing Experimental Verification: Paper is primarily theoretical work lacking actual experimental data verification.
5. Weak Coherent State Approximation: When using weak coherent states, α must be sufficiently small to ensure approximation validity.

Future Directions

1. Active Transformation Extension: Explore possibilities of extending the method to include active operations such as squeezing.
2. Experimental Verification: Verify method effectiveness and advantages on actual optical platforms.
3. Noise Model Generalization: Study handling of non-Markovian or time-correlated noise.
4. Algorithm Optimization: Develop more efficient immanant computation algorithms.
5. Fault-Tolerant Applications: Integrate the method into quality control processes for fault-tolerant quantum computing.

In-Depth Evaluation

Strengths

1. Theoretical Rigor:
   • Based on solid mathematical foundations (Kostant relation)
   • Complete proofs (Theorem 2)
   • Concrete example verification (SU(3) case)
2. Significant Innovation:
   • First application of Kostant relation to quantum benchmarking
   • Cleverly avoids Clebsch-Gordan coefficients, a major bottleneck
   • Fundamentally simplifies problem structure
3. High Practical Value:
   • Significantly reduces computational cost (eliminates exponentially complex components)
   • Substantially simplifies experimental requirements (weak coherent states + intensity measurements)
   • Directly applicable to existing data
4. Method Generality:
   • Applicable to arbitrary n-photon, m-mode systems
   • Robust to loss/gain errors
   • Filtering functions predetermined, independent of specific implementation
5. Clear Presentation:
   • Well-structured, logical flow
   • Detailed mathematical derivations
   • Concrete examples and appendices

Weaknesses

1. Lack of Experimental Verification:
   • Only theoretical analysis and numerical examples
   • Not tested on actual optical systems
   • Actual performance of weak coherent state scheme unknown
2. Limited Applicability:
   • Restricted to passive transformations
   • Extension to active operations difficult
   • Strong noise model assumptions
3. Incomplete Comparisons:
   • Limited numerical comparisons with original scheme
   • Lacking performance analysis under different parameters
   • Missing discussion of relationship to other benchmarking methods
4. Insufficient Technical Details:
   • Specific immanant computation algorithms not detailed
   • Numerical stability not discussed
   • Sampling complexity analysis missing
5. Limited Practical Guidance:
   • Lacking implementation guidelines
   • Insufficient parameter selection recommendations
   • Insufficient error analysis

Impact

1. Contribution to the Field:
   • Provides practical characterization tools for continuous-variable quantum computing
   • Demonstrates new applications of classical mathematics (representation theory) in quantum information
   • May inspire benchmarking methods for other quantum systems
2. Practical Value:
   • Significantly lowers implementation barriers
   • Enables more laboratories to perform quantum device characterization
   • May accelerate continuous-variable quantum computing development
3. Reproducibility:
   • Complete theoretical derivations
   • Clear method description
   • Implementable using existing software tools
   • Requires experimental verification to confirm practicality
4. Potential Impact:
   • May become standard method for bosonic system benchmarking
   • Provides methodological insights for other continuous-variable tasks
   • Promotes application of representation theory tools in quantum information

Applicable Scenarios

1. Ideal Scenarios:
   • Linear optical quantum computing platforms
   • Multi-mode interferometer characterization
   • Resource-constrained laboratories
   • Scenarios requiring rapid characterization
2. Limited Scenarios:
   • Systems requiring characterization of active operations
   • Environments dominated by non-Markovian noise
   • Applications requiring extreme precision
   • Systems with very large photon numbers (computational complexity still grows)
3. Recommended Usage Conditions:
   • Can prepare weak coherent states
   • Possess intensity measurement capability
   • Passive linear optical operations predominate
   • Require frequent characterization or online monitoring

Key References

11 M. Arienzo et al., "Bosonic randomized benchmarking with passive transformations," PRX Quantum 6, 020305 (2025) - Original scheme improved by this work

15 B. Kostant, "Immanant inequalities and 0-weight spaces," J. Am. Math. Soc. 8, 181 (1995) - Original Kostant relation paper

25 H. de Guise et al., "D-functions and immanants of unitary matrices and submatrices," J. Phys. A 49, 09LT01 (2016) - Application of Kostant relation in quantum optics

20 J. J. Wallman and J. Emerson, "Noise tailoring for scalable quantum computation via randomized compiling," Phys. Rev. A 94, 052325 (2016) - Randomized compiling method

Overall Assessment: This is a high-quality theoretical methodology paper that significantly simplifies bosonic randomized benchmarking through clever application of the Kostant relation. The theory is rigorous, innovation is strong, and practical value is high. Main weaknesses are lack of experimental verification and limited applicability scope. If subsequent experiments confirm its effectiveness, this will likely become an important work in the field. Readers are encouraged to follow up on subsequent experimental verification work and possible extension research.