Mathias Mikkelsen and Hubert Okadome Valencia of QunaSys Inc. have detailed a new comparison of methods for simulating bosonic systems on quantum computers. Their work reveals that first quantized mappings can require approximately fewer resources than second quantized mappings when calculating off-diagonal terms of the k-body reduced density matrix.
Detailed numeric analysis shows the binary first quantized mapping requires fewer gates than its second quantized counterparts for realistic system sizes, while also using exponentially fewer qubits than unary mappings. The binary mapping leads to lower one-norms than the unary mapping making it the overall most efficient choice for qubitization-based quantum phase estimation.
First Quantized Mappings Reduce Qubits for Bosonic Systems
QunaSys Inc. researchers detailed a comparison of resource requirements for both approaches when modeling N particles in M modes, focusing on the trade-offs between qubit usage and gate count. The unary first quantized mapping, though generally the most gate-efficient, demands N·M qubits, a substantial requirement that limits its scalability.
In contrast, the binary first quantized mapping achieves comparable gate counts to the most efficient unary approach when M equals , specifically for the Bose-Hubbard model and harmonic trap with short-range interactions. This suggests a performance crossover point dependent on system configuration. However, the team acknowledges that first quantized mappings are inapplicable to problems where the number of bosons is not conserved, necessitating the use of second quantized mappings in those scenarios, and the main challenge for first quantized bosonic simulations is the preparation of a state with bosonic symmetry.
Binary First Quantization Outperforms for Combined N and M Values
The number of qubits needed to represent a system of N particles in M modes differs significantly depending on the chosen quantum mapping technique. While unary first quantized mappings offer gate efficiency, they demand N·M qubits, a substantial resource requirement. This trade-off between gate complexity and qubit usage is a key consideration for near-term quantum simulations.
The researchers at QunaSys Inc. focused on comparing the unary first quantized (U1Q), binary first quantized (B1Q), unary second quantized (U2Q), and binary second quantized (B2Q) mappings to establish these resource trade-offs. This reduction in resources becomes increasingly significant as the complexity of the k-RDM increases.
k-RDM Off-Diagonal Terms Favor First Quantization
This efficiency stems from the drastically reduced number of Pauli strings needed to express these terms, a factor of (4N)^k fewer for the first quantized approach. While first quantized mappings may be less efficient for complex local terms, the benefit for k-RDM off-diagonal terms is substantial, particularly as the number of particles increases.
“In general, our results suggest that first quantized mappings are more gate-efficient at expressing the bosonic problem than the second quantized ones, when using the same basis functions,” the study reports. This advantage renders second-quantized mappings less competitive for qubitization in systems conserving particle number.
Gate Count Analysis: Binary First Quantization Efficiency
This equivalence in gate count for a single Trotter step, specifically for CNOT and gates, highlights a subtle relationship between mapping choice and system configuration. This gate efficiency is coupled with an exponential reduction in qubit requirements when compared to unary mappings, a significant advantage as quantum systems scale.
Qubitization-Based Phase Estimation Benefits from Binary Mapping
This advantage stems from destructive interference between Pauli strings, reducing the overall complexity of the Hamiltonian representation for full quantum phase estimation. While unary mappings can be efficient for individual Trotter steps, the cumulative gate count favors the binary approach when considering the entire algorithm. A key distinction lies in the scaling of Pauli strings required for each term in the Hamiltonian expansion.
The unary mapping necessitates four Pauli strings of length two per term, whereas the binary mapping requires 2^(⌈log₂M⌉) Pauli strings, each with a maximum length of ⌈log₂M⌉. This difference in string length and number directly impacts the number of quantum gates needed to implement the simulation.
The researchers found that the binary mapping’s ability to minimize the one-norm, a measure of circuit complexity, is important for optimizing qubitization-based quantum phase estimation. “Investigating both binary and unary mappings for a Hamiltonian and algorithm of interest is therefore useful, as the former can potentially have a better gate-efficiency for a given algorithm using a much smaller number of qubits,” the paper states. The findings highlight the importance of considering the entire quantum phase estimation process, not just individual steps, when evaluating the resource efficiency of different mappings.
👉 More information
🗞 First and second quantized digital quantum simulations of bosonic systems
✍️ Mathias Mikkelsen and Hubert Okadome Valencia
🧠 DOI: http://link.aps.org/doi/10.1103/s3y8-dptj




See today’s quantum computing news on Quantum Zeitgeist for the latest breakthroughs in qubits, hardware, algorithms, and industry deals.
