---
title: Quantum Paldus Transform (QPT)
url: https://www.emergentmind.com/topics/quantum-paldus-transform-qpt
type: topic
---

# Quantum Paldus Transform (QPT)

The Quantum Paldus Transform (QPT) is an isometry \(U_{\rm QPT}\) that maps the occupation-number basis of a fermionic Fock space with \(2d\) spin-orbitals to a symmetry-adapted basis of Gelfand–Tsetlin (GT) states labelled by total particle number \(N\), total spin \(S\), spin projection \(M\), and \(U(d)\) GT data. In the Unitary Group Approach (UGA), it realises antisymmetric \(U(d)\times SU(2)\) Howe duality and yields a basis that block-diagonalises spin-free Hamiltonians into irreducible \((N,S)\) sectors. In the complete active space (CAS) setting, the same transform is specialised to fixed \((N,S,M)\) sectors and to Shavitt-graph walks \(\delta_i\in\{0,1,2,3\}\), which enables an exact matrix product state (MPS) representation of CAS wavefunctions with bond dimension \(O(d^2)\), together with second- and first-quantized state-preparation routines whose complexity grows as \(O(d^3)\) [2506.09151] [2606.19457].

## 1. Definition of the transform

For \(d\) spatial orbitals with spin states \(\mu\in\{\uparrow,\downarrow\}\), the Fock basis is
\[
|n_1,n_2,\dots,n_{2d}\rangle,\qquad n_p\in\{0,1\},\quad \sum_p n_p=N,
\]
or equivalently
\[
|n_1,\dots,n_{2d}\rangle = \bigl(a_1^\dagger\bigr)^{n_1}\bigl(a_2^\dagger\bigr)^{n_2}\cdots \bigl(a_{2d}^\dagger\bigr)^{n_{2d}}|0\rangle .
\]
The QPT is the isometry
\[
U_{\rm QPT}:\mathcal H_{\rm Fock}\longrightarrow \mathcal H_N\otimes\mathcal H_S\otimes\mathcal H_M\otimes\mathcal H_{\mathbf d},
\]
with
\[
U_{\rm QPT}|n_1,\dots,n_{2d}\rangle = |N\rangle_N|S\rangle_S|M\rangle_M|\mathbf d\rangle_{\mathbf d},
\]
and inverse map
\[
U_{\rm QPT}^\dagger\Bigl(|N\rangle\otimes|S\rangle\otimes|M\rangle\otimes|\mathbf d\rangle\Bigr)=|n_1,\dots,n_{2d}\rangle .
\]
In the CAS-sector formulation, the corresponding map is written
\[
U_p:|N,S,M;\delta_1\dots\delta_d\rangle_{\rm Fock}\longrightarrow |N\rangle|S\rangle|M\rangle|\delta_1\dots\delta_d\rangle,
\]
where \(\delta_i\in\{0,1,2,3\}\) records the Shavitt-graph step at orbital \(i\). The GT basis states admit the expansion
\[
|N,S,M;\delta_1\dots\delta_d\rangle
=
\sum_{m_1,\dots,m_d}
\left(
\prod_{i=1}^d
C^{s_{[i]}m_{[i]}}_{s_{[i-1]}m_{[i-1]};s_i m_i}\cdot O_i^\dagger(m_i,\delta_i)
\right)|0\rangle^{\otimes 2d},
\]
with \(O_i^\dagger(m_i,\delta_i)=1\) for \(\delta_i=0\), \(a_{i,m_i}^\dagger\) for \(\delta_i=1\) or \(2\), and \(a_{i,+1/2}^\dagger a_{i,-1/2}^\dagger\) for \(\delta_i=3\). Once \(N,S,M\) are fixed,
\[
\langle \delta_1\dots\delta_d|U_p|n_1\dots n_d\rangle =
\begin{cases}
1,&\text{if the occupation list and spin-coupling pattern correspond to the walk }\delta,\\
0,&\text{otherwise,}
\end{cases}
\]
so on a fixed \((N,S,M)\) sector the transform acts as a permutation/isometry between the Fock states and the \(\delta\)-register. No simple factorization in terms of one- and two-fermion operators is given; the construction is specified through the GT expansion and the rule that \(U_p\) “reads off” \((n,\text{coupling})\mapsto \delta\) [2506.09151] [2606.19457].

## 2. Representation-theoretic structure

The QPT is grounded in the decomposition of the fermionic Fock space
\[
\bigwedge(\mathbb C^d\otimes\mathbb C^2)\cong
\bigoplus_{N=0}^{2d}\;\bigoplus_{S=0}^{\lfloor N/2\rfloor}
W_{N,S}^{U(d)}\otimes W_S^{SU(2)}.
\]
Here the commuting Lie-algebra actions are generated by
\[
E_{ij}=\sum_\mu a_{i\mu}^\dagger a_{j\mu},\qquad
\mathcal E_{\mu\nu}=\sum_i a_{i\mu}^\dagger a_{i\nu},
\]
with
\[
[E_{ij},E_{kl}]=\delta_{jk}E_{il}-\delta_{il}E_{kj},\qquad
[\mathcal E_{\mu\nu},\mathcal E_{\rho\sigma}]
=
\delta_{\nu\rho}\mathcal E_{\mu\sigma}-\delta_{\mu\sigma}\mathcal E_{\rho\nu},
\]
and \([E_{ij},\mathcal E_{\mu\nu}]=0\). This commuting structure is the algebraic content of Paldus duality and underlies the block structure in \((N,S)\) [2506.09151].

The GT basis is obtained by multiplicity-free recursive branching \(U(i)\times SU(2)\downarrow U(i-1)\times SU(2)\). The resulting states \(|N,S,M;\mathbf d\rangle\) are in one-to-one correspondence with walks on a Shavitt graph, or equivalently with triangular GT patterns. In the CAS construction, this graph-theoretic organisation is described as a transformation from the Fock basis to a “friendlier symmetry-adapted basis.” A plausible implication is that the efficiency of the later MPS construction is not an accident of tensor-network parametrisation alone; it is tied to the combinatorial constraints imposed by the symmetry-adapted branching data [2606.19457].

## 3. Quantum-circuit realisation

A direct circuit implementation of the QPT maintains four registers:
\(\ket N\), \(\ket S\), \(\ket M\), and \(\ket{\mathbf d}\), with \(\lceil\log(2d+1)\rceil\), \(\lceil\log(d+1)\rceil\), \(\lceil\log(4d+1)\rceil\), and \(2d\) qubits respectively. For an input \(\ket{n_1\dots n_{2d}}\), the circuit applies \(d\) successive Clebsch–Gordan subroutines, each processing the pair \(\ket{n_{2i-1},n_{2i}}\). At step \(i\), it updates \(M\mapsto M+\Delta m\in\{-1,0,+1\}\), applies a multiplexed Givens rotation \(G(\theta_{S,M})\) controlled by \(\ket S\ket M\), outputs two new bits \(\mathbf d_{2i-1},\mathbf d_{2i}\) encoding one of the four couplings \(\{0,1,2,3\}\), then updates \(S\mapsto S+\Delta s\in\{0,\pm \tfrac12\}\) and \(N\mapsto N+\Delta n\in\{0,1,2\}\) [2506.09151].

The elementary subroutines are expressed in fault-tolerant terms. Controlled incrementers on \(m\)-bit registers cost \(O(m)\) Toffolis using conditionally clean qubits, or \(3m\) Toffolis with five dirty qubits. Multiplexers \(S(U)=\sum_i |i\rangle\!\langle i|\otimes U_i\) can be implemented in \(I\) Toffolis by unary iteration. Controlled Givens rotations \(R_y(\theta_{S,M})\) are approximated via data-lookup-based or phase-gradient primitives with cost \(O(I_d\,q)\) Toffolis per layer, or via unary iteration in \(O(I_d)\) Toffolis. Since at the \(i\)-th step there are \(I_i=O(i^2)\) distinct \((S,M)\) rotation angles, \(\sum_i I_i=O(d^3)\), while the total incrementer cost is \(O(d\log d)\). The resulting QPT compilation has \(O(d^3)\) Toffoli complexity, with \(O(d)\) sequential Clebsch–Gordan layers and layer depth \(O(\log d+\log(1/\varepsilon))\). The paper also lists three constant-factor optimisations: measurement-based uncomputation, phase-gradient reuse, and multi-index QROM or clean/dirty \({\rm SELECT}\)/\({\rm SWAP}\), the last of which can reduce constants by \(10\times\) [2506.09151].

## 4. CAS wavefunctions in the Paldus basis

A CAS wavefunction in the UGA basis is written as
\[
|\Psi_{\rm CAS}\rangle
=
|N\rangle|S\rangle|M\rangle
\sum_{\delta\in\mathcal D(N,S)}\alpha_\delta |\delta_1\dots\delta_d\rangle,
\]
where \(\mathcal D(N,S)\) is the set of valid walks on the Shavitt graph consistent with total electron number \(N\) and total spin \(S\). The central observation of the 2026 construction is that these amplitudes can be represented exactly as an MPS over the physical alphabet \(\delta_i\in\{0,1,2,3\}\) [2606.19457].

The bond index at site \(i\) is defined as \(v_i\equiv(a_i,b_i,c_i)\), where \(a_i\) counts double occupations in orbitals \(1,\dots,i\), \(b_i\) counts single occupations, and \(c_i\) counts empties, subject to
\[
a_i+b_i+c_i=i,\qquad 2a_i+b_i\le N,\qquad a_i\le a_d,\qquad c_i\le c_d.
\]
With \(v_0=(0,0,0)\) and \(v_d\) fixed by \(N,S\), the amplitudes satisfy
\[
\alpha_{\delta_1\dots\delta_d}
=
\sum_{v_1\dots v_{d-1}}
(A_1^{\delta_1})_{v_0,v_1}
(A_2^{\delta_2})_{v_1,v_2}\cdots
(A_d^{\delta_d})_{v_{d-1},v_d},
\]
where each \(A_i^\delta\) is a very sparse indicator matrix that enforces \(v_i=v_{i-1}+f(\delta_i)\) and multiplies by the scalar \(\alpha\) if that path is in the support of the CAS. Because \(v_i\equiv(a_i,b_i,c_i)\leftrightarrow (a_i,c_i)\), the number of possible bond states obeys
\[
\chi_i=(N/2+1)\cdot\bigl(i-\max(0,(i-N)/2)+1\bigr)=O(N\,i)=O(d^2),
\]
and therefore
\[
\max_i \mathrm{SchmidtRank}_i \le \max_i |\{v_i\}|=O(d^2).
\]
The outcome is an exact MPS with bond dimension \(\chi=O(d^2)\). This directly supports the claim that CAS states expanded in the Paldus basis admit an efficient classical representation, despite the statement that encoding a CAS state classically is traditionally believed to be intractable for chemically relevant systems [2606.19457].

## 5. CAS-state preparation and resource estimates

The CAS-state preparation protocol has two stages. First, the MPS over the \(\delta\)-alphabet is loaded into the registers \(|N\rangle|S\rangle|M\rangle|\delta\rangle\). Second, the inverse transform \(U_p^\dagger\) converts that symmetry-adapted representation back into the standard second-quantized Fock basis. The 2026 paper gives an explicit loading routine and combines it with the inverse-QPT circuit complexity inherited from the earlier QPT construction [2606.19457].

At each site \(i\), the loading routine carries forward a bond register \(|v_{i-1}\rangle\), initialises two qubits \(|00\rangle\) for the physical \(\delta_i\) register, and a temporary QROM register \(|0\rangle^{\otimes b}\) for angle loading. The unitary \(U_i\) acts as
\[
|0\rangle_{p_i}|v_{i-1}\rangle_b|0\rangle_t
\;\to\;
\sum_{\delta_i=0}^3
A_i(\delta_i,v_{i-1})\,|\delta_i\rangle_{p_i}|v_{i-1}+f(\delta_i)\rangle_b|0\rangle_t.
\]
Its implementation consists of four steps: QROAM; a controlled \(R_y(\theta)\), \(R_z(\phi)\) tree preparing \(\sum r_i e^{i\phi_i}|\delta_i\rangle\) on the physical two-qubit register; inverse QROAM; and controlled adders updating \(a,b,c\) conditioned on \(\delta_i\). For fixed \(\varepsilon\), the resource counts per \(U_i\) are \(O(\sqrt{\chi\cdot b})+O(\sqrt{\chi})\) Toffolis for QROAM and its inverse, \(O(b)\) for the \(b\)-bit rotations, and \(O(\log\chi)\) for the adders. With \(\chi=O(d^2)\), \(b=O(\log(d/\varepsilon))\), and \(\sqrt{\chi\cdot b}=O(d\sqrt{\log d})\), the dominant cost per site is \(O(d\sqrt{\log d})\) Toffolis, giving \(O(d^2\sqrt{\log d})\) Toffolis and depth over all \(d\) sites.

| Procedure | Resource estimate | Role |
|---|---:|---|
| MPS load | \(O(d^2\sqrt{\log d})\) Toffolis, depth \(O(d^2\sqrt{\log d})\) | Prepare \(|N\rangle|S\rangle|M\rangle|\delta\rangle\) |
| Inverse QPT \(U_p^\dagger\) | \(O(d^3)\) Toffolis, depth \(O(d^3)\) | Recover the second-quantized GT/Fock state |
| Second-quantized CAS preparation | \(O(d^3)\) total | Combine MPS load and inverse QPT |
| First-quantized CAS preparation | \(O(d^3)\) total | Conversion cost \(O(d^2)\) is subleading |

For the inverse transform, the CAS paper cites a detailed circuit by Burkat et al. with \(O(d^3)\) Toffoli count and \(O(d^3)\) circuit depth. Consequently, the total second-quantized cost is
\[
O(d^2\sqrt{\log d})+O(d^3)=O(d^3),
\]
which the paper describes as, to the best of its knowledge, an exponential improvement over the state of the art for CAS-state preparation. The first-quantized version has the same asymptotic complexity because the determinant-to-antisymmetrized-list conversion costs \(O(d^2)\), which is subleading [2606.19457].

## 6. Applications, scope, and conceptual clarifications

The 2025 QPT paper identifies four principal applications. First, spin-free second-quantized Hamiltonians,
\[
H=\sum_{ij,\mu} h_{ij}\,a_{i\mu}^\dagger a_{j\mu}
+\tfrac12\sum_{ijkl,\mu\nu} v_{ij,kl}\,a_{i\mu}^\dagger a_{j\nu}^\dagger a_{l\nu} a_{k\mu},
\]
lie in the universal-enveloping algebra of \(\mathfrak u(d)\oplus\mathfrak u(2)\), so the Paldus basis block-diagonalises them into irreducible \((N,S)\) sectors and achieves maximal sparsity. The Fermi–Hubbard Hamiltonian is given as an explicit example that acts irreducibly within each \((N,S)\) block. Second, the transform supports efficient preparation of configuration-state functions (CSFs): one prepares \(|N\rangle|S\rangle|M\rangle|\mathbf d\rangle\) and applies \(U_{\rm QPT}^\dagger\), or, conversely, applies \(U_{\rm QPT}\), measures \((N,S,M)\), and post-selects. Third, it yields a direct interpretation of reduced density matrix elements in terms of \(SU(2)\) angular-momentum coupling. Fourth, it enables decoherence-free subsystems for collective \(U(2)\) noise, because in the UGA basis \(U^{\otimes d}\) acts only on the \(|M\rangle\) register while leaving \(|\mathbf d\rangle\) invariant [2506.09151].

Two conceptual clarifications are recurrent. One is terminological: the transform is often described as unitary, but in the CAS presentation it is stated to be “strictly, an isometry on the CAS sector.” The other concerns implementability: the explicit algorithm is a circuit-level isometry between basis descriptions, not a factorized fermionic operator identity, and no simple factorization into one- and two-fermion operators is given. A further implication of the CAS results is that the usefulness of the QPT is not limited to symmetry resolution alone; it also provides the coordinate system in which CAS amplitudes become exactly MPS-compressible with polynomial bond dimension. The open challenges named in the QPT paper are extending the method to full spin-dependent Hamiltonians with \(U(d)\times SU(2)\times SU(2)\), exploiting higher symmetries such as point-group and spatial symmetries, and integrating the transform with error-mitigation and variational algorithms [2606.19457] [2506.09151].

Source: https://www.emergentmind.com/topics/quantum-paldus-transform-qpt