Local fermion-to-qudit mappings: a practical recipe for four-level systems
Rodolfo Carobene, Stefano Barison, Andrea Giachero, Jannes Nys · 2024
Read Local fermion-to-qudit mappings: a practical recipe for four-level systems by Rodolfo Carobene, Stefano Barison, Andrea Giachero, Jannes Nys (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-2412-05616.txt.
Introduction
Rodolfo Carobene, Stefano Barison, Andrea Giachero, and Jannes Nys give a recipe for simulating fermions on quantum hardware. Their twenty twenty-four paper uses qudits, quantum units with more than two levels, in place of plain qubits. The focus is on four-level units, called ququarts.
Fermions must be translated into the hardware's language, and the classic translation spreads operators across many units, which is costly. The new mappings keep the operators local on the qudits. The team gives one mapping for spinless fermions and two for spinful ones, and compares cost in gates and circuit depth. Tests on two standard lattice models show leaner simulations.
Catalog record
| CREATOR | Rodolfo Carobene, Stefano Barison, Andrea Giachero, Jannes Nys |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 52 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-05616
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Local fermion-to-qudit mappings: a practical recipe for four-level systems Rodolfo Carobene, Stefano Barison, Andrea Giachero, Jannes Nys Abstract In this paper, we present a new set of local fermion-to-qudit mappings for simulating fermionic lattice systems. We focus on the use of multi-level qudits, specifically ququarts. Traditional mappings, such as the Jordan-Wigner transformation (JWT), while useful, often result in non-local operators that scale unfavorably with system size. To address these challenges, we introduce mappings that efficiently localize fermionic operators on qudits, reducing the non-locality and operator weights associated with JWT. We propose one mapping for spinless fermions and two mappings for spinful fermions, comparing their performance in terms of qudit-weight, circuit depth, and gate complexity. Notably, for superconducting devices, the control complexity of qudits compared to qubits does not increase prohibitively for four-level systems, making them practical for near-term demonstrations, though control challenges grow with increasing dimensionality. Therefore, we provide solutions to the operator decompositions of the Trotterized quantum dynamics into one and two-qudit gates for all mappings. By leveraging the extended local Hilbert space of qudits, we show that these mappings enable more efficient quantum simulations in terms of two-qudit gates, reducing hardware requirements without increasing computational complexity. We validate our approach by simulating prototypical models such as the spinless t t - V V model and the Fermi-Hubbard model in two dimensions, using Trotterized time evolution. Finally, we show the connection between the plaquette constraint of our mapping and the ℤ 2 \mathbb{Z}_{2} Toric code. This connection can be exploited to prove that the ground state of the localized qudit mapping can be efficiently prepared using quantum circuits or measurement-based feedback control. Our results highlight the potential of qudit-based quantum simulations in achieving scalability and efficiency for fermionic systems on near-term quantum devices. I Introduction Quantum simulation of fermionic systems represents one of the most promising applications of digital quantum devices. However, they present significant challenges due to the anti-commutation relations of fermionic operators. Specifically, to conduct such simulations, one must map anti-commuting fermion operators onto Pauli operators \{\sigma^{x},\sigma^{y},\sigma^{z}\} , which inherently obey commutation relations [\sigma^{\mu},\sigma^{\nu}]=2i\epsilon^{\mu\nu\kappa}\sigma^{\kappa} . The most commonly used mapping between fermionic and spin degrees of freedom is the Jordan-Wigner transformation (JWT), which naturally arises from the second quantization formalism of fermions [ 1 ] . While physical fermionic Hamiltonians are typically local and consist of even-fermion parity operators, the JWT can map these local operators onto non-local Pauli (string) operators. This is due mainly to the introduction of transformation properties for the individual fermionic operators. These string operators scale with the system size for d>1 dimensional lattice Hamiltonians, eventually rendering the calculations challenging for large systems sizes [ 2 , 3 ] . Recent theoretical advancements have focused on new methodologies to simulate 2D Jordan-Wigner-transformed fermionic systems on contemporary noisy devices, prominently using Fermionic-SWAP gates (FSWAP) introduced in Ref. [ 4 ] and further elaborated in Ref. [ 5 ] . However, such methods require an increasing number of FSWAP operations as the system size or dimension grows, posing challenges for digital quantum simulations for large scale quantum simulations. Fortunately, the JWT is not the only mapping available and there is the possibility of searching for efficient encodings that require fewer qubits and/or reduce the operator weights [ 6 , 7 , 8 , 9 , 10 ] . Over the past 40 years, various works have focused on generalizing mappings to higher dimensions (>1D) to maintain locality in the operators and reduce the size of the Jordan-Wigner strings. One of the earliest studies to derive higher-dimensional JWT generalizations is the work of Wosiek [ 11 ] . Ball [ 12 ] and Verstraete-Cirac [ 13 ] independently explored similar ideas, introducing auxiliary fermions to expand the fermionic Hilbert space and then defining an auxiliary Hamiltonian to restrict the reachable Hilbert space. Compared to Ref. [ 11 ] , the Ball-Verstraete-Cirac (BVC) transformations [ 12 , 13 ] explicitly introduced auxiliary fermionic modes to counteract the Jordan-Wigner strings. These auxiliary modes effectively store the parity near the interaction terms, otherwise captured by the Jordan-Wigner string [ 2 ] , resulting in local qubit Hamiltonians. Given the potential benefits for quantum simulations, in recent years there has been a renewed interest in these localized approaches [ 14 , 10 , 15 , 16 , 17 , 18 , 19 , 6 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 ] , in particular through its interpretation as lattice gauge theories [ 15 , 16 , 30 , 31 , 32 ] . Unlike earlier methods (such as Refs. [ 12 , 13 ] ) that explicitly perform JWT, recent techniques define bosonic operators from fermionic ones, which are then mapped directly onto qubit operators without ordering the fermions. Since these novel methods do not require fermion ordering, they can be more easily generalized to different systems and higher dimensions. However, increasing the Hilbert space introduces constraints that must be fulfilled exactly, and it was pointed out in Refs. [ 33 , 34 ] that quantum circuits that minimally fulfill these constraints scale with the system size. Ref. [ 35 ] further highlighted that while local fermion-to-qubit mappings introduce local Hamiltonians, they require a high degree of state non-locality. However, recent work in Ref. [ 36 ] has experimentally demonstrated the feasibility and power of local encodings over Jordan-Wigner mappings. On larger systems, initial states that satisfy the constraints (i.e. a toric code state) can be prepared in constant depth using measurement and feed-forward control [ 37 , 38 ] . The use of multi-level quantum systems for quantum simulation has been gaining a lot of attention [ 39 , 40 , 41 , 42 , 43 ] . Recent work has explored its properties for simulating electronic systems [ 44 , 45 , 46 ] . These systems, known as qudits , are becoming an important alternative due to a number of potential improvements with respect to qubit devices [ 47 , 48 , 49 , 50 , 51 ] . These advantages include increased error resilience, and quantum error corrections with reduced code sizes [ 52 , 53 ] . It has been demonstrated in various works that the use of qudits can significantly reduce the complexity of quantum circuits by reducing the number of two-qudit gates compared to their qubit representations [ 44 ] . Indeed, joining multiple degrees of freedom into the local Hilbert space of a single qudit can decrease the number of inter-qudit operations. In addition, single-qudit gates are generally faster and exhibit higher fidelity compared to multi-qudit entangling operations [ 54 , 55 ] . Specifically for superconducting ququarts, increasing the local dimension tends to increase the number of single-qudit gates required, while simultaneously reducing the total number of entangling gates needed [ 44 ] . From a hardware perspective, employing qudits does not significantly increase the complexity of control electronics. For instance, in superconducting quantum processors, qubits and qudits share essentially the same control hardware and number of control lines. The main additional overhead arises from longer measurement sequences needed to distinguish multiple energy levels, along with modest