Utilizing redundancies in Qubit Hilbert Space to reduce entangling gate counts in the Unitary Vibrational Coupled-Cluster Method

Michal Szczepanik and Emil Zak · 2024

Read Utilizing redundancies in Qubit Hilbert Space to reduce entangling gate counts in the Unitary Vibrational Coupled-Cluster Method by Michal Szczepanik and Emil Zak (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-03955.txt.

Introduction

Michal Szczepanik and Emil Zak wrote this quantum computing paper, published in 2024. It cuts the cost of simulating vibrating molecules on quantum hardware by exploiting wasted space in how vibrations are mapped onto qubits.

The standard method, unitary vibrational coupled cluster, prepares molecular vibration states but burns many two-qubit gates, the error-prone kind. The direct mapping of vibrational modes leaves redundancies in the qubit space. By deleting half the controls in the method's rotations, the authors cut entangling gates by up to half in theory and over a quarter against practical rivals. Runs on Quantinuum's H1-1 machine gave clearly higher fidelities for six and eight qubit systems. For future fault-tolerant machines, some savings move elsewhere as extra gates, so the net gain depends on circuit design.

Catalog record

CREATORMichal Szczepanik and Emil Zak
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE25 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-03955

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Utilizing redundancies in Qubit Hilbert Space to reduce entangling gate counts in the Unitary Vibrational Coupled-Cluster Method Michal Szczepanik and Emil Zak Abstract We present a new method for state preparation using the Unitary Vibrational Coupled-Cluster (UVCC) technique. Our approach utilizes redundancies in the Hilbert space in the direct mapping of vibrational modes into qubits. By eliminating half of the qubit controls required in the Trotterized UVCC ansatz, our method achieves up to a 50% theoretical reduction in the entangling gate count compared to other methods and up to a 28% reduction compared practically useful approaches. This improvement enhances the fidelity of UVCC state preparation, enabling more efficient and earlier implementation of complex quantum vibrational structure calculations on near-term quantum devices. We experimentally demonstrate our method on Quantinuum’s H1-1 quantum hardware, achieving significantly higher fidelities for 6- and 8-qubit systems compared to existing implementations. For fault-tolerant architectures, eliminating half of the control qubits in multi-controlled rotations incurs an additional Toffoli gate overhead elsewhere in the circuit. Thus, the overall performance gain depends on the specific decomposition method used for multi-controlled gates. I Introduction Born-Oppenheimer vibrational energies are obtained as solutions to the Schroedinger equation for the motion of nuclei on a potential energy surface (PES) [ 1 ] . Achieving high-accuracy vibrational energy levels, at 1 \text{cm}^{-1} or better, requires variational or pseudospectral methods and a high-quality PES [ 2 , 3 ] . Both methods transform quantum mechanical differential equations into a matrix eigenvalue problem, the size of which scales exponentially with the number of vibrational modes. For this reason, despite their relevance to industry and fundamental science, high-quality nuclear motion calculations have been limited to a few atoms [ 4 , 3 ] . Quantum computers are expected to alleviate this limitation [ 5 , 6 ] . The development of computational methods, whether for quantum or classical devices, that address the exponential curse of dimensionality in rotational-vibrational calculations is important, given their applications in molecular spectroscopy (reference spectra for experiments [ 7 ] ), astrophysics (e.g., interstellar medium composition [ 8 ] ), exoplanet research (e.g., atmospheric retrirevals [ 9 , 10 ] ), material design (e.g., simulating phonons in crystals), and nanoelectronics (e.g., designing molecular junctions [ 11 , 12 ] and modelling vibrational relaxation [ 13 ] ). Notably, among plurality of other uses, vibrational calculations aid in the design and operations of trapped-ion [ 14 , 15 , 16 , 17 ] and superconducting [ 18 , 19 ] quantum computing architectures. Despite the unfavourable scaling, the variational approach carries the advantage of controlling error by adjusting the basis set size and quality. In this spirit, near-term quantum computing devices adopted the Variational Quantum Eigensolver (VQE) for vibrational calculations [ 20 , 21 , 22 ] . In VQE, a trial state is prepared by executing a quantum computing circuit, parametrised by variational coefficients that are optimised on a classical computer. However, the classical optimization process and the need for sampling multi-dimensional distributions limit the scalability of VQE [ 22 ] . An alternative algorithm suited for fault-tolerant quantum architectures is quantum phase estimation (QPE) [ 23 , 24 , 25 ] , which returns energy levels for a given trial state and a unitary embedding of the Hamiltonian [ 26 , 27 , 28 ] . In both VQE and QPE, the quality of initial trial state preparation is essential for accurate results [ 29 , 30 ] . In VQE, the space spanned by the ansatz sets a lower bound on the calculated energy levels, while in QPE, the ansatz’s overlap with the target eigenstate determines the likelihood of correctly measuring the corresponding eigenvalue. Typically, the larger the basis set representing the trial state vectors, the better the accuracy of the energy levels. For this reason, excitation-based formats for the variational ansatz have gained popularity. Classical computing techniques including vibrational configuration interaction (VCI) [ 31 ] and vibrational coupled-cluster (VCC) [ 32 ] allow control over the number of basis functions, through the choice of excitation level. For this reason, a quantum computing counterpart to VCC, the unitary vibrational coupled-cluster (UVCC) method, has been employed for state preparation in VQE [ 33 , 34 ] , and it is also applicable to QPE [ 35 , 36 , 29 ] . While the quality of the UVCC ansatz can be refined by increasing excitation levels or by enhancing the quality of the reference basis, such improvements lead to rapid increases in the computational complexity, particularly in terms of two-qubit CNOT gate count [ 34 , 6 ] . Quality of the basis set can be improved through solving reduced-dimensionality Schroedinger equations with effective Hamiltonians, as in the multimode procedure, although this sacrifices the Hamiltonian’s sparsity [ 37 , 2 , 3 ] . Upon encoding the UVCC operator into a qubit circuit, irrespective to encoding type, one faces artifacts that affect the efficiency of the quantum computing procedure, when no improvements are made [ 34 ] . For example, in unary encoding each vibrational basis state |n\rangle is represented by a qubit in state |1\rangle of an appropriate index, leaving a grossly large fraction of the Hilbert space nonphysical. This spurious part of the Hilbert space does not improve the solution. On the other hand, unary encoding is advantageous to compact binary encodings due to a lower number of entangling gates [ 34 ] . In this work, we propose a method for constructing the UVCC ansatz that requires fewer CNOT gates compared to other approaches. Our proposed procedure leverages on the awareness of a computational waste carried along the quantum computation using unary encoding. Specifically, we observe that a considerable portion of the qubit Hilbert space does not contribute to the solution. By eliminating control over spurious states, our method achieves up to a 50% reduction in CNOT gates compared to methods reported in previous studies [ 33 , 6 , 38 ] , including techniques based on Givens rotations [ 39 , 40 ] . This advantage is apparent for both low-excitation UVCC cases and in the asymptotic limit, remaining largely insensitive to the decomposition method for multi-controlled NOT gates. However, compared to techniques that employ ancillary qubits to decompose multi-controlled NOT gates into CNOTs, such as the recent method of Khattar and Gidney [ 41 ] , the relative advantage of our approach is typically less pronounced, leveling off at around 28%. In the context of fault-tolerant quantum computing, our approach reduces the number of control qubits needed in controlled rotations within the UVCC circuit by half, though this requires an additional \mathcal{O}(m) Toffoli gate overhead, where m represents the vibrational excitation level. Integrating our method with ancilla-assisted decomposition schemes for multi-controlled rotations offers limited overall reductions in Toffoli gate counts, depending on the chosen decomposition strategy. While optimizing Toffoli gate counts is not the focus of this work, the information-theoretic reductions presented here provide opportunities for such optimizations. II The Unitary Vibrational Coupled-Cluster ansatz In vibrational structure calculations, a fixed number of modes M is associated with a set of internal coordinates \{Q_{i}\} of the system. For simplicity, we choose the harmonic oscillator basis, since it allows for

Read the complete text