Efficient Quantum Circuits for Coherent Conversion Between General First- and Second-Quantized Many-Body Representations

Technical Summary: Efficient Quantum Circuits for Coherent Conversion Between First- and Second-Quantized Many-Body Representations

Problem Statement
Quantum simulation of many-body systems at a fixed particle number $N$ admits two mathematically equivalent but resource-different descriptions: the first-quantized (particle) representation and the second-quantized (occupation-number) representation.

• First Quantization: Embeds the state in $(\mathbb{C}^d)^{\otimes N}$, requiring $N \lceil \log_2 d \rceil$ qubits. It is space-efficient when $N \ll d$ but often incurs high gate complexity for operator implementation due to particle permutation symmetries.
• Second Quantization: Represents the state in a fixed-$N$ sector of Fock space using occupation vectors $(n_1, \dots, n_d)$, typically requiring $O(d)$ qubits. It naturally handles creation/annihilation algebras and particle-number flexibility but can be space-inefficient for small $N$ and large $d$.

While specialized algorithms exist for converting between these representations for specific statistics (e.g., purely fermionic or bosonic), there is a lack of a universal, symmetry-agnostic unitary transformation that can coherently convert general many-body states (including parastatistical sectors) between these two pictures without hard-coding the particle statistics into the circuit logic. This paper addresses the need for such a primitive to enable modular quantum workflows where different stages are optimized for different representations.

Methodology and Core Construction
The authors construct an explicit unitary operator $Q$ (and its inverse $Q^\dagger$) that maps a first-quantized state to its fixed-$N$ second-quantized form. The construction relies on a central structural identification: the occupation-number representation is the generalized group Fourier basis of the first-quantized Hilbert space.

The transformation $Q$ is composed of two main stages:

1. Strong Quantum Schur Transform ($U_{\text{Schur}}$):

   • Based on Schur–Weyl duality, the Hilbert space $(\mathbb{C}^d)^{\otimes N}$ decomposes into sectors labeled by Young diagrams $\lambda$ (irreducible representations of the symmetric group $S_N$) and $U(d)$ irreps.
   • $U_{\text{Schur}}$ performs the non-abelian Fourier transform for the commuting pair $(S_N, U(d))$. It maps the computational basis (particle configurations) to the Schur basis $|\lambda, \mu, \sigma\rangle$, where:
     • $\lambda$: The Young diagram shape (particle statistics sector).
     • $\sigma$: The Young–Yamanouchi word (multiplicity label for $S_N$).
     • $\mu$: The Gelfand–Tsetlin (GT) pattern (basis for the $U(d)$ irrep).
   • This step coherently writes the labels $\lambda, \mu, \sigma$ into dedicated registers.

2. Jordan–Schwinger Arithmetic Unitary ($U_{\text{JS}}$):

   • This stage converts the GT pattern data $\mu$ into the occupation vector $n = (n_1, \dots, n_d)$.
   • The authors establish that the diagonal generators of the Jordan–Schwinger map correspond to mode-number operators. Specifically, the occupation numbers are recovered as successive row-sum differences of the GT pattern: $n_\ell = R_\ell(\mu) - R_{\ell-1}(\mu)$, where $R_\ell$ is the sum of the first $\ell$ entries in the $\ell$-th row of the GT pattern.
   • For general parastatistical sectors where weight spaces are degenerate (multiple GT patterns map to the same occupation vector), the authors impose a canonical Gelfand–Tsetlin promise: the input state is supported only on a distinguished canonical representative $\mu_{\text{can}}(n, \lambda)$ for each weight. Under this promise, the map becomes bijective and reversible.
   • $U_{\text{JS}}$ implements this arithmetic using reversible logic (addition, subtraction, and reconstruction) to compute $n$ from $\mu$ and then uncompute the intermediate GT data.

Key Contributions

• Universal Unitary Converter: The paper provides the first explicit construction of a unitary $Q$ that works for bosons, fermions, and parastatistics (parabosons/parafermions) without prior knowledge of the statistics class. The statistics are diagnosed and stored in the $\lambda$ register during the process.
• Representation-Theoretic Insight: The work rigorously identifies the second-quantized occupation basis as the weight basis of the $U(d)$ representation arising from the Schur–Weyl decomposition. It frames the conversion as a change from the "pre-Fourier" computational basis to the "Fourier" weight basis.
• Complexity Analysis:
  • Quantum Cost: The total gate complexity is $\text{poly}(N, d, \log(1/\epsilon))$. The dominant cost comes from the strong Schur transform (using existing algorithms like Bacon-Chuang-Harrow or Krovi-Burchardt), while the arithmetic stage adds a cost of $O(d^2 \log N)$.
  • Classical Hardness: The paper establishes a sharp separation between coherent quantum conversion and explicit classical conversion.
    • For the forward map (First $\to$ Second), the classical output size is $\Omega(D_{\max}(N, d))$, where $D_{\max}$ is the number of distinct occupation vectors. This is polynomial in $d$ of degree $N$ (for fixed $N$) and exponential in $N$ when $d = \Theta(N)$.
    • For the inverse map (Second $\to$ First), the classical output size is $\Omega(N!)$ for bosonic and fermionic sectors due to the factorial support of Young symmetrizers.
• Sampling Hardness: The authors prove that an efficient classical sampler for the induced occupation-number distribution (for promised inputs) would imply $BQP \subseteq BPP$, suggesting that the conversion task is computationally powerful and likely hard for classical computers to simulate explicitly.

Results and Claims

• Efficiency: The authors demonstrate that a quantum computer can prepare the converted state in quantum memory with polynomial resources, whereas any classical algorithm producing an explicit coefficient list must pay a cost exponential in $N$ (or high-degree polynomial in $d$).
• Modularity: The transform $Q$ acts as a "representation router." It allows quantum workflows to utilize the most efficient representation for specific subroutines (e.g., first quantization for state preparation when $N \ll d$, second quantization for particle-number-flexible dynamics) and switch coherently between them.
• Generality: Unlike previous works focused on fermionic chemistry or bosonic symmetrization, this construction is "symmetry-agnostic." It handles parastatistics naturally, provided the canonical GT promise is met.
• Limitations: The authors acknowledge that for specific, known statistics (e.g., purely fermionic chemistry), specialized converters (like those in Refs. [23, 24]) may be more efficient in terms of constant factors or specific asymptotic regimes. The proposed $Q$ trades some efficiency for universality and modularity.

Significance
The paper reframes the equivalence between first and second quantization from a mere bookkeeping exercise into an operational primitive grounded in representation theory. By identifying the occupation-number representation as the group-Fourier basis of first quantization, the authors provide a rigorous mathematical foundation for coherent representation changes. This enables modular, fault-tolerant quantum simulation strategies where the choice of encoding can be dynamically optimized based on the specific task (preparation, evolution, measurement) without losing coherence or requiring classical re-synthesis of the state. The work also highlights a fundamental complexity-theoretic barrier: while quantum computers can navigate between these representations efficiently, explicitly materializing the result classically is intractable for large systems, reinforcing the potential for quantum advantage in many-body simulation workflows.