Arxiv:2008.13471V3 [Quant-Ph] 12 Apr 2021 with Bosonic Modes 14 Tection They Afford

Quantum Information Processing with Bosonic Qubits in Circuit QED

Atharv Joshi,1 Kyungjoo Noh,2, ∗ and Yvonne Y. Gao1, † 1Centre for Quantum Technologies, National University of Singapore 2AWS Center for Quantum Computing, Pasadena, CA, 91125, USA The unique features of quantum theory offer a powerful new paradigm for information processing. Translating these mathematical abstractions into useful algorithms and applications requires quantum systems with significant complexity and sufficiently low error rates. Such quantum systems must be made from robust hardware that can coherently store, process, and extract the encoded information, as well as possess effective quantum error correction (QEC) protocols to detect and correct errors. Circuit quantum electrodynamics (cQED) provides a promising hardware platform for implementing robust quantum devices. In particular, bosonic encodings in cQED that use multi- photon states of superconducting cavities to encode information have shown success in realizing hardware-efficient QEC. Here, we review recent developments in the theory and implementation of quantum error correction with bosonic codes and report the progress made towards realizing fault-tolerant quantum information processing with cQED devices.

CONTENTS I. INTRODUCTION

I. Introduction 1 A quantum computer harnesses unique features of A. Organization of the article 2 quantum theory, such as superposition and entangle- ment, to tackle classically challenging tasks. To perform II. Concepts of Bosonic Quantum Error faithful computation, quantum information must be Correction 2 protected against errors due to decoherence mechanisms and operational imperfections. While these errors are III. Performance of Bosonic Codes for Loss and relatively insignificant individually, they can quickly Dephasing Errors 4 accumulate to completely scramble the information. A. Rotation-symmetric codes: Binomial and cat To protect quantum information from scrambling, codes 5 the theoretical frameworks of quantum error correction B. Translation-symmetric codes: GKP codes 6 (QEC) [1, 2] and fault-tolerant quantum computation [3] C. Biased-noise bosonic qubits: Two-component were developed in the early days of quantum computing. cat codes 7 Essentially, these frameworks devise encodings which D. Comparison of various bosonic codes for loss map a collection of physical elements onto a single and dephasing errors 8 ‘logical’ bit of quantum information. Such a logical qubit is endowed with cleverly chosen symmetry properties IV. The cQED Hardware for Bosonic Codes 8 that allow us to extract error syndromes and enact error A. Components of the cQED architecture 9 correction without disturbing the encoded information. B. Coherence of superconducting cavities 9 An important metric for evaluating the effectiveness of QEC implementations is the break-even point, which V. Realization of Bosonic Logical Qubits 10 is achieved when the lifetime of a logical qubit exceeds A. Operations on single bosonic modes 10 that of the best single physical element in the system. B. Operations on multiple bosonic modes 12 Achieving the break-even point entails that additional C. Implementations of QEC 13 physical elements and operations introduced to a QEC VI. Towards Fault-tolerant Quantum Computation process do not cause more degradation than the pro- arXiv:2008.13471v3 [quant-ph] 12 Apr 2021 with Bosonic Modes 14 tection they afford. Hence, reaching the break-even point is a critical pre-requisite for implementing fault- VII. Outlook and Perspectives 15 tolerant gates and eventually performing robust quantum information processing on a large scale. References 16 In the conventional approach to QEC, the physical elements are realized by discrete two-level systems. In this approach, even a simple QEC scheme designed to correct single errors, such as the Steane code [2], ∗ This review paper was written before K. Noh joined AWS Center requires tens of two-level systems, ancillary qubits, and for Quantum Computing. measurement elements. Constructing physical devices † Corresponding author: [email protected] that contain these many interconnected elements can be a significant engineering challenge. More crucially, having many interconnected elements often degrades the device 2 performance and introduces new uncorrectable errors last decade. Subsequently, in Sec. V, we explore the latest such as cross-talk due to undesired couplings between the developments in implementing robust universal control elements. Over the last decade, many proof-of-principle on bosonic qubits encoded in superconducting cavities, demonstrations of QEC schemes have been realized both in terms of single-mode gates as well as novel with encoding schemes based on two-level systems [4– two-mode operations. We then describe the different 10]. However, given the practical challenges described strategies for detecting and correcting quantum errors above, these demonstrations have not deterministically on bosonic logical qubits encoded in superconducting extended the performance of the logical qubits beyond cavities. that of the best available physical qubit in the system. In Sec. VI, we discuss the concept of fault-tolerance A promising alternative with the potential to realize and how that might be realized with protected bosonic robust universal quantum computing with effective QEC qubits. Here, we also feature some novel schemes that beyond the break-even point involves encoding logical concatenate bosonic codes with other QEC codes to qubits in continuous variables [11–13]. In particular, protect against quantum errors more comprehensively. superconducting microwave cavities coupled to one or Finally, in Sec. VII, we provide some perspectives for more anharmonic elements in the circuit quantum achieving QEC on a larger scale. We conclude by electrodynamics (cQED) architecture provide a valuable remarking on the appeal of the modular architecture, resource for the hardware-efficient encoding of logical which offers a promising path for practical and robust qubits [14, 15]. These cavities have a large Hilbert quantum information processing with individually pro- space for encoding information in multi-photon states tected bosonic logical elements. compactly and thus form a logical qubit in a single piece of hardware. The anharmonic element, typically in the form of a transmon and henceforth referred to II. CONCEPTS OF BOSONIC QUANTUM as the ‘ancilla’, provides the necessary non-linearity to ERROR CORRECTION control and measure cavity states. This strategy of using multi-photon states of superconducting cavities to The general principle of QEC is to encode logical encode logical information is also known as bosonic codes. quantum information redundantly in a large Hilbert Implementations of bosonic codes in cQED have thus far space with certain symmetry properties, which can not only demonstrated QEC at the break-even point [16], be used to detect errors. In particular, logical code but also robust operations [17–21] and fault-tolerant states are designed such that they can be mapped onto measurement of error syndromes [22], thus making rapid orthogonal subspaces under distinct errors. Crucially, the progress in recent years. logical information can be recovered faithfully only if the In this article, we review the recent developments in mapping between the logical and the error states does bosonic codes in the cQED setting. In particular, we not distort the code words. highlight the progress made in demonstrating effective Mathematically, these requirements can be succinctly QEC and information processing with logical elements described by the Knill-Laflamme condition [23], which implemented using bosonic codes. These recent works states that an error-correcting code C can correct provide compelling evidence for the vast potential of any error operators in the span of an error set ˆ ˆ bosonic codes in cQED as the fundamental building E ≡ {E1, ··· , E|E|} if and only if it satisfies: blocks for robust universal quantum computing. ˆ ˆ† ˆ ˆ ˆ PCE` E`0 PC = α``0 PC (1) A. Organization of the article for all `, `0 ∈ {1, ··· , |E|}, where |E| is the size of the error set, PˆC is the projection operator to the code space C, In Sec. II, we begin by outlining the basic principles and α``0 are matrix elements of a Hermitian and positive of QEC as well as bosonic codes. Here, we highlight semi-definite matrix. A derivation of the Knill-Laflamme their nature by comparing a bosonic code with a multi- condition is given in Ref. [24]. qubit code in the presence of similar errors. In Sec. III, Bosonic codes achieve these requirements by cleverly we present various bosonic encoding schemes proposed configuring the excitations in a harmonic oscillator mode. in literature and compare their respective strengths For instance, in the cQED architecture, information and limitations. In particular, we wish to emphasize is encoded in multi-photon states of superconducting crucial considerations for constructing these codes and microwave cavities. We will illustrate how bosonic QEC evaluating their performance in the presence of naturally- codes work by describing the simplest binomial code [25], occurring errors. also known as the ‘kitten’ code. We will then compare In Sec. IV, we introduce the key hardware building this code with its multi-qubit cousin, the ‘4-qubit code’ blocks required for cQED implementations of bosonic (or the [[4, 1, 2]] code or the distance-2 surface code) [26], encoding schemes. In this section, we also consolidate in the presence of similar errors in the cQED architecture. the progress made in improving the intrinsic quality This comparison, adapted from Ref. [27], emphasizes the factors of superconducting microwave cavities over the hardware-efficient nature of the bosonic QEC approach. 3

(a) (b) ......

4

3

2 1

1 0

0

(c) (d) cavity control

ancilla control readout

FIG. 1. Hardware comparison between the kitten code and the 4-qubit code. (a) Encoding a single logical qubit using multiple energy levels of a single harmonic oscillator, where the only dominant error channel is photon loss. (b) The 4-qubit scheme based on a collection of two-level systems. Individually, each two-level system can experience errors due to spontaneous absorption and emission. Moreover, additional errors can be introduced due to undesired couplings between these systems. (c) The typical hardware for implementing a bosonic logical qubit. The logical information is encoded in a single superconducting cavity (orange) in the kitten code, which can be fully controlled by a single non-linear ancilla (green) and read out via another cavity (gray). (d) Example of a system designed to implement the 4-qubit code, where the logical information is spread out across four data qubits (orange) checked by three ancillary ones (green). The interactions between each two-level system are mediated by coupling cavities (teal), and their respective states are read out via seven planar cavities (gray).

The kitten code is designed to correct only single by mapping the normalized error states |3i and |1i back photon loss events, ora ˆ, which are the dominant to the original code states |0Li and |1Li, we can recover error channel in superconducting cavities. This scheme the input logical state up to an overall normalization encodes logical information in the even photon number constant: parity subspace of a harmonic oscillator (Fig. 1a): √ aˆ|ψLi = 2(α|3i + β|1i) 1 recovery √ |0Li = √ (|0i + |4i), |1Li = |2i. (2) −−−−−→ 2(α|0Li + β|1Li) ∝ |ψLi. (4) 2 Note that we cannot faithfully recover logical informa- With these code words, a single photon loss event maps tion if the logical states have different average√ photon an even parity logical state |ψLi = α|0Li + β|1Li to an √ numbers. For instance, with |0 i = 1 |0i + 3 |4i, which odd parity error statea ˆ|ψ i = 2(α|3i + β|1i). Because L 2 2 L has 3 photons on average, as opposed to 2 photons for of the parity difference, the code space and the error |1 i = |2i, a single photon loss and recovery event yields space are mutually orthogonal. Therefore, single photon L a state which is not proportional to |ψ i: loss events can be detected by measuring the photon L nˆ √ √ number parity operator Πˆ 2 = (−1) . Importantly, aˆ|ψLi = 3α|3i + 2β|1i since the error statesa ˆ|0Li anda ˆ|1Li have the same recovery √ √ normalization constant, −−−−−→ 3α|0Li + 2β|1Li. (5)

† † h0L|aˆ aˆ|0Li = h1L|aˆ aˆ|1Li = 2, (3) Specifically, the error and recovery process distorts the relative phase between |0Li and |1Li. An intuitive way or equivalently, the logical states |0Li and |1Li have the to understand such a phase distortion is by considering same average photon number, a single photon loss event the role of the environment. When |0Li and |1Li have does not distort the encoded information. In other words, different average photon numbers, the environment gains 4 partial information on whether |ψLi was in |0Li or in to store logical information, a single ancilla (typically |1Li. In the example above, |0Li has more photons a transmon) to measure and control the cavity state, than |1Li and hence has a higher probability of losing and a single readout cavity mode to measure the ancilla a photon. When one photon is lost, the environment state. In contrast, a cQED realization of the 4-qubit knows that the encoded state was more likely to be code (Fig. 1d) uses 4 data qubits to encode logical |0Li than |1Li and the weights in Eq. (5) are therefore information, 3 ancillae to measure the three stabilizers, adjusted accordingly. Alternatively, we can say that the and additional cavity modes to connect and measure environment performs a weak measurement on the logical all 7 physical qubits. Apart from the pure complexity state in the |0L/1Li basis, thus leading to the dephasing of realizing such a device, the presence of additional of |ψLi in the |0L/1Li basis. elements introduces other error channels such as cross- Hence, while the even parity structure allows the talk arising from spurious couplings between the physical detection of single photon loss events, it does not qubits. While these effects can be calibrated and guarantee the recoverability of the logical information mitigated on a small scale with clever techniques [28, 29], without any distortion. Faithful recovery is ensured by they can quickly become intractable for more complex selecting |0Li and |1Li to have the same average photon devices. This comparison thus illustrates the advantage number. In the case of the kitten code, average photon of using the multiple levels of a single bosonic mode over numbers are matched by choosing equal coefficients for using multiple two-level systems as a redundant resource the vacuum and the four photon state components in the for QEC. logical zero state. While we focus on bosonic codes that encode a qubit in More generally, a code C can correct single photon a single bosonic mode, there are proposals for encoding loss events if it satisfies the Knill-Laflamme condition for a qubit in many bosonic modes, namely permutation- the error set {I,ˆ aˆ}. That is, the projection operator invariant codes [30–33]. These codes are tailored for to the code space should satisfy PˆCxˆPˆC ∝ PˆC for all excitation loss errors and generalize the simple 4-qubit xˆ ∈ {I,ˆ a,ˆ aˆ†, aˆ†aˆ}. The first condition withx ˆ = Iˆ is code. Other codes of a similar nature have also been trivially satisfied for any code. For even parity codes, studied in Refs. [34–37]. which are composed of logical states of even photon number parity, the second and the third conditions with xˆ =a ˆ andx ˆ =a ˆ† are satisfied due to the parity III. PERFORMANCE OF BOSONIC CODES structure. This implies that even parity codes are capable FOR LOSS AND DEPHASING ERRORS of detecting, but not necessarily correcting, any single photon loss and gain events. Single photon loss events In general, bosonic modes in cQED systems typically can be corrected if the fourth condition is met, that is, undergo both photon loss and dephasing errors. Photon † PˆCaˆ aˆPˆC ∝ PˆC, or equivalently, if all logical states have loss is considered to be the dominant error-channel. The the same average photon number. rate of photon loss is determined by the internal quality Now, using an example inspired by Ref. [27], we factor (Qint) of the superconducting cavity. Intrinsic compare the kitten code with its multi-qubit cousin, the dephasing is usually insignificant for such cavities [38]. 4-qubit code, whose logical states consist of 4 distinct However, as the cavity is dispersively coupled to a two-level systems (Fig. 1b): non-linear ancilla, if the ancilla experiences undesired absorption (Γ↑) or emission (Γ↓) of excitations due to 1 stray radiation, then the encoded logical information |0Li = √ (|0000i + |1111i), 2 undergoes induced dephasing. The rate of this dephasing 1 depends on Γ↑,Γ↓, and the coupling strength χ. In |1Li = √ (|0101i + |1010i). (6) 2 general, a transition of the ancilla state leads to a rotation on the logical qubit by an angle ∼ χt, where t is the The 4-qubit code is stabilized by three stabilizers: time the ancilla spends in the resulting state after the Sˆ1 = Zˆ1Zˆ3, Sˆ2 = Zˆ2Zˆ4, and Sˆ3 = Xˆ1Xˆ2Xˆ3Xˆ4. This transition. In the limit where this rotation is small, scheme is capable of detecting any arbitrary single-qubit we can apply the Markovian approximation and use errors. Moreover, it can correct single excitation loss the following Lindblad master equation to describe the − − − − errors, {σˆ1 , σˆ2 , σˆ3 , σˆ4 }, via approximate QEC [26], evolution of the encoded qubit: which happens when the Knill-Laflamme condition is dρˆ(t) †
approximately satisfied only to a certain low order in = κD[ˆa] + κφD[ˆa aˆ] ρˆ(t), (7) the error parameters. This capability is comparable to dt the protection afforded by the kitten code against single where κ and κφ are the photon loss and dephasing rates photon loss errors. respectively, and D[Aˆ](ˆρ) ≡ AˆρˆAˆ† − 1 {Aˆ†A,ˆ ρˆ} is the Despite their comparable error-correcting capability, 2 dissipation superoperator. Note that loss and dephasing the 4-qubit code and the kitten code incur significantly errors are generated by the jump operatorsa ˆ andn ˆ =a ˆ†aˆ different hardware overheads. A cQED implementation respectively. We say that a system is loss dominated if of the kitten code (Fig. 1c) requires a single bosonic mode κ  κφ and dephasing dominated if κ  κφ. 5

FIG. 2. Visualization and performance of different bosonic codes. (a) - (d) Wigner functions of the code words for the four-component cat, binomial kitten, GKP square lattice, and two-component cat code, respectively. (e) A qualitative description of the robustness of various classes of bosonic codes against dephasing and photon loss errors. Note that for the four-component cat code and the binomial kitten code, the logical states in the |0/1Li basis are invariant under a 90° rotation, up to a global phase. However, a general code state, that is, an arbitrary superposition of |0Li and |1Li), is invariant only under the 180° rotation.

Importantly, in typical experimental regimes, dephas- parity operator and is invariant under the 180° rotation, ing errors induced by transitions of the ancilla state are thus making it robust against single photon loss errors. generally non-Markovian. This limits the performance of However, the modified code is now also higher in energy some current implementations [39] of bosonic QEC codes. as its logical states have 3 photons on average, as opposed In this section, we review the various single-mode to 2 in the case of the kitten code. This additional bosonic codes, namely, cat (Fig. 2a) and bino- redundancy makes the modified binomial code robust mial (Fig. 2b) codes (rotation-symmetric), Gottesman- against single dephasing events as well. In other words, Kitaev-Preskill (GKP) codes (translation-symmetric) the logical states of the modified binomial code satisfy (Fig. 2c), and two-component cat codes (biased-noise the Knill-Laflamme condition for an extended error set bosonic qubits) (Fig. 2d), and discuss their error- {I,ˆ a,ˆ aˆ†aˆ} wherea ˆ†aˆ describes single dephasing errors in correcting capability against both photon loss and the cavity. This means that the two logical code words dephasing errors. also have the same second moment of the photon number probability distribution besides having the same average photon number. Note that when more moments of the A. Rotation-symmetric codes: Binomial and cat photon number probability distribution are the same for codes the two code words, distinguishing them becomes more difficult for the environment, which enhances the error- The class of rotation-symmetric bosonic codes [40] correcting capability of the code. refers to encodings that remain invariant under a set of Another way to generalize rotation-symmetric codes is discrete rotations in phase space. Encoding schemes with by considering different rotation angles. For instance, rotation-symmetry, such as the cat and binomial codes, we can design codes that are invariant under a 120° are stabilized by a photon number super-parity operator rotation instead of a 180° rotation and are thus stabilized i(2π/N)ˆn ˆ i(2π/3)ˆn Πˆ N = e (N ∈ {2, 3, ···}), which is equivalent to by the super-parity modulo 3 operator Π3 = e . the 360/N° rotation operator. As such, these codes are, By taking advantage of a larger spacing in the photon by design, capable of detecting N − 1 photon loss events. number basis, 120°-rotation-symmetric codes can detect Furthermore, these rotation-symmetric codes can also two-photon loss events as well as single photon loss be made robust against dephasing errors in addition to events, thus providing enhanced protection to the logical photon loss errors [41]. For instance, the kitten code may information. For instance, a variant of the binomial code be modified to have the following logical states: with the logical states 1 √ 1 √ |0 i = (|0i + 3|4i), |0 i = (|0i + 3|6i), L 2 L 2 1 √ 1 √ |1 i = ( 3|2i + |6i). (8) |1 i = ( 3|3i + |9i), (9) L 2 L 2 This version of the binomial code is also stabilized by the is invariant under the 120° rotation, or equivalently, 6

2 has photon numbers that are integer multiples of 3. that scales as e−2|α| [46]. However, an extra error- Moreover, the particular coefficients associated with each correcting mechanism, such as a multi-photon engineered photon number state in the code words, which are dissipation [44], is required to exploit the intrinsic error- derived from the binomial coefficients (hence the name correcting capability of the cat codes against dephasing ‘binomial’ code), ensure that the code satisfies the Knill- errors. Moreover, increasing the number of coherent Laflamme condition for an error set {I,ˆ a,ˆ aˆ2, aˆ†aˆ}. This state components in the cat code introduces further indicates that the code is robust against both single protection [36, 47]. For instance, with six coherent and two-photon loss events, as well as single dephasing state components, the code words become 120°-rotation- errors. The binomial code can be further generalized to symmetric and can thus detect up to two-photon loss protect against higher-order effects of loss and dephasing events. Generalizations of cat codes and analyses of their errors [25] as well as used for autonomous QEC [42]. performance can be found in Ref. [45], while a multi- Cat codes are another important example of rotation- mode generalization is available in Ref. [48]. symmetric bosonic codes. The four-component cat codes (or 4-cat codes), which are composed of four coherent states | ± αi, | ± iαi, are the simplest variants of the B. Translation-symmetric codes: GKP codes cat codes that are robust against photon loss errors [43, 44]. In this encoding, the logical states are defined by Another class of bosonic codes are translation- superpositions of coherent states: symmetric, with the Gottesman-Kitaev-Preskill (GKP) codes [49] being a prominent example. The simplest |0Li ∝ |αi + |iαi + | − αi + | − iαi variant of the GKP codes is the square lattice GKP ∞ 4n code, which encodes a logical qubit in the phase space X α ∝ |4ni, of a harmonic oscillator stabilized by two commuting p(4n)! n=0 displacement operators: |1Li ∝ |αi − |iαi + | − αi − | − iαi √ √ ˆ ˆ ∞ Sq ≡ exp[i2 πqˆ] = D(i 2π), X α4n+2 √ √ ∝ |4n + 2i. (10) ˆ ˆ p Sp ≡ exp[−i2 πpˆ] = D( 2π), (12) n=0 (4n + 2)! whereq ˆ andp ˆ are the position and momentum operators Note that the amplitude of the coherent states |α| respectively. determines the size of the cat code. Similar to the A key motivation for choosing these stabilizers is to kitten code (Eq. (2)), the logical states of the 4-cat circumvent the Heisenberg uncertainty principle, which code have an even number of photons. Thus, the 4- dictates that the position and momentum operators cat code is invariant under the 180° rotation and is able cannot be measured simultaneously as they do not to detect single photon loss events. Here, we reinforce commute. Since the two displacement operators in that the parity (or rotation-symmetry) alone does not Eq. (12) commute with each other, we can measure them ensure the recoverability of logical information against simultaneously. Measuring the displacement operators single photon loss errors, which is only guaranteed when √ √ exp[i2 πqˆ] and√ exp[−i2 πpˆ√] is equivalent to measuring the Knill-Laflamme condition is satisfied for the error set their phases 2 πqˆ and −2 πpˆ modulo 2π, which is {I,ˆ aˆ}. For even parity codes, as explained in Sec II, in turn the√ same as simultaneously measuringq ˆ andp ˆ recoverability requires the two logical states to have the modulo π. Therefore, the square lattice GKP code same number of photons on average. is, by design, capable of addressing two non-commuting For large cat codes with |α|  1, the average 2 quadrature operators by simultaneously measuring them photon number is approximately given byn ¯ ' |α| for within a unit cell of a square lattice. The uncertainty both logical states, and the Knill-Laflamme condition is now lies in the fact that we do not know which unit cell approximately fulfilled. More importantly, the average the state is in. photon numbers of the two logical states are exactly The two logical states of the square lattice GKP code the same for certain values of |α| (also known as ‘sweet are explicitly given by: spots’ [45]) such that: X √ X √ |0 i ∝ |qˆ = (2n) πi ∝ |pˆ = n πi, tan |α|2 = − tanh |α|2. (11) L n∈Z n∈Z X √ X n √ The smallest such |α| is given by |α| = 1.538, which |1Li ∝ |qˆ = (2n + 1) πi ∝ (−1) |pˆ = n πi, corresponds to the average photon numbern ¯ = 2.324. n∈Z n∈Z By increasing the size of the cat codes, we can (13) construct logical states that are robust against both √ single photon loss and dephasing errors. In particular, and satisfyq ˆ =p ˆ = 0 modulo π, thus clearly illus- cat codes with a large average photon number |α|2 satisfy trating that a periodic simultaneous quadrature mea- the Knill-Laflamme condition for the dephasing error surement is indeed possible if the spacing is chosen set {I,ˆ n,ˆ nˆ2, ···} approximately modulo an inaccuracy appropriately. Additionally, the code states are invariant 7 √ under discrete translations of length 2 π in both the is relatively straightforward. Any logical Pauli or position and momentum directions, which makes the Clifford operation on GKP states can be realized by a code symmetric under translations. displacement or Gaussian operation (via a linear drive Conceptually, ideal GKP code states consist of or a bilinear coupling). Moreover, magic states [77] infinitely many infinitely squeezed states where each encoded in the GKP code, which are necessary for component is described by a Dirac Delta function. implementing non-Clifford operations, can be prepared However, precisely implementing these ideal states is with only Gaussian operations and GKP states [50, not feasible in realistic quantum systems. In practice, 78]. Thus, the preparation of code words is the only approximate GKP states can be realized where only required non-Gaussian operation for performing each position or momentum eigenstate is replaced by a universal quantum computation with the GKP code. finitely squeezed state and large position and momentum Non-Clifford operations can be directly enacted on GKP components are suppressed by a Gaussian envelope [49– qubits [49], although the cubic phase gate suggested in 51]. Many proposals to realize approximate GKP states the original proposal has been recently shown to perform in various physical platforms [49, 50, 52–65] and ways poorly for this purpose [79]. to simulate approximate states efficiently [66–71] have Furthermore, GKP states can be defined over lattices been explored in the field. The quality of such states other than the square lattice. For instance, hexagonal can be characterized by the degree of squeezing in GKP√ codes√ can correct√ any shift errors of size less than both the position and momentum quadratures. As (2/ 3)1/2 π/2 ' 1.07 π/2, which is larger than the the squeezing, and hence the average photon number, size of shifts that are correctable by the square lattice increases, the approximate state will converge towards GKP code [49, 75]. A recent work has also explored the an ideal code state. Recently, approximate states of rectangular GKP code [80]. More generally, a multi- squeezing 5.5 – 9.5 dB have been realized in trapped mode GKP code can be defined over any symplectic ion [72, 73] and cQED [74] systems. The cQED lattice [81]. Lastly, the GKP code may also be used realization is discussed further in Sec. V. for building robust quantum repeaters for long-distance With a discrete translation-symmetry, GKP codes are quantum communication [82, 83]. naturally robust against random displacement errors as long as the size of the displacement is small compared to the separation between distinct logical states. For C. Biased-noise bosonic qubits: Two-component cat codes instance, the square lattice GKP code√ is robust against any displacements of size less than π/2 as they can √be identified via the quadrature measurements modulo Recently, there has been growing interest in biased- π and then countered accordingly. For moderately noise bosonic qubits, where a quantum system is squeezed approximate GKP states that contain a small engineered to have one type of error occur with a number of photons, photon loss errors can be decomposed much higher probability than other types of errors. as small shift errors and therefore can be effectively This noise bias can simplify the next layer of error addressed by the code [49, 50]. On the other hand, for correction [84]. For instance, we can design a biased- large approximate GKP states that are highly squeezed, noise code that suppresses bit-flips and then correct the even a tiny fraction of photon loss results in large shift dominant phase-flip errors by using repetition codes [85, errors which cannot be corrected by the code. Thus, 86]. Alternatively, we can also tailor the surface code naively using the standard GKP error correction protocol to leverage the advantages of biased-noise models to to decode does not work for large GKP states under increase fault-tolerance thresholds and reduce resource photon loss errors. Nevertheless, studies have observed overheads [87–90]. that if an optimal decoding scheme is adopted, excellent A promising candidate for biased-noise bosonic qubits performance against photon loss errors can be achieved is the two-component cat code [91–95], or 2-cat code, even with large GKP states [46, 75]. This improvement whose logical code words are given by: happens because photon loss errors can be converted ∞ into random displacement errors via amplification. This X α2n |+Li ∝ |αi + | − αi ∝ |2ni, implies that for highly squeezed large GKP codes, p(2n)! a suitable decoding strategy is to first amplify the n=0 ∞ contracted states and then correct the resulting random X α2n+1 |−Li ∝ |αi − | − αi ∝ p |2n + 1i. (14) shift errors√ by measuring the quadrature operators (2n + 1)! modulo π [76]. We note that at present, no analogous n=0 techniques are known for dephasing errors. Thus, highly If α is large enough, 2-cat codes are capable of correcting squeezed GKP codes are not robust against dephasing dephasing errors. One way to implement 2-cat codes is errors since even a small random rotation can result in by autonomously stabilizing them via an engineered dis- 2 2 large shift errors [50]. sipation of the form κ2D[ˆa − α ] [44]. This engineered While the preparation of GKP states is challeng- two-photon dissipation can exponentially suppress the ing, implementing logical operations on GKP states logical bit-flip error in the code states due to dephasing 8

† in superconducting cavities (κφD[ˆa aˆ]), i.e., photon loss and dephasing where each code is designed to perform well. Note that if the loss and dephasing error 2 −2|α|2 γbit-flip ' 2κφ|α| e if κφ  κ2, (15) probabilities are too high, the quantum capacity [99, 100] of the corresponding quantum channel will vanish. When 2 where |α| is the size of the cat code, κ2 is the two-photon this happens, encoding logical quantum information in a dissipation rate, and κφ is the dephasing rate. reliable way becomes impossible even with an optimal However, the dominant error source in superconduct- QEC code [101–105]. ing cavities is photon loss. While dephasing of the While photon loss is the dominant error channel in cavity does not change the photon number parity, a typical cQED implementations, excitation gain events do single photon loss can directly flip the photon number also occur, although at a much lower rate. In general, parity of the state. In other words, a single photon loss codes robust against photon loss errors tend to be robust maps an even parity state |+Li to an odd parity state against photon gain as well. For rotation-symmetric |−Li, thus causing a phase-flip error. As such, phase- codes, parity (or super-parity) measurements can detect flip errors due to single photon loss cannot be mitigated both photon loss and photon gain events. In the case of by the 2-cat code. Nevertheless, bit-flip errors due to translation-symmetric GKP codes, photon gain can be single photon loss can be suppressed exponentially in converted into a random shift error by applying a suitable the photon number α2 provided that the two-photon photon loss channel and hence be monitored with the dissipation is strong enough. Once bit-flip errors are modular quadrature measurement. On the other hand, countered, the next layer of QEC will only need to tackle since the 2-cat codes are not robust against photon loss, the phase-flip errors, which can be induced by spurious they are consequently susceptible to photon gain too, transitions of the ancilla states. In particular, in the which results in phase-flip errors. Nevertheless, in the stabilized 2-cat codes, thermal excitation of the ancilla presence of engineered dissipation, the stabilized 2-cat during idle times can result in a significant rotation of codes can still preserve their noise bias under both photon the cavity state and complete dephasing of the encoded loss and gain errors. qubit. Currently, such events cannot be corrected by Until now, we have focused only on the intrinsic error- the code and are observed to be a limiting factor for the correcting capability of various bosonic codes without logical lifetime of the encoded qubit [39]. considering practical imperfections. In the following Biased-noise 2-cat qubits can also be sections, we will discuss the implementation of bosonic realized by using an engineered Hamiltonian QEC in cQED systems and examine errors that occur in Hˆ = −K(ˆa† − (α∗)2)(ˆa2 − α2), which has two coherent practical situations. states | ± αi as degenerate eigenstates [96–98]. A crucial general consideration for these biased-noise bosonic codes is that the asymmetry in the noise must be IV. THE CQED HARDWARE FOR BOSONIC preserved during the implementation of gates. The CODES theoretical framework to achieve a universal gate set on the 2-cat codes in a bias-preserving manner has been In the preceding sections, we have discussed the proposed recently [85, 98]. concepts and merits of various bosonic QEC codes. These ideas are brought to reality by developing robust quantum hardware with both coherent harmonic modes D. Comparison of various bosonic codes for loss for encoding information as well as non-linear ancillae and dephasing errors for effective control and tomography of the encoded information. Such systems have been realized with the In summary, rotation-symmetric codes (e.g., 4-cat motional degree of freedom in trapped ions [72, 73], codes and binomial codes) are robust against both electromagnetic fields of microwave [106] and optical photon loss and dephasing errors. Translation-symmetric cavities [107], Rydberg atom arrays [108], and flying codes such as the GKP codes can be made highly photons [109], and there are proposals that use other robust against photon loss errors but are susceptible physical systems [110, 111] as well. Among the various to dephasing errors. Thus, translation-symmetric codes platforms, the cQED architecture consisting of super- are suited for loss dominated systems. In contrast, 2- conducting microwave cavities and Josephson junction- cat codes can correct dephasing errors well if they are based non-linear ancillae have enabled many prominent stabilized by an engineered two-photon dissipation, but experimental milestones towards realizing QEC using are not capable of correcting photon loss errors. Hence, bosonic codes. In this section, we will introduce the 2-cat codes are naturally suited for dephasing dominated key hardware building blocks necessary for the successful systems. However, as discussed in Sec. III C, 2-cat realization of bosonic codes in cQED. codes can also be useful in the loss dominated regime as their large noise bias can simplify any higher-level error correction schemes. In Fig. 2(e), we provide a qualitative schematic that represents different regimes of 9

A. Components of the cQED architecture Hamiltonian: α K Hˆ = ω aˆ†aˆ + ω ˆb†ˆb − χ ˆb†ˆbaˆ†aˆ − ˆb†2ˆb2 − aˆ†2aˆ2, Circuit quantum electrodynamics (cQED) explores a b ab 2 2 light-matter interactions by confining quantized electro- (16) magnetic fields in precisely engineered compositions of superconducting inductors and capacitors [15]. These wherea, ˆ ˆb are the annihilation operators associated superconducting circuit elements can be tailor-made by with cavity and artificial atom respectively, χab is the conventional fabrication techniques [112] and controlled dispersive coupling strength between them, α is the by commercially available microwave electronics and anharmonicity of the artificial atom, and K is the non- dilution refrigerators [113] to access strong coupling linearity of the cavity inherited from the atom. Eq. (16) regimes [114–116]. These characteristics make cQED a further illustrates that the coupling between the two compelling platform for universal quantum computation, modes is symmetric. In other words, the frequency of as noted by several review articles [14, 117–123]. the cavity shifts conditioned on the state of the artificial Quantum devices in the cQED framework are typically atom, and vice versa. built by coupling two components, a linear oscillator Superconducting cavities dispersively coupled to an mode and an anharmonic mode, in different configura- artificial atom constitute a highly versatile tool that can tions [15]. The anharmonic modes are discrete few-level be employed to fulfill many different roles for quantum systems that can be implemented using Josephson junc- information processing. For instance, cavities that are tions. They can be designed to interact with one or more strongly coupled to a transmission line are useful for the harmonic modes, akin to atoms in an optical field in the efficient readout of the quantum state of the artificial cavity QED framework [124]. Unlike naturally occurring atom [114, 142, 143]. Conversely, cavities weakly coupled atoms, the parameters of these anharmonic oscillators, to the environment can be used as quantum memories or artificial atoms, can be precisely engineered. For for storing information coherently [136]. Moreover, instance, they may be made to have a fixed resonance when the linewidth of the cavity is narrow compared to frequency in devices with a single Josephson junction the dispersive shift χab, we can selectively address the or have tunable frequencies by integrating multiple individual energy levels of the cavity via the artificial junctions in the presence of an external magnetic flux. atom [115]. In this case, the artificial atoms act as The lowest two or three energy levels of these artificial non-linear ancillae whose role is to enable conditional atoms, typically in the form of transmons [125, 126], operations and perform efficient tomography of the cavity can be used to effectively encode and process quantum state [144]. This configuration where multi-photon states information, as demonstrated by successful realizations of the cavity encode logical information and the ancilla of various NISQ era [127] processors [128–134]. affords universal control [145] has become an increasingly Linear oscillators are typically realized in cQED by prevalent choice for implementing bosonic QEC. superconducting microwave cavities. Commonly used architectures include coplanar waveguide (CPW) [114], three-dimensional (3D) rectangular [135] and cylindrical B. Coherence of superconducting cavities co-axial [136], and micromachined [137] cavities. These quantum harmonic oscillators have well-defined but Superconducting microwave cavities may be realized degenerate energy transitions. Therefore, to selectively in several geometries, with each having their respective address their transitions, we must introduce some non- advantages. Typically, cavities are constructed in two linearities in them. In cQED, non-linearities are intro- main architectures, which are 2-dimensional (2D) and duced by coupling superconducting cavities to artificial 3-dimensional (3D), based on the dimensionality of the atoms in either the resonant or dispersive regimes. electric field distribution. A third possibility combines Under resonant coupling, the transition frequencies of the advantages of both the 2D and 3D designs to realize the artificial atom and the cavity coincide, which allows a compact and highly coherent ‘2.5D’ cavity structure, the direct exchange of energy from one mode to the for instance, using micromachining techniques [137, 146]. other. In this configuration, cavities can act as an on- In the 2D architecture, such as the coplanar waveguide demand single photon source [138] or a quantum bus (CPW), the cavity is defined by gaps between circuit that mediates operations between two isolated artificial elements printed on a substrate which is typically atoms by sequentially interacting with each of them [139– made from silicon or sapphire (Fig. 3a). In such 141]. Dispersive coupling is achieved by detuning the planar structures, the energy is mostly stored in the frequencies of the cavity, ωa, and the artificial atom, substrate, surfaces, and interfaces, all of which suffer ωb, such that the detuning is much larger than the from losses due to spurious two-level systems in the direct interaction strength between them. In this regime, resonator dielectrics [147–150]. Hence, the Qint of there is no resonant energy exchange between the modes. 2D cavities is currently limited to ∼ 105 − 106 but Instead, the coupling translates into a state-dependent can potentially be improved with more sophisticated frequency shift, which can be described by the following cavity design [151–154], materials selection [155–158], and surface treatment [159–162]. Despite their limited 10 coherence properties, 2D cavities are widely featured have only included demonstrations that are compatible in cQED, and especially in NISQ processors, as they with being coupled to non-linear ancillae in the cQED have a small footprint and a straightforward fabrication architecture. From the figure, the 3D co-axial cavities process. Furthermore, in these processors, these cavities emerge as the leading design currently used to realize are typically used as readout or bus modes, which do not bosonic qubits. require long coherence times. In contrast, 3D cavities (Fig. 3c) achieve a higher Qint of ∼ 107 − 108, at the cost of a much larger footprint V. REALIZATION OF BOSONIC LOGICAL than their 2D counterparts, by storing energy in the QUBITS vacuum between the walls of a superconducting box. Among 3D designs, 3D co-axial cavities machined out of The remarkable improvements in the performance of high-purity aluminum (' 99.999%) have shown a Qint as cQED hardware components highlighted in Sec. IV have high as 1.1 × 108 [163]. This is achieved by significantly made realizing protected logical qubits using bosonic suppressing the dissipation due to current flowing along codes a realistic goal. Studies that encode a single logical the seams of the superconducting walls and imperfections mode and protect it against dominant error channels have on the inner surfaces of the structure. Additionally, been reported for the four-component cat [16], binomial this design is particularly effective as a platform for kitten [173], and square and hexagonal GKP [74] codes. implementing bosonic codes as one or more ancillae and Moreover, robust operations on both single [17, 21, 174] readout modes can be conveniently integrated into the and two bosonic modes [19, 20, 180] have also been platform [164]. explored, thus paving the way towards building a fault- One strategy to take advantage of both the long tolerant universal quantum computer based on bosonic lifetimes of 3D designs and the small footprint and logical qubits. In this section, we highlight recent scalable fabrication of 2D geometries is to construct efforts to implement universal control of as well as error compact 2.5D cavities [137, 146, 164–168]. In particular, correction protocols with bosonic modes. the micromachining technique provides a promising method to fabricate lithographically defined 2.5D cavities etched out of silicon wafers [137]. In this configuration, A. Operations on single bosonic modes the energy can be stored primarily in the vacuum but the depth of the cavity is much smaller compared to For processing quantum information encoded in multi- the other dimensions (Fig. 3b). Here, the key challenge photon states of superconducting cavities, we must be is to realize a high-quality contact between the top able to perform effective operations on and charac- wall and the etched region. With carefully optimized terization of the cavity states. The only operation indium bump-bonding methods, internal quality factors available for a standalone cavity mode is a displacement, † ∗ surpassing 300 million have recently been achieved in Dˆ(α) = eαaˆ −α aˆ, which displaces the position and/or the 2.5D architecture [169]. Moreover, the integration momentum of the harmonic oscillator depending on the of a transmon ancilla in this design has also been value of the complex number α. Real values of α cor- demonstrated [146], thus making these 2.5D cavities respond to pure position displacements, while imaginary promising candidates for realizing large-scale quantum values of α correspond to pure momentum ones [181]. devices based on bosonic modes. Displacements can only result in the generation of Regardless of the architecture, the effective implemen- coherent states from vacuum without the possibility to tation of bosonic QEC schemes requires a careful balance selectively address individual photon number states in between the need for isolation and coherence as well as the cavity. Therefore, non-trivial operations on these the ability to effectively manipulate and characterize the bosonic logical qubits are implemented by dispersively cavity. The various loss channels [175, 176] as well as coupling the bosonic mode to a non-linear ancilla in lossy interfaces [177–179] of superconducting microwave combination with simple displacements. cavities have been extensively studied and their intrinsic A key capability enabled by this natural dispersive coherence properties have been improving significantly coupling is a controlled phase shift (CPS). CPS is a over the last 15 years [176]. In Fig. 3d, we compile a unitary operation that imparts a well-defined ancilla non-exhaustive summary of the internal quality factors state-dependent phase on arbitrary cavity states, and is of cavities demonstrated in various geometries over the inχtˆ governed by UˆCPS(t) = |gihg| ⊗ Iˆ+ |eihe| ⊗ e , where last decade. Besides enhancing the Qint of the cavities, nˆ is the cavity photon number operator, χ is the integrating them with ancillary mode(s) is crucial for dispersive coupling strength, and t is the evolution realizing bosonic logical qubits. However, note that time. With this unitary, we can efficiently implement introducing an ancilla results in the degradation of the conditional phase operations by simply adjusting the Qint, as the best ancilla coherence times (∼ 50 − 100µs) evolution time. In particular, when t = π/χ, all the odd are typically about 10 − 20 times lower than those of the photon number states acquire an overall π-phase while state-of-the-art superconducting cavities. Hence, while the even states get none. This allows us to effectively comparing the performance of the cavities in Fig. 3d, we 11

(a) (c) (d) 9 ~cm 10 2D 1 [170]* 4 [164]* 4 2 [171]* 5 [154] 3DR ~µm 3 [159] -βz 108 4 e 3DC 1 1 [135]* 3 [16]* 2.5D 3 2 2 [172]* 4 [173]* 3 int 107 4 (b) Q 4 1 1 [136]* 3 [174]* 3 1 2 5 2 [21]* 4 [163] λ / 4 106 2 3 ~ cm 1 [167]* 3 [168] ~100μm 2 2 [146]* 4 [169] 1 105 ~cm 2010 2015 2020 Year

FIG. 3. Designs of superconducting cavities and their coherence (a)-(c) Illustrations of 2D CPW, 2.5D micromachined, and 3D co-axial cavities respectively. (d) A selected set of Qint for four commonly used cavity designs in cQED extracted from the literature. Overall, the coherence properties of superconducting cavities have been steadily increasing over the last decade, with the 3D co-axial cavities currently being the main architecture for realizing bosonic QEC. The references for each cavity design, in chronological order, are: 2D cavities: Refs. [170], [171], [159], [164], [154]; 3D rectangular cavities (3DR): Refs. [135], [172], [16], [173]; 3D cylindrical co-axial cavities (3DC): Refs. [136], [21], [174], [163]; 2.5D cavities: Refs. [167], [146], [168], [169]. We have highlighted studies that successfully integrated one or more non-linear ancillae with an asterisk (*). map the photon number parity of the bosonic mode onto Wigner tomography, we can reconstruct the full density the state of the ancilla (Fig. 4a). matrix and characterize the action enacted on the cavity Note that the time required for the parity-mapping states in either the Pauli Transfer Matrix [185] or the operation scales inversely with the dispersive coupling process matrix [24]. strength, χ. Naively, one might want to minimize the Another crucial operation arising from the natural operation time by engineering a large χ. However, dispersive coupling is the non-linear selective number- increasing χ can also result in a stronger inherited non- dependent arbitrary phase (SNAP) gate [145, 186]. A iθn(|nihn|) linearity in the cavity (also known as the Kerr effect), SNAP gate (Fig. 4b), defined as Sˆn(θn) = e , which distorts the encoded information. Therefore, the selectively imparts a phase θn to the number state coupling strength between the cavity and its ancilla, |ni. Due to the energy-preserving nature of this as well as the type of the non-linear ancillary mode, operation, we can simultaneously perform Sˆn(θn) on such as the transmon or the SNAIL design [182], are multiple number states. By numerically optimizing the usually carefully optimized for each system to produce linear displacement gates and the phases applied to each the desired Hamiltonian configuration. photon number state in the cavity, we can effectively ˆ With UCPS, we can deterministically create complex cancel out the undesired Fock components via destructive bosonic encodings using analytically designed protocols interference to obtain the intended target state. such as the qcMAP operation, which maps an arbitrary While schemes like qcMAP and SNAP are sufficient qubit state onto a superposition of coherent states in to realize universal control on the cavity state, they the cavity [183]. This scheme is useful for preparing quickly become impractical for handling more complex 4-cat states in both single [172] and multiple [184] bosonic states. For instance, an operation on n photons ˆ 2 cavities. Furthermore, UCPS(π/χ) is also employed as requires O(n ) gates using the SNAP protocol. To the parity-mapping operation which is crucial for the address this challenge, a fully numerical approach using characterization and tomography of the encoded bosonic optimal control theory (OCT) has been developed and qubits and the engineered gates on these qubits. widely adopted in recent years. The OCT framework In general, the quantum states encoded in a cav- provides an efficient general-purpose tool to implement ity can be fully characterized by probing the cav- arbitrary operations. In particular, the Gradient Ascent ity’s quasi-probability distributions, which is commonly Pulse Engineering (GRAPE) method [187, 188] has been achieved by performing Wigner tomography. The successfully deployed in other physical systems [189, 190] Wigner function can be defined as the expectation to implement robust quantum control. By constructing value of the displaced photon number parity operator, an accurate model of the time-dependent Hamiltonian 2 ˆ † ˆ ˆ W (β) = π Tr[D(β) ρD(β)P ]. In cQED, the Wigner of the system in the presence of arbitrary control fields, functions of arbitrary quantum states can be measured we can apply this technique to cQED systems to realize precisely with a well-defined sequence that uses only high-fidelity universal gate sets on any bosonic qubit the cavity displacement, ancilla rotation, and controlled encoded in cavities, as demonstrated in Ref. [17]. A phase shift operations [15]. From the results of the typical example of a set of pulses obtained through the 12

(a) (b)

cavity 0 Cπ 0

ancilla 0 Rπ/2 Rπ/2 cavity 1

e e e 2

......

g g g 0 1 2 ancilla 0 Rπ Rπ Rπ (c) e e e cavity 0

ancilla 0 g g g

FIG. 4. Operations on single bosonic modes. (a) The sequence for mapping the photon number parity of a bosonic mode to the states of its ancilla. (b) Creation of arbitrary quantum states via repeat SNAP operations on the bosonic mode. (c) Implementing universal control on a single bosonic mode by concurrently driving the cavity and its ancilla with numerically optimized pulses. gradient-based OCT framework is shown in (Fig. 4c). gate enacted on two bosonic qubits [195]. This gate is More recently, various concepts from classical machine facilitated by a parametrically-driven sideband transition learning, such as automatic differentiation [191] and between the ancilla and the control mode together with reinforcement learning [192, 193], have been applied to a carefully chosen conditional phase gate between the enhance the efficiency of the numerical optimization. ancilla and the target cavity. By achieving a gate fidelity Crucially, the success of these techniques does not only above 98%, this study showcases the potential of such rely on the robustness of the algorithms, but also depends engineered quantum gates between bosonic modes. More on the choice of boundary conditions. Knowledge recently, a controlled-phase gate has been demonstrated of these boundary conditions requires comprehensive between two binomial logical qubits in Ref. [20]. Here, investigations of the physics of the quantum system and the microwave drives are tailored to induce a geometric the practical constraints of the control and measurement phase that depends on the joint state of the two bosonic apparatus. modes. However, these two types of operations are both customized for a selective set of code words and do not yet generalize readily to other bosonic encoding schemes. B. Operations on multiple bosonic modes A code-independent coupling mechanism between two otherwise isolated bosonic modes is a crucial ingredient Apart from robust single-mode operations, universal for realizing universal control on these logical qubits. quantum computation using bosonic qubits also requires The isolated cavities must be sufficiently detuned from at least one entangling gate between two modes. each other to ensure the coherence of each mode and the Realizing such an operation can be challenging due to absence of undesired cross-talk. Ref. [180] demonstrated the lack of a natural coupling between cavities. Moreover, how the four-wave mixing process in a Josephson junction the individual coherence of each bosonic qubit must be can provide a frequency converting bilinear coupling of iϕ ˆ† −iϕ †ˆ ˆ maintained while maximizing the rate of interactions the form Hint(t)/~ = g(t)(e aˆb + e aˆ b). Here,a, ˆ b between them. One promising strategy to tackle this are the annihilation operations associated with each of issue is to use the non-linear frequency conversion the cavities, the time-dependent coefficient g(t) is the capability of the Josephson junction to provide a driven coupling strength, and ϕ is the relative phase between coupling between two otherwise isolated cavities [194]. the two microwave drives. The coefficient g(t) depends Such operations are fully activated by external microwave on the effective amplitudes of the drives, which satisfy drives which can be tuned on and off on-demand without the frequency matching condition |ω2 − ω1| = |ωa − ωb|. modifying the hardware. This arrangement ensures that Most notably, this coupling can be programmed to the individual bosonic modes remain well-isolated during implement an identity, a 50:50 beamsplitter, or a full idle times and undergo the engineered interaction only SWAP operation between the stationary microwave fields when an operation is enacted. in the cavities by simply adjusting the duration of the Using this strategy, a CNOT gate was the first logical evolution. Such an engineered coupling provides a pow- 13 erful tool for implementing programmable interferometry the break-even point, we can be confident that the between cavity states [180], which is a key building block chosen QEC protocol does not introduce more errors for realizing various continuous-variable information into the system, thus attaining an improvement in the processing tasks such as boson sampling [196], simulation lifetime of the logical qubit. Till date, three studies of vibrational quantum dynamics of molecules [197–200], have approached [74, 173] or achieved [16] the ‘break- and distributed quantum sensing [201–204]. even point’ for QEC with bosonic codes in cQED devices Moreover, this bilinear coupling is also a valuable without any post-selection. resource for enacting gates on two logical elements Broadly, QEC schemes fall into three categories based encoded in GKP states [49]. As mentioned in Sec. III, on how they afford protection to the logical qubit. In GKP encodings rely on the non-linearity of the code active QEC, error syndromes are repeatedly measured words and only require linear or bilinear operations for during the state evolution of the logical qubit and any universal control [78]. Therefore, this engineered bilinear errors detected are subsequently corrected based on the coupling provides a simple and effective strategy for measurement outcomes. In autonomous QEC, errors implementing a deterministic entangling operation for are removed by tailored dissipation or by coupling to the GKP code. an auxiliary system, without repeatedly probing the For other bosonic codes, this bilinear interaction is logical qubit. In passive QEC, the logical information not alone sufficient to generate a universal gate set, is intrinsically protected from decoherence because of which requires at least one entangling gate. In this specifically designed physical symmetries [207]. In case, the exponential SWAP (eSWAP) operation can be this discussion, we have opted to distinguish between designed to provide deterministic and code-independent autonomous and passive QEC to highlight potential entanglement [205]. The eSWAP operation, akin to differences in system design. In the autonomous an exchange operation between spins, implements a approach, the environment is intentionally engineered to programmable unitary of the form: suppress or mitigate errors. Whereas in the passive case, the system Hamiltonian itself is tailored to be robust ˆ ˆ U(θ) = cos (θ)I + sin (θ)SWAP, (17) against certain errors, which often involves constructing a unique physical element in the hardware [208–210]. where θ is the rotation angle on the ancilla and Iˆ is the Typically, active QEC requires robust measurements identity operation. Intuitively, this unitary implements of the error syndrome and real-time feedback. In a weighted superposition of the identity and SWAP superconducting cavities, the dominant source of error is operations between two bosonic modes regardless of single photon loss. For encoding schemes with rotational their specific encodings. The eSWAP unitary has been symmetries, such as the cat and binomial codes, single realized in Ref. [19] between two bosonic modes housed photon loss results in a flip in the parity of the code in 3D co-axial cavities bridged by an ancilla. In this words. Therefore, measuring the parity operator tells us demonstration, an additional ancilla is introduced to one whether a photon jump has taken place, thus allowing us of the cavity modes and the resultant dispersive coupling to detect the error syndrome of these logical qubits [211]. is used to enact a Uˆ operation to provide the tunable CPS In Ref. [16], which corrected a four-component cat state rotation necessary for the eSWAP unitary. The eSWAP under photon loss (Fig. 5), the correct logical state unitary is then enacted on several encoding schemes in was recovered by making appropriate adjustments in the Fock, coherent, and binomial basis. The availability the decoding step based on the number of parity flips of such a deterministic and code-independent entangling detected with real-time feedback. The corrected logical operation is a crucial step towards universal quantum qubit showed an improved lifetime compared to both computation using bosonic logical qubits. A recent study the uncorrected state and the |0, 1i Fock state encoding, has shown that universal control and operations on tens thereby achieving the break-even point. However, this of bosonic qubits can be achieved in a novel architecture error correction strategy is not suitable for protecting comprising a single transmon coupled simultaneously to bosonic qubits for timescales exceeding the intrinsic a multi-mode superconducting cavity [206]. cavity lifetime as the loss of energy from the system is accounted for but not physically rectified. Thus, such C. Implementations of QEC energy attenuation should be physically compensated to significantly surpass the break-even point. In contrast, Ref. [173] admits a photon pumping In general, implementations of QEC with bosonic operation to achieve QEC on a logical qubit encoded in codes suffer from an initial rise in error rates due to the binomial code. The errors are detected by photon the presence of and need to control multi-photon states. number parity measurements, as in Ref. [16]. The errors Hence, the main experimental challenge is to achieve occurring on the logical state are then corrected by an enhancement in the lifetime of the encoded qubit an appropriate recovery operation as soon as they are despite this initial penalty. In the context of bosonic detected. An approximated recovery operation is still codes, the break-even point is defined relative to the required in case no errors are detected as the system |0, 1i Fock states, which are the longest-lived physical evolves under the no-parity-jump operator [25]. In this elements in the cavity without error correction. Beyond 14

0L 0L interactions with the environment. In Refs. [212, 213], the desired dissipation process subtracts photons in pairs from the bosonic mode via four-wave mixing in the Josephson junction of the ancilla. This process defines a stabilized manifold of steady states spanned by superpositions of coherent states (two-component cat 1L 1L a a states). The relative strength between the engineered two-photon dissipation and the intrinsic single photon even odd loss of the cavity defines the separation of the two parity parity components of the cat state, which determines the extent a a of protection against bit-flip errors. In a more recent 0L 0L study, significant improvements in the suppression of bit- flip errors were demonstrated by using a tailored multi- junction ancilla to enact the two-photon dissipation process [39]. Similarly, an engineered four-photon dissipation achieved by an eight-wave mixing process, can protect 1L 1L four-component cat codes against dephasing [214]. How- ever, this strategy only accounts for dephasing errors and FIG. 5. Four-component cat code under photon loss. Every photon loss event (ˆa) changes not only the parity does not recover logical errors caused by photon loss. In of the basis states, but also changes the phase relationship contrast, in another autonomous QEC implementation between them. The encoded state cycles between the even with a truncated 4-cat state [215], a photon was added (logical) and odd (error) parity subspaces, while also rotating to the cavity upon detecting a photon loss using a about the Z-axis by π/2. The decoding sequence must take synthetic dissipation operator. This work effectively both these effects into account to correctly recover the logical corrects photon loss events but does not account for information. dephasing errors. Similarly, another study performed autonomous QEC against photon loss with the kitten code by triggering an engineered jump operation [174]. experiment, the lifetime of the logical qubit was greater This operation recovered the code states whenever the than that of the uncorrected binomial code state, but was system entered an error state without the need for marginally below that of the |0, 1i Fock state. any external probing. Although these current state-of- In addition to the cat and binomial encodings, the-art demonstrations of autonomous QEC techniques active QEC has also been recently demonstrated on only address one type of errors, they provide convincing a high-quality GKP state stored in a superconducting evidence for the viability of the respective bosonic cavity [74]. As explained in Sec. III B, errors occurring encoding schemes, thus paving the way for higher- on a moderately squeezed GKP state simply manifest order QEC protocols that can protect logical information as displacements of the cavity state, and are revealed against both photon loss and dephasing errors. and mitigated by measuring the displacement stabilizers. Finally, QEC with bosonic codes can also be realized While mitigating errors is relatively straightforward, using a passive approach, for instance, by designing experimental challenges in implementing the GKP code logical qubits with highly biased-noise channels to lie in preparing finitely squeezed approximate GKP provide intrinsic protection without the need for probing states and performing modulo quadrature measurement any error syndromes. Very recently, two studies [39, of stabilizers. The QEC protocol used in Ref. [74] 209] verified that the bias between phase and bit-flip suppresses all logical errors for a GKP state by al- errors increases exponentially with α, the size of the ternating between two peak-sharpening and envelope- coherent state components of the 2-cat code. This strong trimming rounds, each consisting of different condi- asymmetry between the different types of errors can be tional displacements on the cavity. The magnitude exploited to significantly reduce the hardware overhead of these conditional displacements is dependent on the for fault-tolerant quantum computation [87–89]. measured state of the ancilla onto which displacement stabilizers are mapped. This sharpen-and-trim technique can be generalized based on the second-order Trotter VI. TOWARDS FAULT-TOLERANT QUANTUM formula [65]. COMPUTATION WITH BOSONIC MODES An alternative strategy for QEC that does not require real-time feedback based on measurement outcomes in- A crucial consideration for a successful QEC imple- volves engineering the dynamics of the system to correct mentation is that even the attempts to correct errors errors autonomously. In practice, autonomous QEC can can themselves be erroneous. Thus, a key requirement be achieved by introducing a tailored dissipation [212] for realizing ‘fault-tolerant’ quantum computation is to such that the logical subspace is stabilized via controlled carefully design error correction circuits such that faults 15 that occur during the execution of a QEC protocol are be fault-tolerant. In performing such a concatenation also tolerated [216]. A simple example of fault-tolerant of codes, additional analog information acquired during circuit design is the transversal implementation of logical the error correction protocol can significantly boost operations in the conventional paradigm of quantum the performance of a next-level QEC scheme based on error correction with multiple two-level systems. For multiple error-corrected GKP qubits [220, 221]. For instance, by implementing logical gates transversally for instance, studies have shown that the fault-tolerance distance-3 codes that can correct any single qubit errors, thresholds of the circuit-based toric code and the surface one can ensure that a single fault in the circuit induces at code can be increased by using the additional analog most a single qubit error in any logical code block which information [222, 223]. Another way to realize fault- remains correctable by the code [217]. tolerant quantum computing with the GKP code is to One key challenge in achieving fault-tolerant opera- combine the GKP code with the idea of fault-tolerant tions on bosonic qubits in the cat and binomial encodings measurement-based quantum computation using cluster is the propagation of uncorrectable errors due to ancilla states, for which the additional analog information also decoherence [16, 173]. The ancilla, often in the form plays an important role [224–230]. From the circuit- of a transmon, is employed during state preparation, based computing studies, we know that fault-tolerant gates, measurements, and error correction on the encoded quantum computing with the GKP code is possible if the qubit via the natural dispersive coupling. Spontaneous squeezing of the GKP states is larger than 11dB. Whereas emission of the ancilla during these processes induces from the measurement-based computing scenarios, a dephasing on the logical qubit, thus irreversibly corrupt- more favorable threshold value of 8dB can be obtained ing the encoded information. Furthermore, two-photon by taking advantage of post-selection. However, as loss and Kerr effects also limit the effectiveness of these the measurement-based schemes use post-selection, they QEC implementations. One technique used to mitigate require a higher resource overhead. Recently, Ref. [59] these imperfections is to optimize the cadence of the made the phase estimation protocol to prepare GKP parity measurements [16]. Frequent parity measurements states in Ref. [50] fault-tolerant by using the concept of allow more accurate tracking of the error syndrome but flag qubits [231, 232]. increase the likelihood of the ancilla relaxing during There have also been general studies on the fault- measurement and thereby dephasing the cavity. On the tolerance of biased-noise qubits [84, 87–89]. These other hand, a larger interval between two measurements studies highlight the potential of achieving higher fault- not only increases the probability of two-photon loss, tolerance thresholds and lower resource overheads by but also results in a higher phase accumulation due to taking advantage of the noise bias. Refs. [85, 86] self-Kerr interactions. As the self-Kerr operator does suggested pieceable fault-tolerant [233] Toffoli circuits not commute with the annihilation operator, photon while Ref. [98] offered a magic state distillation scheme jumps cause cavity dephasing. In Ref. [16], the tailored to biased-noise models [234] as a means of phase accumulated due to the self-Kerr effect was first achieving fault-tolerant universal quantum computation estimated and then accounted for by adjusting the phase with biased-noise 2-cat qubits. A detailed roadmap of subsequent drive pulses. for realizing biased-noise 2-cat qubits using acoustic Several approaches have been devised to reduce the nanomechanical resonators as well as alternative fault- propagation of uncorrectable errors from the ancilla tolerant universal quantum computation schemes were to the encoded information. In Ref. [22], the third recently proposed [90]. energy level of the ancilla is employed in the parity- mapping process to suppress the dephasing of the bosonic mode. A similar strategy can also be applied to VII. OUTLOOK AND PERSPECTIVES general cavity operations, as shown in Ref. [21]. In these schemes, additional levels of the ancilla as well Continuous variables have long been recognized as as carefully engineered cavity-ancilla interactions are promising candidates for the efficient encoding and introduced to prevent incoherent evolution of the bosonic processing of quantum information [235]. Bosonic codes mode in the event of ancilla decay [218]. Alternatively, leverage the advantages afforded by continuous variable error-transparent gates have been realized for binomial quantum devices to realize hardware-efficient QEC and encodings [174] by cleverly engineering the evolution of pave the way towards the eventual realization of a the bosonic qubit in both the code and error space. fault-tolerant quantum computer. In this review, we For GKP codes, early discussions on fault-tolerance have presented some of the most compelling results were focused on whether a small shift error occurring in that attest to the viability of the cQED platform. In the middle of an error correction circuit remains small particular, many milestones, from reaching the break- throughout the entire circuit [49, 219]. Recently, nu- even point to implementing logical operations between merous studies have approached fault-tolerant quantum bosonic logical qubits, have thus far been demonstrated computation with the GKP code by combining the GKP using 3D cQED devices. An important next milestone code with conventional error correction schemes based would be to demonstrate the break-even point for logical on multiple two-level systems that are already shown to operations, i.e., to achieve gate fidelities with error- 16 correctable bosonic qubits that are higher than the of these bosonic quantum systems on the hardware level, highest gate fidelity attainable with the best physical we must also expand our technological repertoire to element without error correction. Achieving the break- effectively control and characterize these more complex even point in both QEC and single or two-mode logical devices. This expansion may involve, for instance, operations forms a critical foundation upon which fault- the development of more sophisticated measurement tolerant gates and algorithms can be constructed. electronics and protocols as well as new theoretical We further emphasize that the techniques and method- frameworks to capture the signatures of multipartite ologies developed in these 3D cQED implementations of entanglement. The small-scale bosonic devices developed bosonic codes are agnostic to the hardware architecture. so far provide an indispensable platform to test and refine They can therefore be readily adapted to more compact these crucial elements. designs, such as planar or 2.5D devices, as their perfor- In recent years, cQED devices have become a mance improves with further developments in materials workhorse for the implementations of quantum error engineering and new fabrication techniques [158]. correction and quantum information processing. Along- Crucially, encoding logical information in bosonic side the developments in cQED, we are also witnessing codes is naturally compatible with a modular approach remarkable progress with other platforms based on to scalability. In the modular architecture, each logical trapped-ions [242, 243], silicon spins [244, 245], neutral element can be encoded, protected, and optimized atoms [246], etc, from both academic and industrial individually, and then connected to other elements initiatives. While the ultimate hardware to achieve fault- via on-demand communication channels [236–238]. In tolerance may not consist of millions of 3D cavities, contrast to directly wiring up an increasing number of the insights gained from and the technical achievements physical elements, modularity has the benefit of reduced demonstrated by the results summarized in this article cross-talk, robustness against local failure modes, and will serve as crucial building blocks for the eventual enhanced re-configurability. Encoding and manipulating realization of a scalable and robust universal quantum individual logical qubits encoded in cavities is a crucial computer. primitive for eventually realizing a large-scale modular ACKNOWLEDGMENTS quantum device. In addition, recent experiments have also demonstrated other key elements of a modular The authors thank S. M. Girvin and L. Sun for fruitful architecture, including programmable quantum com- discussions and comments on the manuscript. Yvonne Y. munication and entanglement generation between two Gao acknowledges funding from the Singapore National distant bosonic modes [239–241] as well as a teleported Research Foundation Fellowship (Class of 2020). This CNOT gate [18]. work is also supported by the Ministry of Education in As we continue to improve the performance and scale Singapore.

[1] P. W. Shor. Scheme for reducing decoherence in quan- [6] T. H. Taminiau, J. Cramer, T. van der Sar, V. V. tum computer memory. Phys. Rev. A, 52(4):R2493– Dobrovitski, and R. Hanson. Universal control and R2496, 1995. https://doi.org/10.1103/PhysRevA. error correction in multi-qubit spin registers in diamond. 52.R2493. Nature Nanotech, 9(3):171–176, 2014. https://doi. [2] A. M. Steane. Error correcting codes in quantum theory. org/10.1038/nnano.2014.2. Phys. Rev. Lett., 77(5):793–797, 1996. https://doi. [7] G. Waldherr, Y. Wang, S. Zaiser, M. Jamali, T. Schulte- org/10.1103/PhysRevLett.77.793. Herbr¨uggen, H. Abe, T. Ohshima, J. Isoya, J. F. [3] J. Preskill. Reliable quantum computers. Proceedings of Du, P. Neumann, and J. Wrachtrup. Quantum error the Royal Society of London. Series A: Mathematical, correction in a solid-state hybrid spin register. Nature, Physical, and Engineering Sciences, 454(1969):385– 506(7487):204–207, 2014. https://doi.org/10.1038/ 410, 1998. https://doi.org/10.1098/rspa.1998. nature12919. 0167. [8] N. M. Linke, M. Gutierrez, K. A. Landsman, C. Figgatt, [4] J. Chiaverini, D. Leibfried, T. Schaetz, M. D. Barrett, S. Debnath, K. R. Brown, and C. Monroe. Fault- R. B. Blakestad, J. Britton, W. M. Itano, J. D. Jost, tolerant quantum error detection. Science Advances, E. Knill, C. Langer, R. Ozeri, and D. J. Wineland. 3(10):e1701074, 2017. https://doi.org/10.1126/ Realization of quantum error correction. Nature, sciadv.1701074. 432(7017):602–605, 2004. https://doi.org/10.1038/ [9] C. K. Andersen, A. Remm, S. Lazar, S. Krinner, nature03074. N. Lacroix, G. J. Norris, M. Gabureac, C. Eichler, [5] M. D. Reed, L. DiCarlo, S. E. Nigg, L. Sun, L. Frunzio, and A. Wallraff. Repeated quantum error detection S. M. Girvin, and R. J. Schoelkopf. Realization of three- in a surface code. Nat. Phys., 16(8):204–207, 2020. qubit quantum error correction with superconducting https://doi.org/10.1038/s41567-020-0920-y. circuits. Nature, 482(7385):382–385, 2012. https:// [10] A. Erhard, H. P. Nautrup, M. Meth, L. Postler, doi.org/10.1038/nature10786. R. Stricker, M. Stadler, V. Negnevitsky, M. Ringbauer, P. Schindler, H. J. Briegel, R. Blatt, N. Friis, and 17

T. Monz. Entangling logical qubits with lattice surgery. University Press, 2000. https://doi.org/10.1017/ Nature, 589(7841):220–224, 2021. https://doi.org/ CBO9780511976667. 10.1038/s41586-020-03079-6. [25] M. H. Michael, M. Silveri, R. T. Brierley, V. V. Albert, [11] S. L. Braunstein. Quantum error correction for commu- J. Salmilehto, L. Jiang, and S. M. Girvin. New class nication with linear optics. Nature, 394(6688):47–49, of quantum error-correcting codes for a bosonic mode. 1998. https://doi.org/10.1038/27850. Phys. Rev. X, 6(3):031006, 2016. https://doi.org/10. [12] S. D. Bartlett and B. C. Sanders. Universal continuous- 1103/PhysRevX.6.031006. variable quantum computation: requirement of op- [26] D. W. Leung, M. A. Nielsen, I. L. Chuang, and tical nonlinearity for photon counting. Phys. Rev. Y. Yamamoto. Approximate quantum error correction A, 65(4):042304, 2002. https://doi.org/10.1103/ can lead to better codes. Phys. Rev. A, 56(4):2567– PhysRevA.65.042304. 2573, 1997. https://doi.org/10.1103/PhysRevA.56. [13] S. D. Bartlett, H. de Guise, and B. C. Sanders. 2567. Quantum encodings in spin systems and harmonic [27] S. M. Girvin. Quantum superconducting circuits and oscillators. Phys. Rev. A, 65(5):052316, 2002. https: error correction. In Les Houches Summer School //doi.org/10.1103/PhysRevA.65.052316. “Quantum information machines”, 2019. https:// [14] M. H. Devoret and R. J. Schoelkopf. Superconducting youtu.be/Na_DB5o4QNY. circuits for quantum information: an outlook. Science, [28] S. Sheldon, E. Magesan, J. M. Chow, and J. M. 339(6124):1169–1174, 2013. https://doi.org/10. Gambetta. Procedure for systematically tuning up 1126/science.1231930. cross-talk in the cross-resonance gate. Phys. Rev. [15] A. Blais, A. L. Grimsmo, S. M. Girvin, and A. Wallraff. A, 93(6):060302, 2016. https://doi.org/10.1103/ Circuit quantum electrodynamics. arXiv:2005.12667 PhysRevA.93.060302. [quant-ph], 2020. https://arxiv.org/abs/2005.12667. [29] M. A. Rol, F. Battistel, F. K. Malinowski, C. C. [16] N. Ofek, A. Petrenko, R. Heeres, P. Reinhold, Z. Legh- Bultink, B. M. Tarasinski, R. Vollmer, N. Haider, tas, B. Vlastakis, Y. Liu, L. Frunzio, S. M. Girvin, N. Muthusubramanian, A. Bruno, B. M. Terhal, and L. Jiang, M. Mirrahimi, M. H. Devoret, and R. J. L. DiCarlo. Fast, high-fidelity conditional-phase gate Schoelkopf. Extending the lifetime of a quantum exploiting leakage interference in weakly anharmonic su- bit with error correction in superconducting circuits. perconducting qubits. Phys. Rev. Lett., 123(12):120502, Nature, 536(7617):441–445, 2016. https://doi.org/ 2019. https://doi.org/10.1103/PhysRevLett.123. 10.1038/nature18949. 120502. [17] R. W. Heeres, P. Reinhold, N. Ofek, L. Frunzio, [30] Y. Ouyang. Permutation-invariant quantum codes. L. Jiang, M. H. Devoret, and R. J. Schoelkopf. Phys. Rev. A, 90(6):062317, 2014. https://doi.org/ Implementing a universal gate set on a logical qubit 10.1103/PhysRevA.90.062317. encoded in an oscillator. Nat. Commun., 8(94), 2017. [31] Y. Ouyang and J. Fitzsimons. Permutation-invariant https://doi.org/10.1038/s41467-017-00045-1. codes encoding more than one qubit. Phys. Rev. [18] K. S. Chou, J. Z. Blumoff, C. S. Wang, P. C. A, 93(4):042340, 2016. https://doi.org/10.1103/ Reinhold, C. J. Axline, Y. Y. Gao, L. Frunzio, M. H. PhysRevA.93.042340. Devoret, L. Jiang, and R. J. Schoelkopf. Deterministic [32] Y. Ouyang. Permutation-invariant qudit codes from teleportation of a quantum gate between two logical polynomials. Linear Algebra and its Applications, qubits. Nature, 561(7723):368–373, 2018. https:// 532:43–59, 2017. https://doi.org/10.1016/j.laa. doi.org/10.1038/s41586-018-0470-y. 2017.06.031. [19] Y. Y. Gao, B. J. Lester, K. S. Chou, L. Frunzio, M. H. [33] Y. Ouyang and R. Chao. Permutation-invariant Devoret, L. Jiang, S. M. Girvin, and R. J. Schoelkopf. constant-excitation quantum codes for amplitude damp- Entanglement of bosonic modes through an engineered ing. IEEE Transactions on Information Theory, exchange interaction. Nature, 566(7745):509–512, 2019. 66(5):2921–2933, 2020. https://doi.org/10.1109/ https://doi.org/10.1038/s41586-019-0970-4. TIT.2019.2956142. [20] Y. Xu, Y. Ma, W. Cai, X. Mu, W. Dai, W. Wang, L. Hu, [34] W. Wasilewski and K. Banaszek. Protecting an X. Li, J. Han, H. Wang, Y. P. Song, Z. B. Yang, S. B. optical qubit against photon loss. Phys. Rev. Zheng, and L. Sun. Demonstration of controlled-phase A, 75(4):042316, 2007. https://doi.org/10.1103/ gates between two error-correctable photonic qubits. PhysRevA.75.042316. Phys. Rev. Lett., 124(12):120501, 2020. https://doi. [35] E. Kapit. Hardware-efficient and fully autonomous org/10.1103/PhysRevLett.124.120501. quantum error correction in superconducting circuits. [21] P. Reinhold, S. Rosenblum, W. Ma, L. Frunzio, L. Jiang, Phys. Rev. Lett., 116(15):150501, 2016. https://doi. and R. J. Schoelkopf. Error-corrected gates on an org/10.1103/PhysRevLett.116.150501. encoded qubit. Nat. Phys., 16(8):822–826, 2020. https: [36] M. Bergmann and P. van Loock. Quantum error //doi.org/10.1038/s41567-020-0931-8. correction against photon loss using multicomponent [22] S. Rosenblum, P. Reinhold, M. Mirrahimi, L. Jiang, cat states. Phys. Rev. A, 94(4):042332, 2016. https: L. Frunzio, and R. J. Schoelkopf. Fault-tolerant detec- //doi.org/10.1103/PhysRevA.94.042332. tion of a quantum error. Science, 361(6399):266–270, [37] E. Kapit. Error-transparent quantum gates for 2018. https://doi.org/10.1126/science.aat3996. small logical qubit architectures. Phys. Rev. Lett., [23] E. Knill and R. Laflamme. Theory of quantum error- 120(5):050503, 2018. https://doi.org/10.1103/ correcting codes. Phys. Rev. A, 55(2):900–911, 1997. PhysRevLett.120.050503. https://doi.org/10.1103/PhysRevA.55.900. [38] M. Reagor, H. Paik, G. Catelani, L. Sun, C. Axline, [24] M. A. Nielsen and I. L. Chuang. Quantum Com- E. Holland, I. M. Pop, N. A. Masluk, T. Brecht, putation and Quantum Information. Cambridge L. Frunzio, M. H. Devoret, L. Glazman, and R. J. 18

Schoelkopf. Reaching 10 ms single photon lifetimes for [51] T. Matsuura, H. Yamasaki, and M. Koashi. Equivalence superconducting aluminum cavities. Appl. Phys. Lett., of approximate Gottesman-Kitaev-Preskill codes. Phys. 102(19):192604, 2013. https://doi.org/10.1063/1. Rev. A, 102(3):032408, 2020. https://doi.org/10. 4807015. 1103/PhysRevA.102.032408. [39] R. Lescanne, M. Villiers, T. Peronnin, A. Sarlette, [52] B. C. Travaglione and G. J. Milburn. Prepar- M. Delbecq, B. Huard, T. Kontos, M. Mirrahimi, ing encoded states in an oscillator. Phys. Rev. and Z. Leghtas. Exponential suppression of bit- A, 66(5):052322, 2002. https://doi.org/10.1103/ flips in a qubit encoded in an oscillator. Nat. PhysRevA.66.052322. Phys., 16(5):509–513, 2020. https://doi.org/10. [53] S. Pirandola, S. Mancini, S. Vitali, and P. Tombesi. 1038/s41567-020-0824-x. Constructing finite-dimensional codes with optical con- [40] A. L. Grimsmo, J. Combes, and B. Q. Baragiola. tinuous variables. EPL, 68(3):323, 2004. https://doi. Quantum computing with rotation-symmetric bosonic org/10.1209/epl/i2004-10203-9. codes. Phys. Rev. X, 10(1):011058, 2020. https: [54] S. Pirandola, S. Mancini, D. Vitali, and P. Tombesi. //doi.org/10.1103/PhysRevX.10.011058. Generating continuous variable quantum codewords in [41] Y. Ouyang and E. Campbell. Trade-offs on number the near-field atomic lithography. J. Phys. B: At. Mol. and phase shift resilience in bosonic quantum codes. Opt. Phys., 39(4):997, 2006. https://doi.org/10. arXiv:2008.12576 [quant-ph], 2020. https://arxiv. 1088/0953-4075/39/4/023. org/abs/2008.12576. [55] H. M. Vasconcelos, L. Sanz, and S. Glancy. All- [42] J. M. Lihm, K. Noh, and U. R. Fischer. Implementation- optical generation of states for “Encoding a qubit in independent sufficient condition of the Knill-Laflamme an oscillator”. Opt. Lett., 35(19):3261–3263, 2010. type for the autonomous protection of logical qu- https://doi.org/10.1364/OL.35.003261. dits by strong engineered dissipation. Phys. Rev. [56] K. R. Motes, B. Q. Baragiola, A. Gilchrist, and A, 98(1):012317, 2018. https://doi.org/10.1103/ N. C. Menicucci. Encoding qubits into oscillators PhysRevA.98.012317. with atomic ensembles and squeezed light. Phys. Rev. [43] Z. Leghtas, G. Kirchmair, B. Vlastakis, R. J. A, 95(5):053819, 2017. https://doi.org/10.1103/ Schoelkopf, M. H. Devoret, and M. Mirrahimi. PhysRevA.95.053819. Hardware-efficient autonomous quantum memory pro- [57] D. J. Weigand and B. M. Terhal. Generating grid tection. Phys. Rev. Lett., 111(12):120501, 2013. https: states from Schr¨odinger-catstates without postselec- //doi.org/10.1103/PhysRevLett.111.120501. tion. Phys. Rev. A, 97(2):022341, 2018. https://doi. [44] M. Mirrahimi, Z. Leghtas, V. V. Albert, S. Touzard, org/10.1103/PhysRevA.97.022341. R. J. Schoelkopf, L. Jiang, and M. H. Devoret. Dynami- [58] J. M. Arrazola, T.R. Bromley, J. Izaac, C. R. Myers, cally protected cat-qubits: a new paradigm for universal K. Br´adler, and N. Killoran. Machine learning quantum computation. New J. Phys., 16(4):045014, method for state preparation and gate synthesis on 2014. https://doi.org/10.1088/1367-2630/16/4/ photonic quantum computers. Quantum Sci. Tech- 045014. nol., 4(2):024004, 2019. https://doi.org/10.1088/ [45] L. Li, C. L. Zou, V.V. Albert, S. Muralidharan, 2058-9565/aaf59e. S. M. Girvin, and L. Jiang. Cat codes with optimal [59] Y. Shi, C. Chamberland, and A. Cross. Fault-tolerant decoherence suppression for a lossy bosonic channel. preparation of approximate GKP states. New J. Phys. Rev. Lett., 119(3):030502, 2017. https://doi. Phys., 21(9):093007, 2019. https://doi.org/10.1088/ org/10.1103/PhysRevLett.119.030502. 1367-2630/ab3a62. [46] V. V. Albert, K. Noh, K. Duivenvoorden, D. J. Young, [60] D. Su, C. R. Myers, and K. K. Sabapathy. Con- R. T. Brierley, P. Reinhold, C. Vuillot, L. Li, C. Shen, version of Gaussian states to non-Gaussian states S. M. Girvin, B. M. Terhal, and L. Jiang. Performance using photon-number-resolving detectors. Phys. Rev. and structure of single-mode bosonic codes. Phys. Rev. A, 100(5):052301, 2019. https://doi.org/10.1103/ A, 97(3):032346, 2018. https://doi.org/10.1103/ PhysRevA.100.052301. PhysRevA.97.032346. [61] M. Eaton, R. Nehra, and O. Pfister. Non-Gaussian and [47] V. V. Albert, C. Shu, S. Krastanov, C. Shen, R. B. Gottesman-Kitaev-Preskill state preparation by photon Liu, Z. B. Yang, R. J. Schoelkopf, M. Mirrahimi, catalysis. New J. Phys., 21(11):113034, 2019. https: M. H. Devoret, and L. Jiang. Holonomic quantum //doi.org/10.1088/1367-2630/ab5330. control with continuous variable systems. Phys. Rev. [62] J. Hastrup, K. Park, J. B. Brask, R. Filip, and U. L. Lett., 116(14):140502, 2016. https://doi.org/10. Andersen. Measurement-free preparation of grid states. 1103/PhysRevLett.116.140502. npj Quantum Information, 7 number = 1, pages = [48] V. V. Albert, S. O. Mundhada, A. Grimm, S. Touzard, 1–8, year = 2021, note = https://doi.org/10.1038/ M. H. Devoret, and L. Jiang. Pair-cat codes: au- s41534-020-00353-3,. tonomous error-correction with low-order nonlinearity. [63] D. J. Weigand and B. M. Terhal. Realizing mod- Quantum Sci. Technol., 4(3):035007, 2019. https: ular quadrature measurements via a tunable photon- //doi.org/10.1088/2058-9565/ab1e69. pressure coupling in circuit QED. Phys. Rev. A, [49] D. Gottesman, A. Kitaev, and J. Preskill. Encoding a 101(5):053840, 2020. https://doi.org/10.1103/ qubit in an oscillator. Phys. Rev. A, 64(1):012310, 2001. PhysRevA.101.053840. https://doi.org/10.1103/PhysRevA.64.012310. [64] J. Hastrup and U. L. Andersen. Improved read- [50] B. M. Terhal and D. Weigand. Encoding a qubit into out of qubit-coupled Gottesman-Kitaev-Preskill states. a cavity mode in circuit QED using phase estimation. arXiv:2008.10531 [quant-ph], 2020. https://arxiv. Phys. Rev. A, 93(1):012315, 2016. https://doi.org/ org/abs/2008.10531. 10.1103/PhysRevA.93.012315. 19

[65] B. Royer, S. Singh, and S. M. Girvin. Stabilization of [78] B. Q. Baragiola, G. Pantaleoni, R. N. Alexander, finite-energy Gottesman-Kitaev-Preskill states. Phys. A. Karanjai, and N. C. Menicucci. All-Gaussian Rev. Lett., 125(26):260509, 2020. https://doi.org/ universality and fault tolerance with the Gottesman- 10.1103/PhysRevLett.125.260509. Kitaev-Preskill code. Phys. Rev. Lett., 123(20):200502, [66] K. H. Wan, A. Neville, and W. S. Koltham- 2019. https://doi.org/10.1103/PhysRevLett.123. mer. Memory-assisted decoder for approximate 200502. Gottesman-Kitaev-Preskill codes. Phys. Rev. Re- [79] J. Hastrup, M. V. Larsen, J. S. Neergaard-Nielsen, N. C. search, 2(4):043280, 2020. https://doi.org/10.1103/ Meniccuci, and U. L. Andersen. Cubic phase gates are PhysRevResearch.2.043280. not suitable for non-Clifford operations on GKP states. [67] I. Tzitrin, J. E. Bourassa, N. C. Menicucci, and K. K. Phys. Rev. A, 103(3):032409, 2021. https://doi.org/ Sabapathy. Progress towards practical qubit com- 10.1103/PhysRevA.103.032409. putation using approximate Gottesman-Kitaev-Preskill [80] L. H¨anggli,M. Heinze, and R. K¨onig.Enhanced noise codes. Phys. Rev. A, 101(3):032315, 2020. https: resilience of the surface-Gottesman-Kitaev-Preskill code //doi.org/10.1103/PhysRevA.101.032315. via designed bias. Phys. Rev. A, 102(5):052408, 2020. [68] B. M. Terhal, J. Conrad, and C. Vuillot. Towards https://doi.org/10.1103/PhysRevA.102.052408. scalable bosonic quantum error correction. Quantum [81] J. Harrington and J. Preskill. Achievable rates Sci. Technol., 5(4):043001, 2020. https://doi.org/10. for the Gaussian quantum channel. Phys. Rev. 1088/2058-9565/ab98a5. A, 64(6):062301, 2001. https://doi.org/10.1103/ [69] G. Pantaleoni, B. Q. Baragiola, and N. C. Menicucci. PhysRevA.64.062301. Modular bosonic subsystem codes. Phys. Rev. Lett., [82] F. Rozpedek, K. Noh, Q. Xu, S. Guha, and L. Jiang. 125(4):040501, 2020. https://doi.org/10.1103/ Quantum repeaters based on concatenated bosonic and PhysRevLett.125.040501. discrete-variable quantum codes. arXiv:2011.15076 [70] B. W. Walshe, B. Q. Baragiola, R. N. Alexander, and [quant-ph], 2020. https://arxiv.org/abs/2011.15076. N. C. Menicucci. Continuous-variable gate telepor- [83] K. Fukui, R. N. Alexander, and P. van Loock. All- tation and bosonic-code error correction. Phys. Rev. optical long-distance quantum communication with A, 102(6):062411, 2020. https://doi.org/10.1103/ Gottesman-Kitaev-Preskill qubits. arXiv:2011.14876 PhysRevA.102.062411. [quant-ph], 2020. https://arxiv.org/abs/2011.14876. [71] L. J. Mensen, B. Q. Baragiola, and N. C. Menicucci. [84] P. Aliferis and J. Preskill. Fault-tolerant quan- Phase-space methods for representing, manipulat- tum computation against biased noise. Phys. Rev. ing, and correcting Gottesman-Kitaev-Preskill qubits. A, 78(5):052331, 2008. https://doi.org/10.1103/ arXiv:2012.12488 [quant-ph], 2020. https://arxiv. PhysRevA.78.052331. org/abs/2012.12488. [85] J. Guillaud and M. Mirrahimi. Repetition cat qubits [72] C. Fluhmann, T. L. Nguyen, M. Marinelli, V. Neg- for fault-tolerant quantum computation. Phys. Rev. nevitsky, K. Mehta, and J. P. Home. Encoding a X, 9(4):041053, 2019. https://doi.org/10.1103/ qubit in a trapped-ion mechanical oscillator. Nature, PhysRevX.9.041053. 566(7745):513–517, 2019. https://doi.org/10.1038/ [86] J. Guillaud and M. Mirrahimi. Error rates and resource s41586-019-0960-6. overheads of repetition cat qubits. arXiv:2009.10756 [73] B. de Neeve, T. L. Nguyen, T. Behrle, and J. Home. [quant-ph], 2020. https://arxiv.org/abs/2009.10756. Error correction of a logical grid state qubit by [87] D. K. Tuckett, S. D. Bartlett, and S. T. Flammia. dissipative pumping. arXiv:2010.09681 [quant-ph], Ultrahigh error threshold for surface codes with biased 2020. https://arxiv.org/abs/2010.09681. noise. Phys. Rev. Lett., 120(5):050505, 2018. https: [74] P. Campagne-Ibarcq, A. Eickbusch, S. Touzard, //doi.org/10.1103/PhysRevLett.120.050505. E. Zalys-Geller, N. E. Frattini, V. V. Sivak, P. Reinhold, [88] D. K. Tuckett, A. S. Darmawan, C. T. Chubb, S. Bravyi, S. Puri, S. Shankar, R. J. Schoelkopf, L. Frunzio, S. D. Bartlett, and S. T. Flammia. Tailoring surface M. Mirrahimi, and M. H. Devoret. Quantum error codes for highly biased noise. Phys. Rev. X, 9(4):041031, correction of a qubit encoded in grid states of an 2019. https://doi.org/10.1103/PhysRevX.9.041031. oscillator. Nature, 584(7821):368–372, 2020. https: [89] D. K. Tuckett, S. D. Bartlett, S. T. Flammia, and //doi.org/10.1038/s41586-020-2603-3. B. J. Brown. Fault-tolerant thresholds for the surface [75] K. Noh, V. V. Albert, and L. Jiang. Quantum code in excess of 5% under biased noise. Phys. Rev. capacity bounds of Gaussian thermal loss channels Lett., 124(13):130501, 2020. https://doi.org/10. and achievable rates with Gottesman-Kitaev-Preskill 1103/PhysRevLett.124.130501. codes. IEEE Transactions on Information Theory, [90] C. Chamberland, K. Noh, P. Arrangoiz-Arriola, E. T. 65(4):2563–2582, 2019. https://doi.org/10.1109/ Campbell, C. T. Hann, J. Iverson, H. Putterman, T. C. TIT.2018.2873764. Bohdanowicz, S. T. Flammia, A. Keller, G. Rafael, [76] D. Gottesman and J. Preskill. Secure quantum J. Preskill, L. Jiang, A. H. Safavi-Naeini, O. PAinter, key distribution using squeezed states. In Quan- and F. G. S. L. Brand˜ao. Building a fault-tolerant tum Information with Continuous Variables, pages quantum computer using concatenated cat codes. 317–356. Springer, 2003. https://doi.org/10.1007/ arXiv:2012.04108 [quant-ph], 2020. https://arxiv. 978-94-015-1258-9_22. org/abs/2012.04108. [77] S. Bravyi and A. Kitaev. Universal quantum com- [91] P. T. Cochrane, G. J. Milburn, and W. J. Munro. putation with ideal Clifford gates and noisy ancillas. Macroscopically distinct quantum-superposition states Phys. Rev. A, 71(2):022316, 2005. https://doi.org/ as a bosonic code for amplitude damping. Phys. 10.1103/PhysRevA.71.022316. Rev. A, 59(4):2631, 1999. https://doi.org/10.1103/ PhysRevA.59.2631. 20

[92] H. Jeong and M. S. Kim. Efficient quantum decoherence. Nature, 455(7212):510–514, 2008. https: computation using coherent states. Phys. Rev. //doi.org/10.1038/nature07288. A, 65(4):042305, 2002. https://doi.org/10.1103/ [107] B. Hacker, S. Welte, S. Daiss, A. Shaukat, L. Ritter, PhysRevA.65.042305. S.and Li, and G. Rempe. Deterministic creation of [93] T. C. Ralph, A. Gilchrist, G. J. Milburn, W. J. Munro, entangled atom–light Schr¨odinger-catstates. Nature and S. Glancy. Quantum computation with optical Photon., 13(2):110–115, 2019. https://doi.org/10. coherent states. Phys. Rev. A, 68(4):042319, 2003. 1038/s41566-018-0339-5. https://doi.org/10.1103/PhysRevA.68.042319. [108] A. Omran, H. Levine, A. Keesling, G. Semeghini, [94] S. Glancy, H. M. Vasconcelos, and T. C. Ralph. T. T. Wang, S. Ebadi, H. Bernien, A. S. Zibrov, Transmission of optical coherent-state qubits. Phys. H. Pichler, S. Choi, J. Cui, M. Rossignolo, P. Rembold, Rev. A, 70(2):022317, 2004. https://doi.org/10. S. Montangero, T. Calarco, M. Endres, M. Greiner, 1103/PhysRevA.70.022317. V. Vuleti´c, and M. D. Lukin. Generation and [95] A. P. Lund, T. C. Ralph, and H. L. Haselgrove. manipulation of Schr¨odingercat states in Rydberg atom Fault-tolerant linear optical quantum computing with arrays. Science, 365(6453):570–574, 2019. https: small-amplitude coherent states. Phys. Rev. Lett., //doi.org/10.1126/science.aax9743. 100(3):030503, 2008. https://doi.org/10.1103/ [109] A. Ourjoumtsev, R. Tualle-Brouri, J. Laurat, and PhysRevLett.100.030503. P. Grangier. Generating optical Schr¨odinger kit- [96] S. Puri, S. Boutin, and A. Blais. Engineering the tens for quantum information processing. Science, quantum states of light in a Kerr-nonlinear resonator by 312(5770):83–86, 2006. https://doi.org/10.1126/ two-photon driving. npj Quantum Inf., 3(1):18, 2017. science.1122858. https://doi.org/10.1038/s41534-017-0019-1. [110] Q. Hou, W. Yang, C. Chen, and Z. Yin. Generation [97] S. Puri, A. Grimm, P. Campagne-Ibarcq, A. Eickbusch, of macroscopic Schr¨odinger cat state in diamond K. Noh, G. Roberts, L. Jiang, M. Mirrahimi, M. H. mechanical resonator. Sci. Rep., 6:37542, 2016. https: Devoret, and S. M. Girvin. Stabilized cat in a //doi.org/10.1038/srep37542. driven nonlinear cavity: a fault-tolerant error syndrome [111] C. Bulutay. Cat-state generation and stabilization for detector. Phys. Rev. X, 9(4):041009, 2019. https: a nuclear spin through electric quadrupole interaction. //doi.org/10.1103/PhysRevX.9.041009. Phys. Rev. A, 96(1):012312, 2017. https://doi.org/ [98] S. Puri, L. St-Jean, J. A. Gross, A. Grimm, N. E. 10.1103/PhysRevA.96.012312. Frattini, P. S. Iyer, A. Krishna, S. Touzard, L. Jiang, [112] L. Frunzio, A. Wallraff, D. Schuster, J. Majer, and A. Blais, S. T. Flammia, and S. M. Girvin. Bias- R. Schoelkopf. Fabrication and characterization of preserving gates with stabilized cat qubits. Sci. Adv., superconducting circuit QED devices for quantum 6(34):5901, 2020. https://doi.org/10.1126/sciadv. computation. IEEE Transactions on Applied Supercon- aay5901. ductivity, 15(2):860–863, 2005. https://doi.org/10. [99] S. Lloyd. Capacity of the noisy quantum channel. Phys. 1109/TASC.2005.850084. Rev. A, 55(3):1613–1622, 1997. https://doi.org/10. [113] R. J. Schoelkopf and S. M. Girvin. Wiring up quantum 1103/PhysRevA.55.1613. systems. Nature, 451(7179):664–669, 2008. https:// [100] I. Devetak. The private classical capacity and quantum doi.org/10.1038/451664a. capacity of a quantum channel. IEEE Transactions on [114] A. Wallraff, D. I. Schuster, A. Blais, L. Frunzio, Information Theory, 51(1):44–55, 2005. https://doi. R. S. Huang, J. Majer, S. Kumar, S. M. Girvin, and org/10.1109/TIT.2004.839515. R. J. Schoelkopf. Strong coupling of a single photon [101] A. S. Holevo and R. F. Werner. Evaluating ca- to a superconducting qubit using circuit quantum pacities of bosonic Gaussian channels. Phys. Rev. electrodynamics. Nature, 431(7005):162–167, 2004. A, 63(3):032312, 2001. https://doi.org/10.1103/ https://doi.org/10.1038/nature02851. PhysRevA.63.032312. [115] D. I. Schuster, A. A. Houck, J. A. Schreier, A. Wallraff, [102] M. M. Wolf, D. P´erez-Garc´ıa, and G. Giedke. Quan- J. M. Gambetta, A. Blais, L. Frunzio, J. Majer, tum capacities of bosonic channels. Phys. Rev. B. Johnson, M. H. Devoret, S. M. Girvin, and R. J. Lett., 98(13):130501, 2007. https://doi.org/10.1103/ Schoelkopf. Resolving photon number states in a PhysRevLett.98.130501. superconducting circuit. Nature, 445(7127):515–518, [103] M. M. Wilde, P. Hayden, and S. Guha. Quantum 2007. https://doi.org/10.1038/nature05461. trade-off coding for bosonic communication. Phys. Rev. [116] T. Niemczyk, F. Deppe, H. Huebl, E. P. Menzel, A, 86(6):062306, 2012. https://doi.org/10.1103/ F. Hocke, M. J. Schwarz, J. J. Garcia-Ripoll, D. Zueco, PhysRevA.86.062306. T. H¨ummer,E. Solano, A. Marx, and R. Gross. Circuit [104] M. M. Wilde and H. Qi. Energy-constrained private quantum electrodynamics in the ultrastrong-coupling and quantum capacities of quantum channels. IEEE regime. Nat. Phys., 6(10):772–776, 2010. https://doi. Transactions on Information Theory, 64(12):7802– org/10.1038/nphys1730. 7827, 2018. https://doi.org/10.1109/TIT.2018. [117] S. M. Girvin, M. H. Devoret, and R. J. Schoelkopf. 2854766. Circuit QED and engineering charge-based super- [105] L. Arqand, A. Memarzadeh and S. Mancini. Quantum conducting qubits. Phys. Scr., 2009(T137):014012, capacity of a bosonic dephasing channel. Phys. Rev. 2009. https://doi.org/10.1088/0031-8949/2009/ A, 102(4):042413, 2020. https://doi.org/10.1103/ T137/014012. PhysRevA.102.042413. [118] T. D. Ladd, F. Jelezko, R. Laflamme, Y. Nakamura, [106] S. Deleglise, I. Dotsenko, C. Sayrin, J. Bernu, M. Brune, C. Monroe, and J. L. O’Brien. Quantum computers. J. M. Raimond, and S. Haroche. Reconstruction of Nature, 464(7285):45–53, 2010. https://doi.org/10. non-classical cavity field states with snapshots of their 1038/nature08812. 21

[119] G. Wendin. Quantum information processing with [130] A. Kandala, A.and Mezzacapo, M. Temme, superconducting circuits: a review. Rep. Prog. K.and Takita, M. Brink, J. M. Chow, and J. M Phys., 80(10):106001, 2017. https://doi.org/10. Gambetta. Hardware-efficient variational quantum 1088/1361-6633/aa7e1a. eigensolver for small molecules and quantum [120] P. Krantz, M. Kjaergaard, F. Yan, T. P. Orlando, magnets. Nature, 549(7671):242–246, 2017. S. Gustavsson, and W. D. Oliver. A quantum https://doi.org/10.1038/nature23879. engineer’s guide to superconducting qubits. Appl. Phys. [131] F. Arute, K. Arya, R. Babbush, D. Bacon, J. C. Rev., 6(2):021318, 2019. https://doi.org/10.1063/1. Bardin, R. Barends, R. Biswas, S. Boixo, F. G. S. L. 5089550. Brandao, D. A. Buell, B. Burkett, Y. Chen, Z. Chen, [121] M. Kjaergaard, M. E. Schwartz, J. Braum¨uller, B. Chiaro, R. Collins, W. Courtney, A. Dunsworth, P. Krantz, J. I. J. Wang, S. Gustavsson, and E. Farhi, B. Foxen, A. Fowler, C. Gidney, M. Giustina, W. D. Oliver. Superconducting qubits: current R. Graff, K. Guerin, S. Habegger, M. P. Harrigan, state of play. Annual Review of Condensed Matter M. J. Hartmann, A. Ho, M. Hoffmann, T. Huang, T. S. Physics, 11(1):369–395, 2020. https://doi.org/10. Humble, S. V. Isakov, E. Jeffrey, Z. Jiang, D. Kafri, 1146/annurev-conmatphys-031119-050605. K. Kechedzhi, J. Kelly, P. V. Klimov, S. Knysh, [122] A. Blais, S. M. Girvin, and W. D. Oliver. Quantum A. Korotkov, F. Kostritsa, D. Landhuis, M. Lindmark, information processing and quantum optics with circuit E. Lucero, D. Lyakh, S. Mandra, J. R. McClean, quantum electrodynamics. Nat. Phys., 16(3):247–256, M. McEwen, A. Megrant, X. Mi, K. Michielsen, 2020. https://doi.org/10.1038/s41567-020-0806-z. M. Mohseni, J. Mutus, O. Naaman, M. Neeley, C. Neill, [123] W. Cai, Y. Ma, W. Wang, C. L. Zou, and L. Sun. M. Y. Niu, E. Ostby, A. Petukhov, J. C. Platt, Bosonic quantum error correction codes in supercon- C. Quintana, E. G. Rieffel, P. Roushan, N. C. Rubin, ducting quantum circuits. Fundamental Research, D. Sank, K. J. Satzinger, V. Smelyanskiy, K. J. 1(1):50–67, 2021. https://doi.org/10.1016/j.fmre. Sung, M. D. Trevithick, A. Vainsencher, B. Villalonga, 2020.12.006. T. White, Z. J. Yao, P. Yeh, A. Zalcman, H. Neven, [124] S. Haroche and J. M. Raimond. Exploring the quantum: and J. M. Martinis. Quantum supremacy using a atoms, cavities, and photons. Oxford University programmable superconducting processor. Nature, Press, 2006. https://doi.org/10.1093/acprof:oso/ 574(7779):505–510, 2019. https://doi.org/10.1038/ 9780198509141.001.0001. s41586-019-1666-5. [125] J. Koch, M. Y. Terri, J. Gambetta, A. A. Houck, D. I. [132] P. Jurcevic, A. Javadi-Abhari, L. S. Bishop, I. Lauer, Schuster, J. Majer, A. Blais, M. H. Devoret, S. M. D. F. Bogorin, M. Brink, L. Capelluto, O. G¨unl¨uk, Girvin, and R. J. Schoelkopf. Charge-insensitive qubit T. Itoko, N. Kanazawa, A. Kandala, K. Keefe, G. design derived from the Cooper pair box. Phys. Rev. A. Krsulich, W. Landers, E. P. Lewandowski, D. T. A, 76(4):042319, 2007. https://doi.org/10.1103/ McClure, G. Nannicini, A. Narasgond, H. M. Nayfeh, PhysRevA.76.042319. E. Pritchett, M. B. Rothwell, S. Srinivasan, N. Sun- [126] J. A. Schreier, A. A. Houck, J. Koch, D. I. Schus- daresan, C. Wang, K. X. Wei, C. J. Wood, J. B. ter, B. R. Johnson, J. M. Chow, J. M. Gambetta, Yau, E. J. Zhang, O. E. Dial, J. M. Chow, and J. M. J. Majer, L. Frunzio, M. H. Devoret, S. M. Girvin, Gambetta. Demonstration of quantum volume 64 on a and R. J. Schoelkopf. Suppressing charge noise superconducting quantum computing system. Quantum decoherence in superconducting charge qubits. Phys. Sci. Technol., 6(2):025020, 2021. https://doi.org/10. Rev. B, 77(18):180502, 2008. https://doi.org/10. 1088/2058-9565/abe519. 1103/PhysRevB.77.180502. [133] Google AI Quantum and Collaborators. Hartree- [127] J. Preskill. Quantum computing in the NISQ era and Fock on a superconducting qubit quantum computer. beyond. Quantum, 2:79, 2018. https://doi.org/10. Science, 369(6507):1084–1089, 2020. https://doi. 22331/q-2018-08-06-79. org/10.1126/science.abb9811. [128] X. Fu, M. A. Rol, C. C. Bultink, J. van Someren, [134] N. Lacroix, C. Hellings, C. K. Andersen, A. Di Paolo, N. Khammassi, I. Ashraf, R. F. L. Vermeulen, J. C. A. Remm, S. Lazar, S. Krinner, G. J. Norris, de Sterke, W. J. Vlothuizen, R. N. Schouten, C. G. Al- M. Gabureac, J. Heinsoo, A. Blais, C. Eichler, and mudever, L. DiCarlo, and K. Bertels. An experimental A. Wallraff. Improving the performance of deep microarchitecture for a superconducting quantum pro- quantum optimization algorithms with continuous gate cessor. In Proceedings of the 50th Annual IEEE/ACM sets. PRX Quantum, 1(2):110304, 2020. https://doi. International Symposium on Microarchitecture, pages org/10.1103/PRXQuantum.1.020304. 813–825, 2017. https://doi.org/10.1145/3123939. [135] H. Paik, D. I. Schuster, L. S. Bishop, G. Kirchmair, 3123952. G. Catelani, A. P. Sears, B. R. Johnson, M. J. Reagor, [129] J. S. Otterbach, R. Manenti, N. Alidoust, A. Best- L. Frunzio, L. I. Glazman, S. M. Girvin, M. H. wick, M. Block, B. Bloom, S. Caldwell, N. Didier, Devoret, and R. J. Schoelkopf. Observation of high E. Schuyler Fried, S. Hong, P. Karalekas, C. B. Osborn, coherence in Josephson junction qubits measured in A. Papageorge, E. C. Peterson, G. Prawiroatmodjo, a three-dimensional circuit QED architecture. Phys. N. Rubin, C. A. Ryan, D. Scarabelli, M. Scheer, E. A. Rev. Lett., 107(24):240501, 2011. https://doi.org/ Sete, P. Sivarajah, R. S. Smith, A. Staley, N. Tezak, 10.1103/PhysRevLett.107.240501. W. J. Zeng, A. Hudson, B. R. Johnson, M. Reagor, M. P. [136] M. Reagor, W. Pfaff, C. Axline, R. W. Heeres, da Silva, and C. Rigetti. Unsupervised machine learning N. Ofek, K. Sliwa, E. Holland, C. Wang, J. Blumoff, on a hybrid quantum computer. arXiv:1712.05771 K. Chou, M. J. Hatridge, L. Frunzio, M. H. Devoret, [quant-ph], 2017. https://arxiv.org/abs/1712.05771. L. Jiang, and R. J. Schoelkopf. Quantum memory with millisecond coherence in circuit QED. Phys. Rev. 22

B, 94(1):014506, 2016. https://doi.org/10.1103/ [148] A. D. O’Connell, M. Ansmann, R. C. Bialczak, PhysRevB.94.014506. M. Hofheinz, N. Katz, E. Lucero, C. McKenney, [137] T. Brecht, M. Reagor, Y. Chu, W. Pfaff, C. Wang, M. Neeley, H. Wang, E. M. Weig, A. N. Cleland, and L. Frunzio, M. H. Devoret, and R. J. Schoelkopf. J. M. Martinis. Microwave dielectric loss at single Demonstration of superconducting micromachined cav- photon energies and millikelvin temperatures. Appl. ities. Appl. Phys. Lett., 107(19):192603, 2015. https: Phys. Lett., 92(11):112903, 2008. https://doi.org/ //doi.org/10.1063/1.4935541. 10.1063/1.2898887. [138] A. A. Houck, D. I. Schuster, J. M. Gambetta, J. A. [149] J. Gao, M. Daal, A. Vayonakis, S. Kumar, J. Zmuidz- Schreier, B. R. Johnson, J. M. Chow, L. Frunzio, inas, B. Sadoulet, B. A. Mazin, P. K. Day, and H. G. J. Majer, M. H. Devoret, S. M. Girvin, and R. J. Leduc. Experimental evidence for a surface distribution Schoelkopf. Generating single microwave photons in of two-level systems in superconducting lithographed a circuit. Nature, 449(7160):328–331, 2007. https: microwave resonators. Appl. Phys. Lett., 92(15):152505, //doi.org/10.1038/nature06126. 2008. https://doi.org/10.1063/1.2906373. [139] J. Majer, J. M. Chow, J. M. Gambetta, J. Koch, B. R. [150] C. Muller, J. H. Cole, and J. Lisenfeld. Towards Johnson, J. A. Schreier, L. Frunzio, D. I. Schuster, A. A. understanding two-level-systems in amorphous solids: Houck, A. Wallraff, A. Blais, M. H. Devoret, S. M. insights from quantum circuits. Rep. Prog. Phys., Girvin, and R. J. Schoelkopf. Coupling superconducting 82(12):124501, 2019. https://doi.org/10.1088/ qubits via a cavity bus. Nature, 449(7161):443–447, 1361-6633/ab3a7e. 2007. https://doi.org/10.1038/nature06184. [151] J. M. Sage, V. Bolkhovsky, W. D. Oliver, B. Turek, [140] M. A. Sillanp¨a¨a, J. I. Park, and R. W. Simmonds. and P. B. Welander. Study of loss in superconducting Coherent quantum state storage and transfer between coplanar waveguide resonators. Journal of Applied two phase qubits via a resonant cavity. Nature, Physics, 109(6):063915, 2011. https://doi.org/10. 449(7161):438–442, 2007. https://doi.org/10.1038/ 1063/1.3552890. nature06124. [152] M. R Vissers, J. S. Kline, J. Gao, D. S. Wisbey, and [141] L. DiCarlo, J. M. Chow, J. M. Gambetta, L. S. Bishop, D. P. Pappas. Reduced microwave loss in trenched B. R. Johnson, D. I. Schuster, J. Majer, A. Blais, superconducting coplanar waveguides. Appl. Phys. L. Frunzio, S. M. Girvin, and R. J. Schoelkopf. Demon- Lett., 100(8):082602, 2012. https://doi.org/10. stration of two-qubit algorithms with a superconducting 1063/1.3683552. quantum processor. Nature, 460(7252):240–244, 2009. [153] K. Geerlings, S. Shankar, E. Edwards, L. Frunzio, https://doi.org/10.1038/nature08121. R. J. Schoelkopf, and M. H. Devoret. Improving [142] A. Wallraff, D. I. Schuster, A. Blais, L. Frunzio, the quality factor of microwave compact resonators by J. Majer, M. H. Devoret, S. M. Girvin, and R. J. optimizing their geometrical parameters. Appl. Phys. Schoelkopf. Approaching unit visibility for control of Lett., 100(19):192601, 2012. https://doi.org/10. a superconducting qubit with dispersive readout. Phys. 1063/1.4710520. Rev. Lett., 95(6):060501, 2005. https://doi.org/10. [154] G. Calusine, A. Melville, W. Woods, R. Das, C. Stull, 1103/PhysRevLett.95.060501. V. Bolkhovsky, D. Braje, D. Hover, D. K. Kim, [143] J. Johansson, S. Saito, T. Meno, H. Nakano, M. Ueda, X. Miloshi, D. Rosenberg, A. Sevi, J. L. Yoder, K. Semba, and H. Takayanagi. Vacuum Rabi oscillations E. Dauler, and W. D. Oliver. Analysis and mitigation in a macroscopic superconducting qubit oscillator of interface losses in trenched superconducting coplanar system. Phys. Rev. Lett., 96(12):127006, 2006. https: waveguide resonators. Appl. Phys. Lett., 112(6):062601, //doi.org/10.1103/PhysRevLett.96.127006. 2018. https://doi.org/10.1063/1.5006888. [144] A. Blais, R. S. Huang, A. Wallraff, S. M. Girvin, [155] H. Wang, M. Hofheinz, J. Wenner, M. Ansmann, and R. J. Schoelkopf. Cavity quantum electrody- R. C. Bialczak, M. Lenander, E. Lucero, M. Neeley, namics for superconducting electrical circuits: an A. D. O’Connell, D. Sank, M. Weides, A. N. Cleland, architecture for quantum computation. Phys. Rev. and J. M. Martinis. Improving the coherence time A, 69(6):062320, 2004. https://doi.org/10.1103/ of superconducting coplanar resonators. Appl. Phys. PhysRevA.69.062320. Lett., 95(23):233508, 2009. https://doi.org/10.1063/ [145] S. Krastanov, V. V. Albert, C. Shen, C. L. Zou, R. W. 1.3273372. Heeres, B. Vlastakis, R. J. Schoelkopf, and L. Jiang. [156] R. Barends, N. Vercruyssen, A. Endo, P. J. de Visser, Universal control of an oscillator with dispersive cou- T. Zijlstra, T. M. Klapwijk, P. Diener, S. J. C. pling to a qubit. Phys. Rev. A, 92(4):040303, 2015. Yates, and J. J. A. Baselmans. Minimal resonator https://doi.org/10.1103/PhysRevA.92.040303. loss for circuit quantum electrodynamics. Appl. Phys. [146] T. Brecht, Y. Chu, C. Axline, W. Pfaff, J. Z. Lett., 97(2):023508, 2010. https://doi.org/10.1063/ Blumoff, K. Chou, L. Krayzman, L. Frunzio, and 1.3458705. R. J. Schoelkopf. Micromachined integrated quantum [157] M. R. Vissers, J. Gao, D. S. Wisbey, D. A. Hite, C. C. circuit containing a superconducting qubit. Phys. Tsuei, A. D. Corcoles, M. Steffen, and D. P. Pappas. Rev. Applied, 7(4):044018, 2017. https://doi.org/10. Low loss superconducting titanium nitride coplanar 1103/PhysRevApplied.7.044018. waveguide resonators. Appl. Phys. Lett., 97(23):232509, [147] J. M. Martinis, K. B. Cooper, R. McDermott, M. Stef- 2010. https://doi.org/10.1063/1.3517252. fen, M. Ansmann, K. D. Osborn, K. Cicak, S. Oh, D. P. [158] A. P. M. Place, L. V. H. Rodgers, P. Mundada, B. M. Pappas, R. W. Simmonds, and C. C. Yu. Decoherence Smitham, M. Fitzpatrick, Z. Leng, A. Premkumar, in Josephson qubits from dielectric loss. Phys. Rev. J. Bryon, S. Sussman, G. Cheng, T. Madhavan, H. K. Lett, 95(21):210503, 2005. https://doi.org/10.1103/ Babla, B. Jaeck, A. Gyenis, N. Yao, R. J. Cava, PhysRevLett.95.210503. N. P. de Leon, and A. A. Houck. New material 23

platform for superconducting transmon qubits with Phys. Lett., 116(15):154002, 2020. https://doi.org/ coherence times exceeding 0.3 milliseconds. Nature 10.1063/5.0003907. Communications, 12(1):1–6, 2021. https://doi.org/ [170] M. Hofheinz, H. Wang, M. Ansmann, R. C. Bialczak, 10.1038/s41467-021-22030-5. E. Lucero, M. Neeley, A. D. O’Connell, D. Sank, [159] A. Megrant, C. Neill, R. Barends, B. Chiaro, Y. Chen, J. Wenner, J. M. Martinis, and A. N. Cleland. Synthe- L. Feigl, J. Kelly, E. Lucero, M. Mariantoni, P. J. J. sizing arbitrary quantum states in a superconducting O’Malley, D. Sank, A. Vainsencher, J. Wenner, T. C. resonator. Nature, 459(7246):546–549, 2009. https: White, Y. Yin, J. Zhao, C. J. Palmstrøm, John M. //doi.org/10.1038/nature08005. Martinis, and A. N. Cleland. Planar superconducting [171] P. J. Leek, M. Baur, J. M. Fink, R. Bianchetti, resonators with internal quality factors above one L. Steffen, S. Filipp, and A. Wallraff. Cavity quantum million. Appl. Phys. Lett., 100(11):113510, 2012. electrodynamics with separate photon storage and qubit https://doi.org/10.1063/1.3693409. readout modes. Phys. Rev. Lett., 104(10):100504, 2010. [160] M. Sandberg, M. R. Vissers, J. S. Kline, M. Weides, https://doi.org/10.1103/PhysRevLett.104.100504. J. Gao, D. S. Wisbey, and D. P. Pappas. Etch induced [172] B. Vlastakis, G. Kirchmair, Z. Leghtas, S. E. Nigg, microwave losses in titanium nitride superconducting L. Frunzio, S. M. Girvin, M. Mirrahimi, M. H. Devoret, resonators. Appl. Phys. Lett., 100(26):262605, 2012. and R. J. Schoelkopf. Deterministically encoding https://doi.org/10.1063/1.4729623. quantum information using 100-photon Schr¨odingercat [161] A. Bruno, G. de Lange, S. Asaad, K. L. van der Enden, states. Science, 342(6158):607–610, 2013. https:// N. K. Langford, and L. DiCarlo. Reducing intrinsic loss doi.org/10.1126/science.1243289. in superconducting resonators by surface treatment and [173] L. Hu, Y. Ma, W. Cai, X. Mu, Y. Xu, W. Wang, deep etching of silicon substrates. Appl. Phys. Lett., Y. Wu, H. Wang, Y. P. Song, C. L. Zou, S. M. Girvin, 106(18):182601, 2015. https://doi.org/10.1063/1. L. M. Duan, and L. Sun. Quantum error correction 4919761. and universal gate set operation on a binomial bosonic [162] A. Melville, G. Calusine, W. Woods, K. Serniak, logical qubit. Nat. Phys., 15(5):503–508, 2019. https: E. Golden, B. M. Niedzielski, D. K. Kim, A. Sevi, //doi.org/10.1038/s41567-018-0414-3. J. L. Yoder, E. A. Dauler, and W. D. Oliver. [174] Y. Ma, Y. Xu, X. Mu, W. Cai, L. Hu, W. Wang, X. Pan, Comparison of dielectric loss in titanium nitride and H. Wang, Y. P. Song, C. L. Zou, and L. Sun. Error- aluminum superconducting resonators. Appl. Phys. transparent operations on a logical qubit protected by Lett., 117(12):124004, 2020. https://doi.org/10. quantum error correction. Nat. Phys., 16(8):827–831, 1063/5.0021950. 2020. https://doi.org/10.1038/s41567-020-0893-x. [163] M. Kudra, J. Bizn´arov´a,F. Roudsari, J. J. Burnett, [175] W. D Oliver and P. B. Welander. Materials in supercon- D. Niepce, S. Gasparinetti, B. Wickman, and P. Delsing. ducting quantum bits. MRS Bulletin, 38(10):816–825, High quality three-dimensional aluminum microwave 2013. https://doi.org/10.1557/mrs.2013.229. cavities. Appl. Phys. Lett., 117(7):070601, 2020. https: [176] C. R. H. McRae, H. Wang, J. Gao, M. Vissers, //doi.org/10.1063/5.0016463. T. Brecht, A. Dunsworth, D. Pappas, and J. Mutus. [164] C. Axline, M. Reagor, R. Heeres, P. Reinhold, C. Wang, Materials loss measurements using superconducting K. Shain, W. Pfaff, Y. Chu, L. Frunzio, and R. J. microwave resonators. Rev. Sci. Instrum., 91(9):091101, Schoelkopf. An architecture for integrating planar and 2020. https://doi.org/10.1063/5.0017378. 3D cQED devices. Appl. Phys. Lett., 109(4):042601, [177] J. Wenner, R. Barends, R. C. Bialczak, Y. Chen, 2016. https://doi.org/10.1063/1.4959241. J. Kelly, E. Lucero, M. Mariantoni, A. Megrant, P. J. J. [165] Z. K. Minev, I. M. Pop, and M. H. Devoret. Planar O’Malley, D. Sank, A. Vainsencher, H. Wang, T. C. superconducting whispering gallery mode resonators. White, Y. Yin, J. Zhao, A. N. Cleland, and J. M. Appl. Phys. Lett., 103(14):142604, 2013. https://doi. Martinis. Surface loss simulations of superconducting org/10.1063/1.4824201. coplanar waveguide resonators. Appl. Phys. Lett., [166] T. Brecht, W. Pfaff, C. Wang, Y. Chu, L. Frunzio, 99(11):113513, 2011. https://doi.org/10.1063/1. M. H. Devoret, and R. J. Schoelkopf. Multilayer 3637047. microwave integrated quantum circuits for scalable [178] C. Wang, C. Axline, Y. Y. Gao, T. Brecht, Y. Chu, quantum computing. npj Quantum Inf., 2(1):1–4, 2016. L. Frunzio, M. H. Devoret, and R. J. Schoelkopf. Surface https://doi.org/10.1038/npjqi.2016.2. participation and dielectric loss in superconducting [167] Z. K. Minev, K. Serniak, I. M Pop, Z. Leghtas, qubits. Appl. Phys. Lett., 107(16):162601, 2015. https: K. Sliwa, M. Hatridge, L. Frunzio, R. J Schoelkopf, //doi.org/10.1063/1.4934486. and M. H Devoret. Planar multilayer circuit quantum [179] W. Woods, G. Calusine, A. Melville, A. Sevi, E. Golden, electrodynamics. Phys. Rev. Applied, 5(4):044021, D. K. Kim, D. Rosenberg, J. L. Yoder, and W. D. 2016. https://doi.org/10.1103/PhysRevApplied.5. Oliver. Determining interface dielectric losses in 044021. superconducting coplanar-waveguide resonators. Phys. [168] D. Zoepfl, P. R. Muppalla, C. M. F. Schneider, Rev. Applied, 12(1):014012, 2019. https://doi.org/ S. Kasemann, S. Partel, and G. Kirchmair. Char- 10.1103/PhysRevApplied.12.014012. acterization of low loss microstrip resonators as a [180] Y. Y. Gao, B. J. Lester, Y. Zhang, C. Wang, building block for circuit QED in a 3D waveguide. AIP S. Rosenblum, L. Frunzio, L. Jiang, S. M. Girvin, Advances, 7(8):085118, 2017. https://doi.org/10. and R. J. Schoelkopf. Programmable interference 1063/1.4992070. between two microwave quantum memories. Phys. [169] Chan U. Lei, L. Krayzman, S. Ganjam, L. Frunzio, Rev. X, 8(2):021073, 2018. https://doi.org/10.1103/ and R. J. Schoelkopf. High coherence superconducting PhysRevX.8.021073. microwave cavities with indium bump bonding. Appl. 24

[181] R. J. Glauber. Coherent and incoherent states of the https://doi.org/10.1038/s41534-019-0201-8. radiation field. Phys. Rev., 131(6):2766, 1963. https: [194] N. Bergeal, F. Schackert, M. Metcalfe, R. Vijay, //doi.org/10.1103/PhysRev.131.2766. V. E. Manucharyan, L. Frunzio, D. E. Prober, R. J. [182] N. E. Frattini, U. Vool, S. Shankar, A. Narla, K. M. Schoelkopf, S. M. Girvin, and M. H. Devoret. Phase- Sliwa, and M. H. Devoret. 3-wave mixing Josephson preserving amplification near the quantum limit with dipole element. Appl. Phys. Lett., 110(22):222603, 2017. a Josephson ring modulator. Nature, 465(7294):64–68, https://doi.org/10.1063/1.4984142. 2010. https://doi.org/10.1038/nature09035. [183] Z. Leghtas, G. Kirchmair, B. Vlastakis, M. H. Devoret, [195] S. Rosenblum, Y. Y. Gao, P. Reinhold, C. Wang, R. J. Schoelkopf, and M. Mirrahimi. Deterministic C. J. Axline, L. Frunzio, S. M. Girvin, L. Jiang, protocol for mapping a qubit to coherent state super- M. Mirrahimi, M. H. Devoret, and R. J. Schoelkopf. positions in a cavity. Phys. Rev. A, 87(4):042315, 2013. A CNOT gate between multiphoton qubits encoded in https://doi.org/10.1103/PhysRevA.87.042315. two cavities. Nat. Commun., 9(1):652, 2018. https: [184] C. Wang, Y. Y. Gao, P. Reinhold, R. W. Heeres, //doi.org/10.1038/s41467-018-03059-5. N. Ofek, K. Chou, C. Axline, M. Reagor, J. Blumoff, [196] S. Aaronson and A. Arkhipov. The computational and K. M. Sliwa. A Schr¨odingercat living in two complexity of linear optics. In Proceedings of the boxes. Science, 352(6289):1087–1091, 2016. https: Forty-Third Annual ACM Symposium on Theory of //doi.org/10.1126/science.aaf2941. Computing, pages 333–342, 2011. https://doi.org/ [185] J. M. Chow, J. M. Gambetta, A. D. C´orcoles, S. T. 10.1145/1993636.1993682. Merkel, J. A. Smolin, C. Rigetti, S. Poletto, G. A. [197] J. Huh, G. Gi. Guerreschi, B. Peropadre, J. R. McClean, Keefe, M. B. Rothwell, J. R. Rozen, M. B. Ketchen, and and A. Aspuru-Guzik. Boson sampling for molecular M. Steffen. Universal quantum gate set approaching vibronic spectra. Nature Photon., 9(9):615–620, 2015. fault-tolerant thresholds with superconducting qubits. https://doi.org/10.1038/nphoton.2015.153. Phys. Rev. Lett., 109(6):060501, 2012. https://doi. [198] C. Sparrow, E. Mart´ın-L´opez, N. Maraviglia, org/10.1103/PhysRevLett.109.060501. C. Neville, A.and Harrold, J. Carolan, Y. N. [186] R. W. Heeres, B. Vlastakis, E. Holland, S. Krastanov, Joglekar, T. Hashimoto, N. Matsuda, J. L. V. V. Albert, L. Frunzio, L. Jiang, and R. J. Schoelkopf. O’Brien, D. P. Tew, and A. Laing. Simulating Cavity state manipulation using photon-number selec- the vibrational quantum dynamics of molecules tive phase gates. Phys. Rev. Lett., 115(13):137002, using photonics. Nature, 557(7707):660–667, 2018. 2015. https://doi.org/10.1103/PhysRevLett.115. https://doi.org/10.1038/s41586-018-0152-9. 137002. [199] W. R. Clements, J. J. Renema, A. Eckstein, A. A. [187] N. Khaneja, T. Reiss, C. Kehlet, T. Schulte- Valido, A. Lita, T. Gerrits, S. W. Nam, W. S. Koltham- Herbruggen, and S. J. Glaser. Optimal control of mer, J. Huh, and I. A. Walmsley. Approximating coupled spin dynamics: design of NMR pulse sequences vibronic spectroscopy with imperfect quantum optics. by gradient ascent algorithms. J. Magn. Reson., J. Phys. B: At. Mol. Opt. Phys., 51(24):245503, 2018. 172(2):296–305, 2005. https://doi.org/10.1016/j. https://doi.org/10.1088/1361-6455/aaf031. jmr.2004.11.004. [200] C. S. Wang, J. C. Curtis, B. J. Lester, Y. Zhang, [188] P. de Fouquieres, S. G. Schirmer, S. J. Glaser, Y. Y. Gao, J. Freeze, V. S. Batista, P. H. Vaccaro, and I. Kuprov. Second order gradient ascent pulse I. L. Chuang, L. Frunzio, L. Jiang, S. M. Girvin, engineering. J. Magn. Reson., 212(2):412–417, 2011. and R. J. Schoelkopf. Efficient multiphoton sampling https://doi.org/10.1016/j.jmr.2011.07.023. of molecular vibronic spectra on a superconducting [189] F. Dolde, V. Bergholm, Y. Wang, I. Jakobi, B. Nay- bosonic processor. Phys. Rev. X, 10(2):021060, 2020. denov, S. Pezzagna, J. Meijer, F. Jelezko, P. Neumann, https://doi.org/10.1103/PhysRevX.10.021060. T. Schulte-Herbruggen, J. Biamonte, and J. Wrachtrup. [201] Q. Zhuang, Z. Zhang, and J. H. Shapiro. Distributed High-fidelity spin entanglement using optimal control. quantum sensing using continuous-variable multipartite Nat. Commun., 5(1):3371, 2014. https://doi.org/10. entanglement. Phys. Rev. A, 97(3):032329, 2018. 1038/ncomms4371. https://doi.org/10.1103/PhysRevA.97.032329. [190] B. E. Anderson, H. Sosa-Martinez, C. A. Riofr´ıo,I. H. [202] Q. Zhuang, J. Preskill, and L. Jiang. Distributed Deutsch, and P. S. Jessen. Accurate and robust unitary quantum sensing enhanced by continuous-variable error transformations of a high-dimensional quantum system. correction. New J. Phys., 22(2):022001, 2020. https: Phys. Rev. Lett., 114(24):240401, 2015. https://doi. //doi.org/10.1088/1367-2630/ab7257. org/10.1103/PhysRevLett.114.240401. [203] Y. Xia, W. Li, W. Clark, D. Hart, Q. Zhuang, and [191] M. Abdelhafez, D. I. Schuster, and J. Koch. Gradient- Z. Zhang. Demonstration of a reconfigurable entangled based optimal control of open quantum systems using radio-frequency photonic sensor network. Phys. Rev. quantum trajectories and automatic differentiation. Lett., 124(15):150502, 2020. https://doi.org/10. Phys. Rev. A, 99(5):052327, 2019. https://doi.org/ 1103/PhysRevLett.124.150502. 10.1103/PhysRevA.99.052327. [204] K. Noh, S. M. Girvin, and L. Jiang. Encoding [192] M. Y. Niu, S. Boixo, V. N. Smelyanskiy, and H. Neven. an oscillator into many oscillators. Phys. Rev. Universal quantum control through deep reinforcement Lett., 125(8):080503, 2020. https://doi.org/10. learning. npj Quantum Inf., 5(1):33, 2019. https:// 1103/PhysRevLett.125.080503. doi.org/10.1038/s41534-019-0141-3. [205] H. K. Lau and M. B. Plenio. Universal quantum [193] X. M. Zhang, Z. Wei, R. Asad, X. C. Yang, and computing with arbitrary continuous-variable encoding. X. Wang. When does reinforcement learning stand Phys. Rev. Lett., 117(10):100501, 2016. https://doi. out in quantum control? A comparative study on org/10.1103/PhysRevLett.117.100501. state preparation. npj Quantum Inf., 5(1):85, 2019. 25

[206] S. Chakram, A. E. Oriani, R. K. Naik, A. V. Dixit, dl.acm.org/doi/10.5555/2011665.2011666. K. He, A. Agrawal, H. Kwon, and D. I. Schuster. [218] W. L. Ma, M. Zhang, Y. Wong, K. Noh, S. Rosenblum, Seamless high-Q microwave cavities for multimode P. Reinhold, R. J. Schoelkopf, and L. Jiang. Path- circuit QED. arXiv:2010.16382 [quant-ph], 2020. independent quantum gates with noisy ancilla. Phys. https://arxiv.org/abs/2010.16382. Rev. Lett., 125(11):110503, 2020. https://doi.org/ [207] B. M. Terhal. Quantum error correction for quantum 10.1103/PhysRevLett.125.110503. memories. Rev. Mod. Phys., 87(2):307, 2015. https: [219] S. Glancy and E. Knill. Error analysis for encoding a //doi.org/10.1103/RevModPhys.87.307. qubit in an oscillator. Phys. Rev. A, 73(1):012325, 2006. [208] A. Gyenis, P. S. Mundada, A. Di Paolo, T. M. https://doi.org/10.1103/PhysRevA.73.012325. Hazard, X. You, D. I. Schuster, J. Koch, A. Blais, [220] K. Fukui, A. Tomita, and A. Okamoto. Analog quantum and A. A. Houck. Experimental realization of error correction with encoding a qubit into an oscillator. an intrinsically error-protected superconducting qubit. Phys. Rev. Lett., 119(18):180507, 2017. https://doi. arXiv:1910.07542 [quant-ph], 2019. https://arxiv. org/10.1103/PhysRevLett.119.180507. org/abs/1910.07542. [221] K. Fukui, A. Tomita, and A. Okamoto. Tracking quan- [209] A. Grimm, N. E. Frattini, S. Puri, S. O. Mundhada, tum error correction. Phys. Rev. A, 98(2):022326, 2018. S. Touzard, M. Mirrahimi, S. M. Girvin, S. Shankar, and https://doi.org/10.1103/PhysRevA.98.022326. M. H. Devoret. Stabilization and operation of a Kerr- [222] C. Vuillot, H. Asasi, Y. Wang, L. P. Pryadko, and cat qubit. Nature, 584(7820):205–209, 2020. https: B. M. Terhal. Quantum error correction with the //doi.org/10.1038/s41586-020-2587-z. toric Gottesman-Kitaev-Preskill code. Phys. Rev. [210] W. C. Smith, M. Villiers, A. Marquet, J. Palomo, M. R. A, 99(3):032344, 2019. https://doi.org/10.1103/ Delbecq, T. Kontos, P. Campagne-Ibarcq, B. Doucot, PhysRevA.99.032344. and Z. Leghtas. Magnifying quantum phase fluctuations [223] K. Noh and C. Chamberland. Fault-tolerant with Cooper-pair pairing. arXiv:2010.15488 [quant-ph], bosonic quantum error correction with the surface– 2020. https://arxiv.org/abs/2010.15488. Gottesman-Kitaev-Preskill code. Phys. Rev. A, [211] L. Sun, A. Petrenko, Z. Leghtas, B. Vlastakis, 101(1):012316, 2020. https://doi.org/10.1103/ G. Kirchmair, K. M. Sliwa, A. Narla, M. Hatridge, PhysRevA.101.012316. S. Shankar, J. Blumoff, L. Frunzio, M. Mirrahimi, [224] N. C. Menicucci. Fault-tolerant measurement-based M. H. Devoret, and R. J. Schoelkopf. Tracking quantum computing with continuous-variable cluster photon jumps with repeated quantum non-demolition states. Phys. Rev. Lett., 112(12):120504, 2014. https: parity measurements. Nature, 511(7510):444–448, 2014. //doi.org/10.1103/PhysRevLett.112.120504. https://doi.org/10.1038/nature13436. [225] K. Fukui, A. Tomita, A. Okamoto, and K. Fujii. [212] Z. Leghtas, S. Touzard, I. M. Pop, A. Kou, B. Vlastakis, High-threshold fault-tolerant quantum computation A. Petrenko, K. M. Sliwa, A. Narla, S. Shankar, M. J. with analog quantum error correction. Phys. Rev. Hatridge, M. Reagor, L. Frunzio, R. J. Schoelkopf, X, 8(2):021054, 2018. https://doi.org/10.1103/ M. Mirrahimi, and M. H. Devoret. Confining the state PhysRevX.8.021054. of light to a quantum manifold by engineered two- [226] B. W. Walshe, L. J. Mensen, B. Q. Baragiola, and photon loss. Science, 347(6224):853–857, 2015. https: N. C. Menicucci. Robust fault tolerance for continuous- //doi.org/10.1126/science.aaa2085. variable cluster states with excess antisqueezing. Phys. [213] S. Touzard, A. Grimm, Z. Leghtas, S. O. Mundhada, Rev. A, 100(1):010301, 2019. https://doi.org/10. P. Reinhold, C. Axline, M. Reagor, K. Chou, J. Blumoff, 1103/PhysRevA.100.010301. K. M. Sliwa, S. Shankar, L. Frunzio, R. J. Schoelkopf, [227] K. Fukui. High-threshold fault-tolerant quantum M. Mirrahimi, and M. H. Devoret. Coherent oscillations computation with the GKP qubit and realistically noisy inside a quantum manifold stabilized by dissipation. devices. arXiv:1906.09767 [quant-ph], 2019. https: Phys. Rev. X, 8(2):021005, 2018. https://doi.org/ //arxiv.org/abs/1906.09767. 10.1103/PhysRevX.8.021005. [228] H. Yamasaki, K. Fukui, Y. Takeuchi, S. Tani, [214] S. O. Mundhada, A. Grimm, J. Venkatraman, Z. K. and M. Koashi. Polylog-overhead highly fault- Minev, S. Touzard, N. E. Frattini, V. V. Sivak, tolerant measurement-based quantum computation: K. Sliwa, P. Reinhold, S. Shankar, M. Mirrahimi, and all-Gaussian implementation with Gottesman-Kitaev- M. H. Devoret. Experimental implementation of a Preskill code. arXiv:2006.05416 [quant-ph], 2020. Raman-assisted eight-wave mixing process. Phys. Rev. https://arxiv.org/abs/2006.05416. Applied, 12(5):054051, 2019. https://doi.org/10. [229] J. E. Bourassa, R. N. Alexander, M. Vasmer, A. Patil, 1103/PhysRevApplied.12.054051. I. Tzitrin, T. Matsuura, D. Su, B. Q. Baragi- [215] J. M. Gertler, B. Baker, J. Li, S. Shirol, J. Koch, ola, S. Guha, G. Dauphinais, K. K. Sabapathy, and C. Wang. Protecting a bosonic qubit with N. C. Menicucci, and I. Dhand. Blueprint for a autonomous quantum error correction. Nature, scalable photonic fault-tolerant quantum computer. 590(7845):243–248, 2021. https://doi.org/10.1038/ Quantum, 5:392, 2021. https://doi.org/10.22331/ s41586-021-03257-0. q-2021-02-04-392. [216] D. Gottesman. An introduction to quantum error [230] M. V. Larsen, C. Chamberland, K. Noh, J. S. correction and fault-tolerant quantum computation. Neergaard-Nielsen, and U. L. Andersen. A fault-tolerant arXiv:0904.2557 [quant-ph], 2009. https://arxiv.org/ continuous-variable measurement-based quantum com- abs/0904.2557. putation architecture. arXiv:2101.03014 [quant-ph], [217] P. Aliferis, D. Gottesman, and J. Preskill. Quantum 2021. https://arxiv.org/abs/2101.03014. accuracy threshold for concatenated distance-3 codes. [231] R. Chao and B. W. Reichardt. Quantum error Quantum Info. Comput., 6(2):97–165, 2006. https:// correction with only two extra qubits. Phys. Rev. 26

Lett., 121(5):050502, 2018. https://doi.org/10. [240] P. Campagne-Ibarcq, E. Zalys-Geller, A. Narla, 1103/PhysRevLett.121.050502. S. Shankar, P. Reinhold, L. Burkhart, C. Axline, [232] C. Chamberland and M. E. Beverland. Flag fault- W. Pfaff, L. Frunzio, R. J. Schoelkopf, and M. H. tolerant error correction with arbitrary distance codes. Devoret. Deterministic remote entanglement of su- Quantum, 2:53, 2018. https://doi.org/10.22331/ perconducting circuits through microwave two-photon q-2018-02-08-53. transitions. Phy. Rev. Lett., 120(20):200501, 2018. [233] T. J. Yoder, R. Takagi, and I. L. Chuang. Universal https://doi.org/10.1103/PhysRevLett.120.200501. fault-tolerant gates on concatenated stabilizer codes. [241] L. D. Burkhart, J. Teoh, Y. Zhang, C. J. Axline, Phys. Rev. X, 6(3):031039, 2016. https://doi.org/ L. Frunzio, M. H. Devoret, L. Jiang, S. M. Girvin, 10.1103/PhysRevX.6.031039. and R. J. Schoelkopf. Error-detected state transfer and [234] P. Webster, S. D. Bartlett, and D. Poulin. Reducing entanglement in a superconducting quantum network. the overhead for quantum computation when noise is arXiv:2004.06168 [quant-ph], 2020. https://arxiv. biased. Phys. Rev. A, 92(6):062309, 2015. https:// org/abs/2004.06168. doi.org/10.1103/PhysRevA.92.062309. [242] K. A. Landsman, Y. Wu, P. H. Leung, D. Zhu, N. M. [235] S. Lloyd and S. L. Braunstein. Quantum computation Linke, K. R. Brown, L. Duan, and C. Monroe. Two- over continuous variables. Phys. Rev. Lett., 82(8):1784– qubit entangling gates within arbitrarily long chains of 1787, 1999. https://doi.org/10.1103/PhysRevLett. trapped ions. Phy. Rev. A, 100(2):022332, 2019. https: 82.1784. //doi.org/10.1103/PhysRevA.100.022332. [236] L. Jiang, J. M. Taylor, A. S. Sørensen, and M. D. Lukin. [243] L. Egan, D. M. Debroy, C. Noel, A. Risinger, D. Zhu, Distributed quantum computation based on small D. Biswas, M. Newman, M. Li, K. R. Brown, M. Cetina, quantum registers. Phy. Rev. A, 76(6):062323, 2007. and C. Monroe. Fault-tolerant operation of a quantum https://doi.org/10.1103/PhysRevA.76.062323. error-correction code. arXiv:2009.11482 [quant-ph], [237] H. J. Kimble. The quantum internet. Nature, 2020. https://arxiv.org/abs/2009.11482. 453(7198):1023–1030, 2008. https://doi.org/10. [244] M. Veldhorst, C. H. Yang, J. C. Hwang, W. Huang, J. P. 1038/nature07127. Dehollain, J. T. Muhonen, S. Simmons, A. Laucht, F. E. [238] C. Monroe, R. Raussendorf, A. Ruthven, K. R. Hudson, K. M. Itoh, A. Morello, and A. S. Dzurak. A Brown, P. Maunz, L. M. Duan, and J. Kim. Large- two-qubit logic gate in silicon. Nature, 526(7573):410– scale modular quantum-computer architecture with 414, 2015. https://doi.org/10.1038/nature15263. atomic memory and photonic interconnects. Phy. Rev. [245] T. F. Watson, S. G. J. Philips, E. Kawakami, D. R. A, 89(2):022317, 2014. https://doi.org/10.1103/ Ward, P. Scarlino, M. Veldhorst, D. E. Savage, M. G. PhysRevA.89.022317. Lagally, M. Friesen, S. N. Coppersmith, M. A. Eriksson, [239] C. J. Axline, L. D. Burkhart, W. Pfaff, M. Zhang, and L. M. K. Vandersypen. A programmable two-qubit K. Chou, P. Campagne-Ibarcq, P. Reinhold, L. Frunzio, quantum processor in silicon. Nature, 555(7698):633– S. M. Girvin, L. Jiang, M. H. Devoret, and R. J. 637, 2018. https://doi.org/10.1038/nature25766. Schoelkopf. On-demand quantum state transfer and [246] H. Levine, A. Keesling, G. Semeghini, A. Omran, T. T. entanglement between remote microwave cavity mem- Wang, S. Ebadi, H. Bernien, M. Greiner, V. Vuletic, ories. Nature Phys., 14(7):705–710, 2018. https: H. Pichler, and M. D. Lukin. Parallel implementation of //doi.org/10.1038/s41567-018-0115-y. high-fidelity multiqubit gates with neutral atoms. Phys. Rev. Lett., 123(17):170503, 2019. https://doi.org/ 10.1103/PhysRevLett.123.170503.