Fast Bosonic Control via Multiphoton Qubit-Oscillator Interactions

arXiv:2510.27035 · quant-ph · Submitted 2025-10-30 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Quantum Radio. Generated commentary on the latest quantum physics and condensed matter papers.

Kai: I'm Kai, and with me are Mira and Lev, guest researcher.

Mira: Today's paper: "Fast Bosonic Control via Multiphoton Qubit-Oscillator Interactions".

Kai: Multiphoton control protocols, such as those involving n-photon Law-Eberly interactions, substantially reduce state preparation times for various bosonic codewords compared to schemes relying on standard linear interactions.

Mira: First, who's behind it and why it matters.

Paper summary: Kai: To summarize what they're saying, the paper focuses on how to control the infinite-dimensional Hilbert space of a harmonic oscillator using an auxiliary qubit. They introduce a specific protocol, the n-photon Law-Eberly (nLE) protocol, which uses an nJC interaction to perform qubit-oscillator swaps of the form "e⟩ l⟩ ↔ g⟩ l + n〉. They also present a general approach called the Fine-tune-then-populate (FTP) protocol for arbitrary state synthesis by combining different interaction orders and selective qubit rotations. The central thesis is that these multiphoton control protocols substantially reduce preparation times for binomial, cat, and Gottesman-Kitaev-Preskill codewords compared to schemes relying on standard linear interactions.

Mira: That's a significant claim because it directly addresses the scalability issue where preparation time was the limiting factor in earlier demonstrations. The paper provides an upper bound for the n-photon interaction time scaling, which they state is TK,n = Kπomega + XKj=one πgnr((jn)!((j−one)n)!).

Lev: From an error correction perspective, if we can get those preparation times down significantly, it opens up possibilities for implementing larger logical states faster on a physical platform, even if the exact time bound is quite complex. We need to see how robust these protocols are when we factor in the decoherence times of planar resonators versus those in three dee cavities.

Kai: It sounds like they're setting up a framework that moves away from just linear qubit-oscillator interactions toward more complex, higher-order ones to achieve faster control. This suggests a shift in how we think about synthesizing these states on current hardware setups.

Mira: Indeed, the FTP protocol offers flexibility by allowing combinations of different interaction orders and photon-number selective qubit rotations for arbitrary state synthesis. It's about building a more versatile toolkit for state preparation rather than relying on one specific interaction type.

Lev: If the FTP protocol can handle arbitrary synthesis, that gives us more freedom when we think about constructing complex error-correcting codes that require many different types of states. We need to make sure the assumptions they used for their analysis hold up when we try to map this onto real superconducting circuits with realistic parameters.

Conclusion: Kai: So, looking at "Fast Bosonic Control via Multiphoton Qubit-Oscillator Interactions," the authors are essentially showing a method to prepare bosonic states much more efficiently using these multiphoton interactions. The implication is that we can build larger quantum codes faster on planar hardware, which is where most of our current experimental work is happening.

Mira: It boils down to the fact that they found a way to beat the time scaling limitations imposed by standard linear interactions for preparing states like cat states and Gottesman-Kitaev-Preskill codewords. The authors, Gorgichuk et al., are proposing practical ways to manage the state preparation bottleneck that has been plaguing superconducting circuit implementations for a while.

Lev: If this actually translates to hardware, it means we can start testing error-correcting codes that require more complex initial states much sooner than previously thought, which is a big step for experimental quantum information science. We still have to figure out the practical limits they mentioned regarding the rotating-wave approximation and the sideband control gate times at larger photon cutoffs.

Institute for Quantum Computing, University of Waterloo · Red Blue Quantum Inc. · Advanced ICT Research Institute, National Institute of Information and Communications Technology · Research Institute for Science and Technology, Tokyo University of Science · Department of Physics, The University of Tokyo

quant-ph

Submitted: 2025-10-30

Updated: 2026-09-15

Comments: Substantially revised with new sections comparing different control schemes and employing optimal control theory

Journal ref: Quantum 10, 2224 (2026)

DOI: 10.22331/q-2026-10-01-2224

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 83/100

The gist: Multiphoton control protocols, such as those involving n-photon Law-Eberly interactions, substantially reduce state preparation times for various bosonic codewords compared to schemes relying on

Key concepts

n-photon Law-Eberly (nLE) protocol
This is a specific control scheme that uses an nJC interaction to swap states between the qubit and the oscillator. It involves alternating qubit drives and these interactions in a sequence to achieve desired state transformations, offering a faster way to prepare bosonic codewords.
Fine-tune-then-populate (FTP) protocol
This is a general method for creating any arbitrary quantum state. It works by combining different interaction orders and selective qubit rotations with photon-number control. This flexibility allows for the synthesis of complex states by navigating between different subspaces in the system.
Bosonic codewords
These are specific target states in the harmonic oscillator Hilbert space that researchers aim to prepare. Examples include binomial, cat, and Gottesman-Kitaev-Preskill (GKP) states. Preparing these states efficiently is crucial for scalable bosonic quantum computation.
Time Scaling TK,n
This formula provides an upper bound on the time required to perform n-photon interactions. It shows how the preparation time scales with the interaction strength ($\Omega$), the number of photons ($n$), and other system parameters, helping researchers estimate how long it will take to prepare a specific state.

Terminology

Summary

Multiphoton control protocols, such as those involving n-photon Law-Eberly interactions, substantially reduce state preparation times for various bosonic codewords compared to schemes relying on standard linear interactions.

Core Problem and Motivation

Bosonic quantum computation leverages the infinite-dimensional Hilbert space of a harmonic oscillator to protect quantum information, but achieving arbitrary control over this space is necessary for manual synthesis of logical codewords. Previous demonstrations were often restricted to three-dimensional cavities due to long preparation times, which is a major bottleneck for scalable architectures on planar superconducting hardware. This paper addresses this bottleneck by proposing the use of nonlinear qubit-oscillator interactions combined with a qubit drive for preparing multiphoton states in the oscillator, specifically bosonic codewords, aiming to shorten state preparation and gate operation times.

Protocols and Control Schemes

The paper introduces several control schemes:

  1. The n-photon Law-Eberly (nLE) protocol: This generalizes the LE scheme to use an nJC interaction that performs qubit-oscillator swaps of the form e⟩ l⟩ ↔ g⟩ l + n⟩. The protocol involves a sequence of 2M operations, where each step alternates between a qubit drive and an nJC interaction.

  2. The Fine-tune-then-populate (FTP) protocol: This is introduced as a general approach for arbitrary state synthesis using combinations of different interaction orders and (photon-number) selective qubit rotations to introduce transitions between different subspaces.

  3. Conditional displacement control: This scheme relies on the dispersive regime, allowing for the synthesis of arbitrary states in the oscillator via operations like conditional displacements combined with qubit rotations.

State Preparation and Time Scaling

The effectiveness of these protocols is demonstrated through numerical simulations using realistic planar superconducting circuit parameters, including decoherence. Key findings include:

multiphoton control protocols substantially reduce preparation times for binomial, cat, and Gottesman-Kitaev-Preskill codewords compared to schemes relying on standard linear interactions.

For example, preparing a two-component even cat state using the two-photon protocol achieved a time of 26 ns compared to 110 ns for the one-photon protocol.

The time scaling for n-photon interactions is given by an upper bound:

TK,n = Kπomega + XKj=1 πgnr((jn)!((j−1)n)!).

The paper shows that two-photon state preparation outperforms the one-photon counterpart for larger circuit depths in certain regimes.

Arbitrary State Synthesis and Optimization

The FTP protocol allows for arbitrary state synthesis by combining different interaction orders. The total number of steps required is given by:

Karb(n, L) = Jn + L − (n − 1).

The protocol is advantageous when the support is spread thinly across subspaces, as it reduces the per-step time cost compared to the LE protocol. Optimal control theory (OCT) calculations using GRAPE algorithm are also employed to find numerically optimized pulses, which generally recover or slightly improve upon the nLE protocol's performance for specific states.

Robustness and Practical Limitations

The robustness of these schemes is validated by simulating a realistic circuit QED system including spurious terms and decoherence. However, practical limitations exist:

  1. Sideband control gate times become too short at larger Fock cutoffs, restricting the number of photons used (e.g., cutoff below 20 photons for n=2).

  2. The validity of the effective Hamiltonian generating the operation for larger circuit depths is limited by the rotating-wave approximation (RWA), where counter-rotating terms become significant for higher Fock states.

  3. Selective qubit rotations are limited by the dispersive regime, setting a bound on qubit driving strength that must be weaker than the multiphoton dispersive shift to remain in the perturbative description's validity.

Generalization to Multiple Oscillators

The single-mode protocols can be generalized to multiple oscillators coupled to the same auxiliary qubit. The protocol for synthesizing an arbitrary state in the two-oscillator subspace is given by a sequence of unitaries:

"Uˆ FTP = Uˆ(2) Y2−1 i=0 Uˆ 2,i L Y1−1 k6=0 Qˆ(n1) k4,k5,k6 Cˆk6n1+k5,k4> k4,k5,k6."

The upper bound on the number of steps for this protocol is:

**"Karb(n1, L1; n2, L2) = Jn1,n2 + n2L1 − (n1 − 1) + (L1 + 1)L2 − (n2 − 1).

Improvements for AI systems

Here are specific improvements to AI systems that can be derived from the concepts presented in this scientific paper:

  1. A new class of Quantum Error Correction (QEC) codes capable of operating on planar superconducting hardware with significantly reduced state preparation times.

  2. The ability for AI-driven quantum control algorithms to optimize pulse sequences (like GRAPE or OCT methods described) to achieve target quantum states faster than traditional analytical protocols (e.g., achieving a 48% speedup for two-photon vs. one-photon protocols in specific state preparation tasks).

  3. The development of Fine-Tune-Then-Populate (FTP) control algorithms that dynamically select the optimal sequence of interaction orders (linear, two-photon, three-photon, etc.) and selective qubit rotations based on the real-time structure and sparsity of the target quantum state to minimize preparation time.

  4. A protocol for generating arbitrary states in multi-oscillator systems (e.g., two coupled resonators) that is optimized by calculating a complex multidimensional punch card height matrix, allowing the AI to predict and execute the shortest possible sequence of operations required to reach a specific entangled state distribution across multiple modes.

  5. The ability for AI-designed control pulses to generate quantum states with high fidelity in realistic, noisy environments (including decoherence and spurious terms from circuit QED) by incorporating qubit resets or error conjugation strategies directly into the optimal control landscape.

Specific Capabilities of the Improved AI System:

  1. The improved system can perform complex bosonic state preparation (like GKP states, cat states, or binomial codewords) on superconducting qubits in a fraction of the time required by current linear interaction protocols.

  2. It can autonomously determine whether using a single-photon (linear) interaction or an n-photon (nonlinear) interaction is optimal for a given target state structure and hardware parameters, selecting the fastest protocol dynamically.

  3. It can synthesize highly complex, multi-mode entangled states across multiple oscillators by intelligently combining different types of multiphoton interactions to navigate the Hilbert space efficiently.

  4. The system can provide real-time feedback on the fidelity of a state preparation process and suggest necessary corrective operations (like selective rotations or phase corrections) to compensate for decoherence during execution, thereby maximizing the success rate in open quantum systems.

  5. It can predict the speed limit (Fock cutoff) for which a given control scheme remains viable under realistic circuit QED constraints, guiding experimentalists on whether to push the system toward higher photon numbers or use more robust (but slower) methods.

Abstract

We study the problem of n-fold rotationally symmetric bosonic state preparation, which is of great importance to bosonic quantum error correction, using a multiphoton interaction between an oscillator and an auxiliary qubit. We present an n-photon Law-Eberly (n LE) protocol that serves as an analytic baseline, alongside numerical optimal control calculations that further reduce state preparation time. We find that multiphoton control protocols substantially reduce preparation times for binomial, cat, and Gottesman-Kitaev-Preskill codewords compared to schemes relying on standard linear interactions. Further, we achieve arbitrary control over the oscillator's Hilbert space by combining different multiphoton interaction orders. We also extend these control improvements to the preparation of rotationally symmetric multi-oscillator states. Lastly, numerical simulations using realistic planar superconducting circuit parameters validate the robustness of our scheme against qubit and oscillator decoherence. Our findings can significantly enhance the performance of bosonic codes on planar superconducting hardware, an important ingredient for scalable fault-tolerant quantum computers.

Sources

Related papers