Selected topics of quantum computing for nuclear physics

Dan-Bo Zhang,1, 2 Hongxi Xing,3, 4 Hui Yan,1, 2 Enke Wang,3, 4 and Shi-Liang Zhu1, 2 1Guangdong Provincial Key Laboratory of Quantum Engineering and Quantum Materials, School of Physics and Telecommunication Engineering, South China Normal University, Guangzhou 510006, China 2Guangdong-Hong Kong Joint Laboratory of Quantum Matter, Frontier Research Institute for Physics, South China Normal University, Guangzhou 510006, China 3Guangdong Provincial Key Laboratory of Nuclear Science, Institute of Quantum Matter, South China Normal University, Guangzhou 510006, China 4Guangdong-Hong Kong Joint Laboratory of Quantum Matter, South China Normal University, Guangzhou 510006, China (Dated: November 4, 2020) Nuclear physics, whose underling theory is described by quantum gauge field coupled with matter, is fundamentally important and yet is formidably challenge for simulation with classical computers. Quantum computing provides a perhaps transformative approach for studying and understanding nuclear physics. With rapid scaling-up of quantum processors as well as advances on quantum algorithms, the digital quantum simulation approach for simulating quantum gauge fields and nuclear physics has gained lots of attentions. In this review, we aim to summarize recent efforts on solving nuclear physics with quantum computers. We first discuss a formulation of nuclear physics in the language of quantum computing. In particular, we review how quantum gauge fields (both Abelian and non-Abelian) and its coupling to matter field can be mapped and studied on a quantum computer. We then introduce related quantum algorithms for solving static properties and real-time evolution for quantum systems, and show their applications for a broad range of problems in nuclear physics, including simulation of lattice gauge field, solving nucleon and nuclear structure, quantum advantage for simulating scattering in quantum field theory, non-equilibrium dynamics, and so on. Finally, a short outlook on future work is given.

I. INTRODUCTION been an incentive for building quantum computers. Ba- sically, quantum simulation can be divided into simulat- ing static properties and real-time evolution of a physi- Understanding how elementary particles form nuclear cal system [16]. The corresponding quantum algorithms matter is fundamentally important. While the underling have been developed at the early age of quantum comput- basic physical theory, namely quantum chromodynam- ing based on ideal quantum computers. Recent scaling- ics (QCD), can be formulated concisely, it is notoriously up of quantum processors enables us to simulate static hard to solve [1]. This is because interactions between properties and real-time evolution of a quantum system quarks and gluons follow very distinct behavior: quarks to a larger size with noisy qubits. In such near-term are asymptotically free at the high-energy scale, while noisy intermediate quantum(NISQ) era [17], quantum al- strong couplings exist among quarks and gluons at low- gorithms are designed suitable for near-term quantum de- energy. In this regard, QCD can be non-perturbative, vices. Variational quantum algorithms are remarkable as leading to a failure of the perturbative expansion of Feyn- candidates for fully exploiting the power of NISQ quan- man diagrams which have gained remarkable success in tum computers. It receives special interests as it pro- (QED). Lattice QCD has been vides a practical approaches for solving quantum chem- developed [2–4], which can rely on computational meth- istry [18–23], quantum many-body systems [24–26], and ods such as quantum Monto Carlo [5, 6] and tensor net- many other applications [27–29].

arXiv:2011.01431v1 [quant-ph] 3 Nov 2020 works [7–14]. The former suffers from the fermion sign Nuclear physics can be considered as quantum many- problem. The tensor network approach expresses many- body physics on and below the subatomic scale, and it is body wavefunction in a compressed way which is free of of great advantages in using quantum computers for ul- sign problem, but it may have an exponential growth of timate solutions. However, simulation of nuclear physics complexity for evaluating physical observations. While meet some special challenges that are different than those those numeral efforts are still pushing the limits on what of quantum chemistry and quantum many-body physics quantum many-body problems can be solved on classical in condensed matter. Firstly, nuclear physics is studied computers, they will meet some intrinsic difficulties that via quantum field theory on continuous space-time, pos- is guaranteed by computational complexity, which claims sible with infinite dimensional degrees of freedom per vol- quantum many-body problems are NP-hard [6]. ume, and they should be encoded efficiently into an finite To overcome the intrinsic difficulty of simulating a number of qubits on lattice. Moreover, the underling the- quantum world with classical computers, Feynman pro- ory for nuclear physics should respect gauge-invariance, posed quantum hardware as simulation platforms in and in the corresponding quantum simulation, a physi- 1983 [15]. Since then, simulation of physical systems has cal Hilbert space should be kept. Those two challenges 2 not only call for a formulation of lattice systematic formulations of lattice gauge theory suitable for the purpose of efficient quantum computing [30, 31], on quantum computers have been investigated, espe- but also the issue of realization of reliable simulation on cially on how to reduce the desired quantum resource noisy quantum processors becomes tremendous impor- for gauge field which has much redundancy. The ap- tant [30, 32, 33]. proach from lattice gauge theory, however, is still too Efforts of simulating nuclear physics often benefit from resource demanding for simulating nuclear physics in- advances on quantum simulations of condensed matter terested in experimental observations. Another practi- or electronic structures of molecules. There are differ- cal strategy, mostly pushed by the community of high- ent approaches for tackling this problem, and we may energy and nuclear physics, is to incorporate quantum loosely divide it into two types: analog and digital quan- algorithms as subroutines and wisdom of classical meth- tum simulation, which may be developed with different ods are exploited. This permits valuable quantum re- motivations. Analog quantum simulation stands for early sources concentrated on classical difficulty parts, in order attempts for simulating gauge lattice theory, when peo- to exploit near-term noisy quantum processors to solve ple have realized that cold atoms can simulate quan- complicated nuclear physics. Such a hybrid quantum- tum many-body problems in condensed matter. Differ- classical strategy has been used in studying parton distri- ent platforms for quantum simulation of lattice gauge butions [65], evolution of non-equilibrium thermal states field have been proposed, including cold atoms [34–45], such as quark-gluon plasma [63], and so on. trapped ions [25, 46–50], superconducting circuit [51, 52], Although sufficient advances have been made, quan- and Rydberg atoms [53–57]. Existing reviews can refer tum simulation of high-energy nuclear physics is still a to Ref. [58] and Ref. [42] for details, and related arti- young field, and it is just beginning to bring more re- ficial gauge fields for ultracold atoms were reviewed in searchers into this field. In this review, we aim to sum- Ref. [59, 60]. Those proposals stress on how to realize marize recent advances, give basic concepts on quantum dynamical gauge field with controllable atomic interac- computing and quantum algorithms, introduce several tions. It is realized all proposals would inevitably re- representative works, and propose future directions. The quire very complicated setup, suggesting that simulation review will be organized as follows. of gauge field and nuclear physics is in general difficult In Sec. II, we first discuss how nuclear physics prob- and rather resource consuming. To date, a building block lems can be mapped into formulas of lattice field theory for simulating a coupling of gauge-field with fermion has expressed with qubits, which can be solved on a quantum been implemented in cold-atom experiment [45], paving computer. Specifically, we reveal that the way of treating the way for realistic analog quantum simulation of gauge gauge field may be the key ingredient. We then introduce lattice theory. quantum algorithms related for simulating quantum sys- The approach of digital quantum simulation for gauge tems, including both static properties and real-time evo- field theory and nuclear physics begins later but receives lution. more attention recently [25, 30, 31, 48, 61–87]. The In Sec. III, we give some specific topics and representa- digital way is programmable by compiling all operators tive examples on the applications of quantum computing into basic quantum gates and thus is much more flexible for nuclear physics. We first elaborate on a prototypi- for simulating different quantum systems [88]. In 2012, cal example, the 1+1D . This model, Preskill et al proves an exponential quantum speed-up although describing quantum electrodynamics (QED), in simulation of scattering problem for scalar relativis- shares lots of key features with QCD, such as confine- tic quantum field theory with self-interactions (φ4 the- ment and a topological theta vacuum. Thus, this mini- ory) [61] , with time complexity scaling up polynomial in mal model can be used as a testbed for simulating lattice the number of particles, their energy and the desired pre- gauge field. We also give examples on hybrid quantum- cision, remarkably in the non-perturbative regime where classical approach which uses quantum algorithms as a classical algorithms fail to work. Although elaborated on subroutine, including examples of parton physics and a concrete model, it puts a solid base for simulating high- evolution of non-equilibrium thermal states. energy physics with quantum advantages from the aspect Finally, in Sec. IV, we give outlooks and summaries. of computational complexity. More quantum algorithms have proposed to simulate lattice gauge field and nuclear physics on near-term quantum processors. While lots techniques can inherit from those of quantum computa- tional chemistry, one key difference is the requirement to II. NUCLEAR PHYSICS ON A QUANTUM COMPUTER deal with gauge field and gauge-invariance. A very suc- cessful example is that gauge degree of freedom for the 1+1D Schwinger model can be eliminated [25, 48, 89]. In this section, we first discuss how to map nuclear With this reduction, simulations of ground state and physics problems on to a quantum computer, by refor- real-time evolution of 1+1D Schwinger model have been mulating gauge quantum field in the language of qubits. demonstrated on trapped-ion platforms [25, 48]. How- Then, we introduce quantum algorithms for solving static ever, in general such an elimination is impossible, and and dynamical properties of nuclear physics. 3

A. Map nuclear physics onto a quantum computer map it into a task solvable on a quantum computer with a finite number of qubits under desired precision [61, 66]. The underling theory for nuclear physics is quantum Another important ingredient in the encoding stage is field theory describing fermionic matter coupled with how to express gauge field with qubits. It is impor- bosonic gauge fields on a continuous space-time back- tant to enforce local gauge-invariance (general Gauss law) ground. However, a quantum computer consists of ar- for simulating lattice gauge theory on a quantum com- rays of qubits, and each qubit owns 2-dim Hilbert space. puter [14, 30–32] which corresponds to keep the simula- To simulate nuclear physics on a quantum computer, it tion in the physical Hilbert space. demands for a mapping of the original physical degree of Unitary evolution. Physical processes in nuclear freedom onto qubits at the first stage. physics can be both unitary and nonunitary, while states On a quantum computer, the basic unit to encode in- on a quantum computer evolve in a unitary process. For formation is quantum bit, which can store a superposed an unitary evolution such as evolutions of the Hamil- state of |0i and |1i. In general, information is expressed tonian, one can decompose the unitary operator into a as a quantum state of multiple qubtis. With several hun- sequence of basic single and two-qubit quantum gates dreds of qubits, a quantum computer can store a quan- with product-formula, by cutting the evolution time into tum state that is beyond the capacity of classical com- short time periods. As an nonunitary process can be em- puters. The quantum state is manipulated on a quantum bedded in an unitary process of a larger system, one can computer with unitary evolutions, which can be decou- introduce some auxiliary qubits to implement the nonuni- pled into a sequence of quantum gates from a universal tary evolution. A common applied algorithmic design is set of basic quantum gates. Then, information or in- by linear-combination-of-unitaries, using ancillary qubits terested quantities can be extracted by repeatedly mea- [90–95] or continuous-variables [95–99]. suring the final quantum states. Further post-processing Measurements. In nuclear physics, we often need on a classical computer may be required. The physical to get distributions of particles, such as in the scatter- process and its corresponding simulation on a quantum ing problems. On a quantum computer, however, mea- z computer is illustrated in Fig. 1. surement often refer to computational basic (σ measure- Nuclear physics includes problems related to steady ments), which may not directly correspond to find par- states, such as structure of nucleons and phases of matter, ticles in specified directions or momentums. One can and real-time evolution, such as scattering problem and synthesize such physical measurements by unitary evolu- evolution of nuclear matter. While preparation of steady tions followed by measurements on computational basis. states and simulating the real-time evolution have been Digital vs analog quantum computer. It should standard techniques in quantum computing and variants be reminded of a difference between digital and analog of approaches have been developed, it is important to quantum computers. Analog quantum computers are de- firstly convert nuclear physics into formulas that a quan- signed specifically for simulating quantum systems. Un- tum computer can handle with. Keep in mind that basic like a digital quantum computer that uses an univer- components of a quantum computer are qubits for en- sal set of quantum gates to construct all unitary pro- coding information, universal set of single and two-qubit cess, an analog quantum computer can directly simu- quantum gates for manipulating information, and mea- late a Hamiltonian and its real-time evolution, by en- surements for extracting information. To study nuclear gineering the desired interactions. This limits its capac- physics on a quantum computer, those field operators ity for simulating different systems, while on the other should be converted into qubit operators, and the space- hand, it can be easier to implement. Representative time should be discretized, and some symmetries should physical systems for analog quantum computers include be considered, especially the local gauge invariance (gen- cold atoms [59], trapped ions, and superconducting cir- eral Gauss law). We discuss them separately. cuits. Nevertheless, analog quantum simulator can be Encode fields into qubits on lattice. Quantum programmable, enlarging its capacity of simulating a dif- fields usually are described on a continuous space-time ferent target Hamiltonian with a controllable Hamilto- and they can be bosonic (infinite dimension) or fermionic, nian on the physical platform using variational methods. while qubits on a quantum computer is defined on a lat- For instance, ground state of the 1+1D Schwinger model tice with each site resides one qubit. To encode quantum has been simulated in the trapped-ion system with long- state of a quantum field on a quantum computer, we interaction between qubits [25]. should discretize the space as a lattice, encode each local bosonic field into a finite number of qubits by a trunca- tion, and encode a local fermionic field into a sequence B. Quantum algorithms for simulation of qubits by nonlocal transformations. Remarkably, to reduce the computational resource yet retain the essen- To exploit the power of quantum computers for sim- tial physics, the discretization of space and the cut-off ulating nuclear physics, quantum algorithms are indis- bosonic field should respect some physical arguments. In pensable. After having converting a nuclear physics prob- other words, the original nuclear physics problem may lem on a quantum computer, the remaining task is to have an infinite degree of freedoms and it is important to design a quantum algorithm to solve it efficiently. Al- 4

FIG. 1. Illustration of solving nuclear physics on a quantum computer. The upper is a physical process of scattering in the experiment of Large Hadron Collide (from https://www.symmetrymagazine.org/article/lhc-scientists-detect-most-favored- higgs-decay), while the bottom depicts a corresponding simulation with a quantum computer. though quantum algorithms can be counter-intuitive and a destructive error interference that errors will cancel in requires special efforts to learn in general, it can be very different short time periods [103]. This indicates product- natural for physicists when using quantum algorithms for formula may be more efficient than expected. solving quantum systems. Those quantum algorithms may be classified as two kinds: preparing steady states Solving steady states, including eigenstates and ther- of a quantum system and simulating the real-time evolu- mal states, is important for understanding static proper- tion. In the following, we give a brief introduction. ties of quantum systems, but can be formidable challenge Simulating Hamiltonian evolutions e−iHt is typically on a classical computer. Such a basic task is compara- hard for classical computers, due to the growth of com- tively harder on a quantum computer than simulation of plexity of quantum states with increasing t. However, it real-time evolution, and lots of efforts have been devoted can be very directly implemented on a quantum com- into it. The first textbook algorithm is quantum phase es- puter, by decomposing e−iHt into a set of quantum timation [88], where eigenvalues of Hamiltonian are writ- PM ten onto ancillary qubits and then ancillary qubits are gates [16]. For a Hamiltonian H = Hi with lo- i=1 measured to read eigenvalue as well as associated eigen- cal terms Hi that [Hi,Hj] = 0 for i 6= j, the most basic states. Concretely, consider H|uni = En|uni and the technique is to use a product formula ( also named as P Trotter formula), which is initial state is |ψi = n cn|uni, then each eigenstate can 2 be obtained with a probability |cn| . Thus it is important " M #N that the target eigenstate should have a large weighting −iHt Y −iH t e ≈ e i N . (1) in the initial state. Quantum phase estimation can be i=1 understood as signal processing from the time domain to the frequency domain. It involves Hamiltonian evolution For an desired precision , the time complexity is O(t2/). at different time periods, and then uses quantum Fourier It is interesting to note that while such a formula is transformation to find eigenvalues at frequency domain. simple, it is powerful even on the near-term quantum Both Hamiltonian evolution and quantum Fourier trans- devices, and the performance can be comparable with formation are resource-costing, and consequently, quan- more advanced quantum algorithms for Hamiltonian sim- tum phase estimation is not suitable on the near-term ulations [100–102]. Remarkably, it is found that time- quantum devices. Quantum adiabatic algorithm pro- complexity can be reduced to O(t/) using the product- vides an approach to prepare ground state of a Hamilto- formula when the Hamiltonian can be written as H = nian [104]. It starts from a Hamiltonian whose ground HA + HB, where local terms in HA or HB commute to state is easy to prepare, and by adiabatically tuning the each other but [HA,HB] 6= 0. This is because there is Hamiltonian into the target Hamiltonian, the final state 5 will be the ground state for the target Hamiltonian. The been demonstrated on current quantum devices, includ- performance of the quantum adiabatic algorithm relies ing both ground state [25] and real-time evolution [48]. on a gap between the first excited state and the ground Moreover, the Schwinger model can reveal some impor- state along the adiabatic path. Thus it requires a good tant features shared with QCD. We thus use this model design for the adiabatic path which can be nontrivial. In to illustrate a work flow for studying nuclear physics on a practice, the adiabatic evolution should be very slow, and quantum computer: how the original formula of quantum may be hard to finish within the coherent time. Thus, field problem should be mapped into a lattice spin model, quantum adiabatic algorithm, although can be universal, what quantum algorithm should be chosen to solve the may not be suitable in the era of NISQ. lattice spin model, and how desired physical quantities Another important class of quantum algorithms is vari- can be accessed with measurements on a quantum com- ational quantum eigensolver (VQE) [18, 19, 21, 22, 24, puter, possible with post-processing. 25, 105–107], which is considered promising for fully ex- We introduce the Hamiltonian for the Schwinger model ploiting the power of NISQ quantum devices. It can rely with fixing gauge A0(x) = 0, on a shallow quantum circuit and a moderate number Z † 0 1 of qubits for solving classical intractable problems. The Hˆ = dx[Ψ (x)γ γ (−i∂1 + gAˆ1(x))Ψ(x) quantum circuit is parameterized and the parameters can 1 be obtained by minimizing the energy. The optimization +mΨ†(x)γ0Ψ(x) + Eˆ2(x)], (2) is a hybrid quantum-classical one. In this sense, varia- 2 tional quantum algorithms give a new paradigm for algo- where Ψ†(x), Ψ(x) are fermionic field operators, the elec- rithmic design: rather than directly designing a quantum ˆ ˆ algorithm, it trains a quantum algorithm for a given task, tric field E(x) and the vector potential A1(x) satisfy ˆ ˆ ˆ ˆ 0 0 by optimizing a cost function. The VQE is designed for −E(x) = ∂0A1(x) and [A1(x), E(x )] = −iδ(x − x ), 0 z 1 y solving the ground state by minimizing the energy, and γ = σ and γ = iσ . In addition, the local variants of VQE have been developed for obtaining ex- conservation (Gauss law) should be respected, ∂1E = † cited states [108–111], thermal states at finite tempera- gΨ (x)Ψ(x), which is a consequence of local U(1) gauge ture [112–116], quantum imaginary time evolution [117], symmetry. and general quantum processes [118]. A recasting of Eq. (2) onto a lattice is the Kogut- Susskind formula [119], which puts fermions on the sites and gauge field on bonds. Concretely, two-component III. APPLICATIONS FOR NUCLEAR PHYSICS fermions are defined on the even and odd sites respec- √ √ † tively: Φ2j = aΨ(x2j) and Φ2j−1 = aΨ (x2j−1). Gauge fields on the bonds are θ = −agA(x + a ) Although at an early stage, quantum computing has j,j+1 j 2 1 a ˆ ˆ had a broad applications for nuclear physics problems, and Lj,j+1 = g E(xj + 2 ), which satisfy [θj,j+1, Lj0,j0+1] = ˆ and it is beyond the scope of this review for a throughout iδj,j0 . By those representations, the lattice Hamiltonian investigation. Rather, we attend to use some prototypi- under open boundary reads, cal examples to illustrate how nuclear physics problems N−1 can be solved on a quantum computer. With concrete 1 X ˆ Hˆ = [Φ†eiθj,j+1 Φ + h.c.] examples, the basic concepts and procedures may be re- 2a j j+1 vealed. We start with the 1+1D Schwinger model, which j=1 N N−1 is a prototypical model for simulating gauge field with X g2a X + m (−1)jΦ†Φ + Lˆ2 . (3) a state-of-art realistic quantum computer. Other exam- j j 2 j,j+1 ples include interesting problems ranging from scatter- j=1 j=1 ing, evolution of non-equilibrium thermal states, nuclear The lattice Hamiltonian has already allowed for nu- structure at both high-energy and low-energy, and so on. meral simulation. Still, a challenge is to simulate the con- In addition, we introduce a work that give the time com- tinuous variable gauge field with discrete-variable qubits. plexity of solving scattering for a scalar quantum field, Approximations can be applied, either by reduction the which shows quantum advantage. U(1) to Zn gauge field, or by a truncation into finite- dimension Hilbert space. Those can be standard schemes for all Abelian and non-Abelian gauge fields. Fortu- A. Simulation of real-time evolution and ground nately, the 1+1D Schwinger model is special as there is state of lattice gauge theory: the Schwinger model essentially no magnetic field. Moreover, the electric field and fermions are locally dependent into each other. This The capacity of quantum simulation of lattice gauge allows us to eliminate the gauge field totally by using the j field implies that nuclear physics can be studied on a Gaussian law, Lˆ − Lˆ = Φ†Φ − 1−(−1) , which quantum computer. However, the required quantum j,j+1 j−1,j j j 2 ˆ Pj † 1−(−1)i resource is too demanding in the NISQ era, and it is determines Lj,j+1 = 0 + i=1[Φi Φi − 2 ], with 0 more practical to start with some minimal models. The as the value of Lˆ at the left boundary. Now, the lattice Schwinger model is such a prototypical model that has model can be represented solely with fermionic operators, 6

† Q z + and by Jordan-Wigner transformation, Φj = i

Schwinger model for a large range of parameters, espe- An interesting demonstration of quantum computing cially around the quantum critical point where quantum for atomic nucleus was reported in Ref. [120], which correlations is strong. The work opens a new direction adopts the VQE to solve the binding energy of the for simulating large-size quantum system exploiting ana- deuteron on a cloud quantum computer. With pionless log quantum simulators with great flexibility by using the EFT, the deuteron Hamiltonian with a truncation of N variational method. basis reads,

We comment that elimination of the gauge field com- N−1 pletely is only possible for 1+1D U(1) gauge theory. X † HN = tnn0 an0 an, (6) Nevertheless, the idea of reducing quantum resources for nn0=0 representing gauge field is important and possible, since † gauge field has a redundancy in description. Many efforts where an(an) create (annihilate) a deuteron in the have been denoted for this mapping stage, including non- harmonic-oscillator s-wave sate |ni, and the hopping 0 Abelian gauge field in 1+1D [12, 67], 1+2D Abelian [69] strength tnn0 = hn |(T + V )|ni can be calculated un- and non-Abelian gauge field, and so on . der the harmonic oscillator basis. As the deuteron is described in EFT via a contact interaction, only nearest hopping exists (see Ref. [120] for detailed values). By the Jordan-Wigner transformation, the Hamiltonian with in- B. Nuclear structure creasing N (up to N = 3) can be written as, z z One center motivation of nuclear physics is to explore H2 = 5.906709I + 0.218291σ0 − 6.125σ1 x x y y the internal structure of nucleus, which relies on an in- − 2.143304(σ0 σ1 + σ0 σ1 ) z creasing of energy scale of probes for resolving finer struc- H3 = H2 + 9.625(I − σ2 ) tures. Protons and neutrons are basic ingredients of a + 3.913119(σxσx + σyσy). nucleus at low-energy scales, but a proton or a neutron 1 2 1 2 itself is a collection of quarks and gluons at large en- (7) ergy scales. Colliders have been built for detecting those With the harmonic oscillator basis, the deuteron structures at different energy scales from scattering cross ground state will be a superposition state at dif- sections. Yet, numeral methods meet challenges that are ferent basis, which is clear due to the hopping common for quantum many-body systems, especially for terms. The VQE approach uses an unitary-coupled- solving the structure of hardrons, which in nature is non- cluster (UCC) ansatz. Defining single excitation process, † † x y y x perturbative. We review recent efforts for solving nuclear θ(a a1−a a0) iθ/2(σ σ −σ σ ) U01(θ) = e 0 1 = e 0 1 0 1 and U02(η) = structure on a quantum computer [65, 120], for an atomic † † x z y y z x η(a a2−a a0) iη/2(σ σ σ −σ σ σ ) nucleus and an hadron respectively, which adopt quite e 0 2 = e 0 1 2 0 1 2 , then we can employ different strategies. ansatz U(θ) = U01(θ) for H2, U(η, θ) = U01(θ)U02(η) for H3, both performing on the initial state |100i. Param- eters η and θ are optimized by minimizing the energy of H2 and H3 with their corresponding variational wave- 1. Binding energy of atomic nucleus funciton, respectively. Further, the extrapolation to the infinite space should be employed with data obtained at From the aspect of low-energy nuclear physics, an finite size. The numeral results are shown to be very atom consists of interacting protons and neutrons that accurate with exact diagonalization. are bounded together. Although lattice QCD can be We point out that solving the Hamiltonian in Eq. (6) is applied, a better starting point is to use effective field actually easy by exact diagonalization since it involves a theory (EFT) that protons and neutrons are relevant de- tridiagonal matrix. Why bother to use a quantum com- gree of freedoms. For light nucleus, pionless EFT pro- puter? Moreover, a N × N matrix is easy to diagonalize vides systematically improvable approach for modeling on a classical computer for N ∼ 100, but can be challenge nuclear interaction. For this bound-state problem, a for the current quantum devices. It should be reminded second-quantization of the Hamiltonian can be obtained that the VQE approach becomes valuable when solving by using proper local basis, e.g., a common choice can nuclear structure becomes a many-body problem, and a be the harmonic oscillator basis. As a truncation of lo- simple exact diagonalization will face a M ⊗ M matrix cal basis can be applied, the problem can be solved in (M grows exponentially with N). It is expected that a a finite-dimensional Hilbert space. Further, the second- road map for extending the scheme for quantum many- quantization fermionic Hamiltonian can be mapped into body nuclear physics is required. a qubit Hamiltonian. Such a procedure is very familiar in quantum chemistry and quantum many-body problems in condensed matter. And one can immediately recog- 2. Parton physics nize that the problem is much like quantum chemistry, and it is expected that an exponential wall prevents clas- Solving internal structure of a hadron is quite a differ- sical methods when the system size is becoming large. ent story. The partonic structure of a hadron depends 8 on the energy scale to see it. For instance, a proton is a Noted that there is no Wilson line in the definition of bounded state of three valence quarks at low energy scale, Eq. (10) as gauge field is not involved in the Thirring and no individual quark has been observed in experi- model in Eq. (8). |P i stands for the hadron state at mo- ments due to color confinement. On the other hand, at mentum P . In this sense, PDF is the Fourier transform higher energy scale in deep inelastic scattering (DIS) with of a time-separated correlator of the hadron. Two key large momentum transfer, the hadron shall be modeled procedures on a quantum computer are clear: firstly pre- as a collection of charged point-like constituents, namely pare |P i which is a quantum many-body state for the parton. The parton contributes to the cross section with with quantum numbers for representing momentum xpµ, where x is the momentum fraction of the the hadron at momentum P ; then calculate the correla- proton carried by the parton, and pµ is the momentum of tor. Since there is only one flavor of fermion, we may ex- the hadron. Understanding hadronic structure in terms press a hadron with a quantum number of three fermions, of the parton distribution functions (PDFs) over xpµ and plus fluctuations of fermion-antifermion pairs. Such a its generalization is a research frontier [121–124]. scheme seems very straightforward, but some technical The PDFs f(x) serve as an input for explaining scat- issues prevents it from being useful. Since light cone can tering cross sections in DIS experiments. However, the not be defined on the lattice until the continuum limit, parton distribution itself is hard to compute from ab ini- time-separated correlators should be evaluated with cau- tio methods due to its nonperturbative nature. There are tion. The issue can be even more complicated for gauge different approaches from first principle, such as light- theory where the Wilson line should be incorporated. front Hamiltonian and lattice QCD, which still awaits The second scheme exploits a connection between PDF for exascale supercomputer to verify. Here we introduce and the hadronic tensor. The hadronic tensor for the d- a recent work that points out an avenue on a near-term dimension theory is defined as correlations of currents quantum computer. J µ, Ref. [65] presents two different schemes for calculating Z parton distribution on a quantum computer: direct cal- µν d iqy µ ν culation of PDFs using operator formula and by evalua- W (q) = Re d xe hP |T {J (y)J (0)}|P i, (11) tion of the hadronic tensor. For illustration purpose two schemes are demonstrated with 1+1D Thirring model. which can characterize the many-body wavefunction The lattice Hamiltonian writes, |P i through correlations and consequently reveal the hadronic structure to some degree. The connection be- X 1 H = (−1)j[c†c + h.c.] tween PDF and W µν (q) is established as (via collinear 2 j j+1 j factorization), j † 2 + m(−1) cjcj − g njnj+1 (8) µν X ˆ µν W = fi ⊗ Wi . (12) † where nj = cjcj. This model is not gauge theory and thus i is different from QCD. However, evaluating the hadronic tensor will not depend on the gauge field explicitly in Here ⊗ represents convolution, i stands for parton species ˆ µν µν lepton-hadron DIS at the leading order, and thus the and Wi is the partonic tensor. Thus, with W , the second scheme is plausible and promising for studying parton distribution fi of species i can be derived. The PDF on NISQ quantum processors. The Hamiltonian nontrivial relation between PDF and hadronic tensor is shall be mapped into a spin model, indicated in the convolution. The remaining task is to evaluate W µν (q) on the state N−1 X (−1)j+1 |P i, which can refer to linear response method on a quan- H = (σxσx + σyσy ) 4 j j+1 j j+1 tum computer. Still, the preparation of the state |P i is j=1 quite involved, and Ref. [65] points out that quantum N X (−1)jm g2 adiabatic state preparation may be employed. + ( σz + σzσz ). (9) 2 j 4 j j+1 The above proposal suggests a way for studying par- j=1 ton physics on a quantum computer, but a demonstra- Similar to the Schwinger model, the bare vacuum is set at tion is still awaited. This requires to efficiently prepare m → ∞ limit, and |1i(|0i) on the even (odd) site repre- |P i (possible with VQE), evaluating the hadronic tensor, sents occupation of a fermion (antifermion), respectively. and extracting PDF through global fitting to the results The operator formula for a definition of PDF (without of hadronic tensor obtained from quantum computing. gauge link) is given by, Another important question is which lattice Hamiltonian should be a good starting point. The light-front Hamilto- Z + nian of QCD looks promising as it already have provided f(x) = dyeixP yhP |ψ¯(y)γ+ψ(0)|P i, (10) a framework for ab initio calculation of proton structure. √ √ Remarkably, VQE based on the light-front Hamiltonian where γ+ = (γ0 + γ1)/ 2 and P + = (P 0 + P 1)/ 2 has just been demonstrated for calculating structure of is the light-cone momentum for the incoming hadron. pion [75]. 9

C. Quantum advantage: scattering in scalar the lattice. The initial state can be prepared in an adi- quantum field theory. abatic way, which involves two steps: firstly, two sepa- rated wave packets are produced on the vacuum of the Scattering is central to nuclear physics as it is almost non-interacting Hamiltonian H0, where the vacuum is a the only available experimental method. Consequently, Gaussian state and can be constructed exactly; Secondly, calculating scattering amplitude is what in theory needed interaction is turned on and the system evolves adia- to do for a comparison with experiments. Different ma- batically with Hamiltonian parameterized by H(s(t)) = chinery of computational methods have been developed s(t)H0 + (1 − s(t))H , where H(0) = H0, H(1) = H and but there are some fundamental challenges. The diffi- s(t) is the adiabatic path. Note that the wave packets culty can be revealed through a minimal model of quan- will propagate and broaden since it is not an eigenstate tum field theory, the φ4 theory, which is a scalar the- of H0, additional backward evolution is required to undo ory with quartic self-interactions. When the coupling unnecessary dynamical phases. The time complexity for 1+d/2 approaches the phase transition point, the perturba- the initial state procedure scales as O((1/ε) ). The tive method becomes unreliable; and even in the weak- system of two incoming particles then evolves for a period coupling regime, precision can be not controllable with that scattering occurs. Then, the interaction is adiabatic the Feynman diagram calculation. Also, it is beyond the turn-off and the scattering result is sampled by measuring ability of lattice field theory which is good at calculat- the number operator of momentum modes defined in the ing static properties, while scattering is a time-evolution free theory. With scaling analysis via effective field the- problem. On the other hand, time-evolution can be im- ory, the algorithm shows a complexity polynomial with 1 −1.5−o(1) plemented on a quantum computer and it is thus ex- the ε, e.g., in 1D its O(( ε ) ). Remarkably, at pected that scattering can be investigated efficiently. strong coupling, the time complexity still scales polyno- A technical challenge is that the number of degree of mial with 1/(λc −λ0) (λc is the quantum phase transition freedom per unit volume is infinite, and it is necessary to point), the momentum of incoming particles p, the maxi- give a mapping that can assign qubits to physical degree mum kinematically allowed number of outgoing particles of freedom, in a controllable way under desired accuracy. nout and the precision ε. This is in contrast to the known For instance, truncation of Fock space can be applied, classical algorithm scaling exponentially with 1/(λc −λ0) since the number of particles is constrained by the en- and 1/ε, thus showing an exponentially quantum speed- ergy of incoming particles. Then, given the number of up. qubits, preparation of incoming two particles as two wave The merit of this work is to make clear the point of packets, and their scattering, should be formulated in the solving quantum field theory on a quantum computer language of quantum circuit. The quantum advantage is with quantum advantage. A scheme to reach this goal that the depth of the quantum circuit scales only polyno- has also been pointed out. Still, some subroutines may mial with the number of particles, energy and the desired be improved to make the algorithm more implementable precision, for both weak and strong couplings [61]. in the NISQ era, and there indeed are some following jobs The space is discretized as a d-dimensional L × ... × L with more detailed algorithms [66]. lattice Ω with lattice spacing a. Let us start with the lattice Hamiltonian, D. Non-equilibrium dynamics X 1 1 H = ad[ π(x)2 + (5 φ)2(x) 2 2 a x∈Ω In physics we often need to study non-equilibrium 1 λ quantum systems at finite temperature. For nuclear + m φ(x)2 + 0 φ(x)4]. (13) 2 0 4! physics, quark-gluon plasma, which is produced in heavy- ion collision or in the expansion of the early Universe, Here φ(x) is field operator on the site x, π(x) is the is an outstanding example but still limited knowledge −d conjugate field satisfying [φ(x), π(y)] = ia δx,y, 5a is is known, as simulation with classical methods are very a finite-difference operator, m0 is the particle mass for hard. Basically, studies of such quantum systems involves the non-interacting theory H0 corresponding to λ0 = 0. time evolution of density matrix. As a quantum com- The first issue is to represent the infinite dimensional puter essentially manipulate pure quantum states with field φ(x) in terms of qubits. Setting φmax as a cutoff unitary operation, an extension to density matrix needs and δφ as increments, nb = O(log2(φmax/)δφ) qubits are additional treatment. One can either refer to a subsystem required per site. With a constraint of energy E, the by tracing out the ancillary, or view the density matrix p d d 2 cutoff is φmax = O( L E/a m0ε) for desired fidelity as an ensemble of pure quantum states with a classical 1 − ε to the original state. The energy constraint also distribution. p d d introduce a cutoff π(x) to πmax = O( L E/a ε). Note Consider a density matrix ρ0 as the initial state for a d that πmax = 1/(a δφ) as φ(x) and π(x) are conjugate. quantum system, then its real-time evolution under H1 d iH1t −iH1t Then, it follows that nb = O(log2(L E/m0ε)). is given by ρ(t) = e ρ0e . The initial state can Two incoming particles should be expressed as a many- be set as equilibrium state for the system under H at body initial state with two separated wave packets on inverse temperature β = 1/T , namely quantum Gibbs 10

−βH state, ρ0 ≡ ρ(β) = e , which satisfying the symmetric anticipated that more efficient schemes of expressing de- Bloch equation, gree of freedom of gauge field and its coupling to matter field would be developed. On the other hand, incorpo- dρ(β) 1 = − (H ρ + ρH ), (14) rating quantum algorithms as subroutines into a hybrid dβ 2 0 0 quantum-classical algorithm can be a very promising ap- proach for solving nuclear physics with practical results. with initial condition ρ(0) = 1. In Ref. [63], a hy- Another observation is that there is still a lack of con- brid quantum-classical scheme has been developed for sensus on the goal; perhaps it is inspiring that a road-map the real-time evolution of density matrix in the context can be proposed for demonstrating quantum advantages of nuclear physics. The initial state is obtained by solv- for nuclear physics on a quantum computer, in accor- ing Eq. (14) with density matrix quantum Monte Carlo dance with the scaling-up of near-term quantum proces- algorithm (DMQMC), leading to an approximation of sors and error mitigation techniques . Moreover, soft- P ρ˜0 ≈ ρ0, with ρ˜0 = p χp|apihbp|. Such an approxima- ware methodology is important, as it enables systematic tion is completed on a classical computer with enough developments and rapid innovations on both quantum samples of |apihbp|. With the approximated initial state, algorithms and applications to nuclear physics. Com- time-dependent observable is evaluated as (with a sub- pared with quantum computing for chemistry [125] and 1 ∗ stitution ρ˜0 → 2 (ρ ˜0 +ρ ˜0 ) ), quantum machine learning [126], open-source packages are still lacked for nuclear physics, and it awaits for such iH1t −iH1t hO(t)i = Tr(Oe ρ0e )/Trρ0 packages for making an attempt to study nuclear physics P iH1t ∗ −iH1t Tr(Oe [χp|apihbp| + χ |bpihap|e ]) on a quantum computer more friendly. ≈ p p . 2Trρ˜0 In summary, we have given an brief review on recent (15) advances on quantum computing for nuclear physics. We have clarified two different approaches from analog and The task thus can be decomposed to evaluate observable digital quantum simulator, and the review has focused O for each component, including the diagonal term, e.g., on the latter. We have outlined how degrees of free- iH1t −iH1t hap|e Oe |api, and the off-diagonal term, such as dom of nuclear physics problems can be mapped into a iH1t −iH1t hap|e Oe |bpi. The former corresponds to perform formula of qubits that is solvable on a quantum com- −iH t measurement O on a state evolved with e 1 from ini- puter, and the corresponding quantum algorithms for tial state |api or |bpi. And for the later, one can replace both static properties and real-time evolution have been initial state with √1 (|a i ± |b i). 2 p p shortly discussed. Concretely, we have given some exam- In total, the hybrid quantum-classical approach adopts ples from recent outstanding works, ranging from simu- a strategy that assign quantum resource for the real-time lation of lattice gauge theory, nuclear structure as well as evolution part which is hard for the classical computer, quantum advantage in terms of time complexity for scat- while the classical part can utilize the state-of-art classi- tering problem. Lastly, we have pointed out that lots of cal algorithm since it is good at solving steady states. A efforts are still required to make quantum computing a combination of both makes this approach promising on playground for investigating nuclear physics on near-term NISQ quantum processors for studying non-equilibrium quantum devices. dynamics.

IV. OUTLOOK AND SUMMARY ACKNOWLEDGMENTS At this stage, we have given an introductory review of quantum computing for nuclear physics, but a through- This work was supported by the Key-Area Research out investigation is still beyond the scope of this review, and Development Program of GuangDong Province given recent rapid expanding of this field. We further (Grant No. 2019B030330001), the National Key Re- briefly give an outlook that focuses on this field within search and Development Program of China (Grant No. the NISQ era. Although formulations of lattice gauge 2016YFA0301800), the National Natural Science Foun- theory for quantum computing have been developed in dation of China (Grants No. 12005065, No.12022512, a few different approaches, it is clear that there still is No.12035007, and No. 12074180), and the Key Project a considerable gap for simulating QCD on NISQ quan- of Science and Technology of Guangzhou (Grants No. tum processors with limited quantum resources, and it is 201804020055 and No. 2019050001).

[1] M. Peskin and D. Schroeder, “An introduction to quan- [2] Kenneth G. Wilson, “Confinement of quarks,” Phys. tum field theory,” (1995). Rev. D 10, 2445–2459 (1974). 11

[3] J. Kogut and L. Susskind, “Hamiltonian formulation tion of molecular energies,” Physical Review X 6, 031007 of wilson’s lattice gauge theories,” Phys. Rev. 11, 395 (2016). (1975). [21] Abhinav Kandala, Antonio Mezzacapo, Kristan Temme, [4] John B. Kogut, “The lattice gauge theory approach to Maika Takita, Markus Brink, Jerry M. Chow, and ,” Rev. Mod. Phys. 55, 775– Jay M. Gambetta, “Hardware-efficient variational quan- 836 (1983). tum eigensolver for small molecules and quantum mag- [5] L. D. McLerran and B. Svetitsky, “A monte carlo study nets,” Nature 549, 242–246 (2017). of su(2) yang-mills theory at finite temperature,” Phys. [22] Harper R. Grimsley, Sophia E. Economou, Edwin Lett. 98, 195 (1981). Barnes, and Nicholas J. Mayhall, “An adaptive vari- [6] M. Troyer and U. J. Wiese, “Computational complex- ational algorithm for exact molecular simulations on a ity and fundamental limitations to fermionic quantum quantum computer,” Nature Communications 10, 3007 monte carlo simulations,” Phys. Rev. Lett. 94 (2005). (2019). [7] Pietro Silvi, Enrique Rico, Tommaso Calarco, and [23] Frank Arute et al., “Hartree-fock on a superconducting Simone Montangero, “Lattice gauge tensor networks,” qubit quantum computer,” (2020), arXiv:2004.04174 New Journal of Physics 16, 103015 (2014). [quant-ph]. [8] L. Tagliacozzo, A. Celi, and M. Lewenstein, “Ten- [24] Jin-Guo Liu, Yi-Hong Zhang, Yuan Wan, and Lei sor networks for lattice gauge theories with continuous Wang, “Variational quantum eigensolver with fewer groups,” Physical Review X 4, 041024 (2014). qubits,” Phys. Rev. Research 1, 023025 (2019). [9] E. Rico, T. Pichler, M. Dalmonte, P. Zoller, and [25] C. Kokail, C. Maier, R. van Bijnen, T. Brydges, M. K. S. Montangero, “Tensor networks for lattice gauge theo- Joshi, P. Jurcevic, C. A. Muschik, P. Silvi, R. Blatt, ries and atomic quantum simulation,” Phys. Rev. Lett. C. F. Roos, and P. Zoller, “Self-verifying variational 112, 201601 (2014). quantum simulation of lattice models,” Nature 569, [10] Stefan Kühn, Erez Zohar, J. Ignacio Cirac, and 355–360 (2019). Mari Carmen Bañuls, “Non-abelian string breaking phe- [26] Pierre-Luc Dallaire-Demers, Michał Stęchły, Jerome F. nomena with matrix product states,” J. High Energy Gonthier, Ntwali Toussaint Bashige, Jonathan Phys. 2015, 130 (2015). Romero, and Yudong Cao, “An application bench- [11] Boye Buyens, Frank Verstraete, and Karel mark for fermionic quantum simulations,” (2020), Van Acoleyen, “Hamiltonian simulation of the schwinger arXiv:2003.01862 [quant-ph]. model at finite temperature,” Phys. Rev. D 94, 085018 [27] Eric R. Anschuetz, Jonathan P. Olson, Alán Aspuru- (2016). Guzik, and Yudong Cao, “Variational quantum factor- [12] Mari Carmen Bañuls, Krzysztof Cichy, J. Ignacio Cirac, ing,” arXiv:1808.08927 (2018). Karl Jansen, and Stefan Kühn, “Efficient basis formu- [28] Xiaosi Xu, Jinzhao Sun, Suguru Endo, Ying Li, Si- lation for (1+1)-dimensional su(2) lattice gauge theory: mon C. Benjamin, and Xiao Yuan, “Variational al- Spectral calculations with matrix product states,” Phys- gorithms for linear algebra,” (2019), arXiv:1909.03898 ical Review X 7, 041046 (2017). [quant-ph]. [13] Pietro Silvi, Yannick Sauer, Ferdinand Tschirsich, and [29] Michael Lubasch, Jaewoo Joo, Pierre Moinier, Martin Simone Montangero, “Tensor network simulation of an Kiffner, and Dieter Jaksch, “Variational quantum al- su(3) lattice gauge theory in 1d,” Phys. Rev. D 100, gorithms for nonlinear problems,” Phys. Rev. A 101, 074512 (2019). 010301 (2020). [14] Patrick Emonts and Erez Zohar, “Gauss Law, Minimal [30] Q. S. Collaboration Nu, Henry Lamm, Scott Lawrence, Coupling and Fermionic PEPS for Lattice Gauge The- and Yukari Yamauchi, “General methods for digital ories,” SciPost Phys. Lect. Notes , 12 (2020). quantum simulation of gauge theories,” Phys. Rev. D [15] R. Feynman, “Simulating physics with computers,” Int. 100, 034518 (2019). J. Theor. Phys. 21, 467 (1982). [31] Erez Zohar, Alessandro Farace, Benni Reznik, and [16] S. Lloyd, “Universal quantum simulators,” Science 273, J. Ignacio Cirac, “Digital lattice gauge theories,” Phys. 1073 (1996). Rev. A 95, 023604 (2017). [17] John Preskill, “Quantum Computing in the NISQ era [32] Jesse R. Stryker, “Oracles for gauss’s law on digital and beyond,” Quantum 2, 79 (2018), arXiv: 1801.00862. quantum computers,” Phys. Rev. A 99, 042301 (2019). [18] M. H. Yung, J. Casanova, A. Mezzacapo, J. McClean, [33] Jad C. Halimeh and Philipp Hauke, “Reliability of Lat- L. Lamata, A. Aspuru-Guzik, and E. Solano, “From tice Gauge Theories,” Phys. Rev. Lett. 125, 030503 transistor to trapped-ion computers for quantum chem- (2020). istry,” Scientific Reports 4, 3589 (2014). [34] H. P. Büchler, M. Hermele, S. D. Huber, Matthew P. A. [19] Jarrod R. McClean, Jonathan Romero, Ryan Babbush, Fisher, and P. Zoller, “Atomic quantum simulator for and Alán Aspuru-Guzik, “The theory of variational lattice gauge theories and ring exchange models,” Phys. hybrid quantum-classical algorithms,” New Journal of Rev. Lett. 95, 040402 (2005). Physics 18, 023023 (2016). [35] Sanpera A. Ahufinger V. Damski B. Sen A. Lewen- [20] P. J J O’Malley, R. Babbush, I. D Kivlichan, J. Romero, stein M and U. Sen, “Ultracold atomic gases in opti- J. R McClean, R. Barends, J. Kelly, P. Roushan, cal lattices: mimicking condensed matter physics and A. Tranter, N. Ding, B. Campbell, Y. Chen, Z. Chen, beyond,” Adv. Phys. 56, 243 (2007). B. Chiaro, A. Dunsworth, A. G Fowler, E. Jeffrey, [36] Celi A. Latorre J. I. Boada O and V. Pico, “Simula- E. Lucero, A. Megrant, J. Y Mutus, M. Neeley, C. Neill, tion of gauge transformations on systems of ultracold C. Quintana, D. Sank, A. Vainsencher, J. Wenner, T. C atoms,” New J. Phys. 12 (2010). White, P. V Coveney, P. J Love, H. Neven, A. Aspuru- [37] A. Bermudez, L. Mazza, M. Rizzi, N. Goldman, Guzik, and J. M Martinis, “Scalable quantum simula- M. Lewenstein, and M. A. Martin-Delgado, “Wilson 12

fermions and axion electrodynamics in optical lattices,” Ann. Phys-new. York. 351, 634–654 (2014). Phys. Rev. Lett. 105, 190404 (2010). [53] Celi A. Zamora A. Tagliacozzo L and M. Lewenstein, [38] Dalmonte M. Müller M. Rico E. Stebler P. Wiese U. J. “Optical abelian lattice gauge theories,” Ann. Phys. Banerjee D and P. Zoller, “Atomic quantum simulation 330, 160 (2013). of dynamical gauge fields coupled to fermionic matter: [54] L. Tagliacozzo, A. Celi, P. Orland, M. W. Mitchell, and from string breaking to evolution after a quench,” Phys. M. Lewenstein, “Simulation of non-abelian gauge theo- Rev. Lett. 109 (2012). ries with optical lattices,” Nature Communications 4, [39] Cirac J. I. Zohar E and B. Reznik, “Simulating compact 2615 (2013). quantum electrodynamics with ultracold atoms: Prob- [55] Jin Zhang, J. Unmuth-Yockey, J. Zeiher, A. Bazavov, ing confinement and nonperturbative effects,” Phys. S. W. Tsai, and Y. Meurice, “Quantum simulation of Rev. Lett. 109 (2012). the universal features of the polyakov loop,” Phys. Rev. [40] Cirac J. I. Zohar E and B. Reznik, “Simulating (2+1)- Lett. 121, 223201 (2018). dimensional lattice qed with dynamical matter using ul- [56] Hannes Bernien, Sylvain Schwartz, Alexander Keesling, tracold atoms,” Phys. Rev. Lett. 110 (2013). Harry Levine, Ahmed Omran, Hannes Pichler, Soon- [41] Erez Zohar, J. Ignacio Cirac, and Benni Reznik, “Cold- won Choi, Alexander S. Zibrov, Manuel Endres, Markus atom quantum simulator for su(2) yang-mills lattice Greiner, Vladan Vuletić, and Mikhail D. Lukin, “Prob- gauge theory,” Phys. Rev. Lett. 110, 125304 (2013). ing many-body dynamics on a 51-atom quantum simu- [42] Erez Zohar, J. Ignacio Cirac, and Benni Reznik, “Quan- lator,” Nature 551, 579–584 (2017). tum simulations of lattice gauge theories using ultracold [57] Federica M. Surace, Paolo P. Mazza, Giuliano Giudici, atoms in optical lattices,” Rep. Prog. Phys. 79, 014401 Alessio Lerose, Andrea Gambassi, and Marcello Dal- (2015). monte, “Lattice gauge theories and string dynamics in [43] K. Stannigel, P. Hauke, D. Marcos, M. Hafezi, S. Diehl, rydberg atom quantum simulators,” Physical Review X M. Dalmonte, and P. Zoller, “Constrained dynamics 10, 021041 (2020). via the zeno effect in quantum simulation: Implement- [58] U. J. Wiese, “Ultracold quantum gases and lattice sys- ing non-abelian lattice gauge theories with cold atoms,” tems: quantum simulation of lattice gauge theories,” Phys. Rev. Lett. 112, 120406 (2014). Ann. Phys. 525, 777 (2013). [44] Frederik Görg, Kilian Sandholzer, Joaquín Minguzzi, [59] Dan-Wei Zhang, Yan-Qing Zhu, Y. X. Zhao, Hui Yan, Rémi Desbuquois, Michael Messer, and Tilman and Shi-Liang Zhu, “Topological quantum matter with Esslinger, “Realization of density-dependent peierls cold atoms,” Advances in Physics 67, 253–402 (2018). phases to engineer quantized gauge fields coupled to ul- [60] D. W. Zhang, Z. D. Wang, and S.-L. Zhu, “Relativistic tracold matter,” Nat. Phys. 15, 1161–1167 (2019). quantum effects of dirac particles simulated by ultracold [45] Alexander Mil, Torsten V. Zache, Apoorva Hegde, Andy atoms,” Frontiers of Physics 7, 31–53 (2011). Xia, Rohit P. Bhatt, Markus K. Oberthaler, Philipp [61] Stephen P. Jordan, Keith S. M. Lee, and John Preskill, Hauke, Jürgen Berges, and Fred Jendrzejewski, “A scal- “Quantum algorithms for quantum field theories,” Sci- able realization of local u(1) gauge invariance in cold ence 336, 1130–1133 (2012). atomic mixtures,” Science 367, 1128 (2020). [62] Ercolessi E. Facchi P. Marmo G. Pascazio S. Notarni- [46] R. Blatt and C. F. Roos, “Quantum simulations with cola S and F. V. Pepe, “Discrete abelian gauge theo- trapped ions,” Nat. Phys. 8, 277 (2012). ries for quantum simulations of qed,” J. Phys. A: Math. [47] Marcos D. Dalmonte M. Hauke P and P. Zoller, “Quan- Theor. 48 (2015). tum simulation of a lattice schwinger model in a chain [63] Henry Lamm and Scott Lawrence, “Simulation of of trapped ions,” Phys. Rev. 3 (2013). nonequilibrium dynamics on a quantum computer,” [48] Esteban A. Martinez, Christine A. Muschik, Philipp Phys. Rev. Lett. 121, 170501 (2018). Schindler, Daniel Nigg, Alexander Erhard, Markus [64] N. Klco, E. F. Dumitrescu, A. J. McCaskey, T. D. Mor- Heyl, Philipp Hauke, Marcello Dalmonte, Thomas ris, R. C. Pooser, M. Sanz, E. Solano, P. Lougovski, Monz, Peter Zoller, and Rainer Blatt, “Real-time dy- and M. J. Savage, “Quantum-classical computation of namics of lattice gauge theories with a few-qubit quan- schwinger model dynamics using quantum computers,” tum computer,” Nature 534, 516–519 (2016). Phys. Rev. A 98, 032331 (2018). [49] Logan W. Clark, Brandon M. Anderson, Lei Feng, [65] Q. S. Collaboration Nu, Henry Lamm, Scott Lawrence, Anita Gaj, K. Levin, and Cheng Chin, “Observa- and Yukari Yamauchi, “Parton physics on a quantum tion of density-dependent gauge fields in a bose-einstein computer,” Physical Review Research 2, 013272 (2020). condensate based on micromotion control in a shaken [66] Natalie Klco and Martin J. Savage, “Digitization of two-dimensional lattice,” Phys. Rev. Lett. 121, 030402 scalar fields for quantum computing,” Phys. Rev. A 99, (2018). 052335 (2019). [50] Zohreh Davoudi, Mohammad Hafezi, Christopher Mon- [67] Natalie Klco, Martin J. Savage, and Jesse R. Stryker, roe, Guido Pagano, Alireza Seif, and Andrew Shaw, “Su(2) non-abelian gauge field theory in one dimen- “Towards analog quantum simulations of lattice gauge sion on digital quantum computers,” Phys. Rev. D 101, theories with trapped ions,” Physical Review Research 074512 (2020). 2, 023015 (2020). [68] Kübra Yeter-Aydeniz, Eugene F. Dumitrescu, Alex J. [51] Rabl P. Rico E. Marcos D and P. Zoller, “Superconduct- McCaskey, Ryan S. Bennink, Raphael C. Pooser, and ing circuits for quantum simulation of dynamical gauge George Siopsis, “Scalar quantum field theories as a fields,” Phys. Rev. Lett. 111 (2013). benchmark for near-term quantum computers,” Phys. [52] D. Marcos, P. Widmer, E. Rico, M. Hafezi, P. Rabl, Rev. A 99, 032306 (2019). U. J. Wiese, and P. Zoller, “Two-dimensional lattice [69] Erez Zohar, Alessandro Farace, Benni Reznik, and gauge theories with superconducting quantum circuits,” J. Ignacio Cirac, “Digital quantum simulation of z2 lat- 13

tice gauge theories with dynamical fermionic matter,” gopalan, “Deeply inelastic scattering structure functions Phys. Rev. Lett. 118, 070501 (2017). on a hybrid quantum computer,” Phys. Rev. D 102, [70] Erez Zohar and J. Ignacio Cirac, “Eliminating fermionic 016007 (2020). matter fields in lattice gauge theories,” Phys. Rev. B 98, [86] Khadeejah Bepari, Sarah Malik, Michael Spannowsky, 075119 (2018). and Simon Williams, “Towards a Quantum Comput- [71] Christian W. Bauer, Wibe A. de Jong, Benjamin Nach- ing Algorithm for Helicity Amplitudes and Parton man, and Davide Provasoli, “A quantum algorithm Showers,” arXiv:2010.00046 [hep-ex, physics:hep-ph, for high energy physics simulations,” arXiv:1904.03196 physics:quant-ph] (2020), arXiv: 2010.00046. [hep-ph, physics:quant-ph] (2019), arXiv: 1904.03196. [87] Michael Kreshchuk, William M. Kirby, Gary Goldstein, [72] Indrakshi Raychowdhury and Jesse R. Stryker, “Loop, Hugo Beauchemin, and Peter J. Love, “Quantum Simu- string, and hadron dynamics in su(2) hamiltonian lat- lation of in the Light-Front For- tice gauge theories,” Phys. Rev. D 101, 114502 (2020). mulation,” arXiv:2002.04016 [hep-th, physics:quant-ph] [73] Weijie Du, James P. Vary, Xingbo Zhao, and Wei Zuo, (2020), arXiv: 2002.04016. “Quantum Simulation of Nuclear Inelastic Scattering,” [88] Michael A. Nielsen and Isaac L. Chuang, Quantum arXiv:2006.01369 [nucl-th, physics:quant-ph] (2020), Computation and : 10th Anniver- arXiv: 2006.01369. sary Edition (Cambridge University Press, 2010). [74] Alessandro Roggero, Andy C.Y. Li, Joseph Carlson, Ra- [89] Christine Muschik, Markus Heyl, Esteban Martinez, jan Gupta, and Gabriel N. Perdue, “Quantum comput- Thomas Monz, Philipp Schindler, Berit Vogell, Mar- ing for neutrino-nucleus scattering,” Phys. Rev. D 101, cello Dalmonte, Philipp Hauke, Rainer Blatt, and Pe- 074038 (2020). ter Zoller, “U(1) wilson lattice gauge theories in digi- [75] Michael Kreshchuk, Shaoyang Jia, William M. Kirby, tal quantum simulators,” New Journal of Physics 19, Gary Goldstein, James P. Vary, and Peter J. 103020 (2017). Love, “Light-Front Field Theory on Current Quantum [90] Long Gui-Lu, “General quantum interference principle Computers,” arXiv:2009.07885 [hep-lat, physics:hep-th, and duality computer,” Commun. Theor. Phys. 45, 825– physics:quant-ph] (2020), arXiv: 2009.07885. 844 (2006). [76] Wibe A. de Jong, Mekena Metcalf, James Mulligan, Ma- [91] Andrew M. Childs, Robin Kothari, and Rolando D. teusz Płoskoń, Felix Ringer, and Xiaojun Yao, “Quan- Somma, “Quantum algorithm for systems of linear equa- tum simulation of open quantum systems in heavy- tions with exponentially improved dependence on pre- ion collisions,” arXiv:2010.03571 [hep-ex, physics:hep- cision,” Siam. J. Comput. 46, 1920–1950 (2017). ph, physics:nucl-ex, physics:nucl-th, physics:quant-ph] [92] J. Van Apeldoorn, A. Gilyen, S. Gribling, and (2020), arXiv: 2010.03571. R. de Wolf, “Quantum sdp-solvers: better upper and [77] Zohreh Davoudi, Indrakshi Raychowdhury, and An- lower bounds,” 2017 IEEE 58th Annual Symposium on drew Shaw, “Search for Efficient Formulations for Foundations of Computer Science (FOCS). Proceedings Hamiltonian Simulation of non-Abelian Lattice Gauge , 403–414 (2017). Theories,” (2020), arXiv:2009.11802 [hep-lat]. [93] Anirban Narayan Chowdhury and Rolando D. Somma, [78] Annie Y. Wei, Preksha Naik, Aram W. Harrow, and “Quantum algorithms for gibbs sampling and hitting- Jesse Thaler, “Quantum algorithms for jet clustering,” time estimation,” Quantum Information & Computa- Phys. Rev. D 101, 094015 (2020). tion 17, 41–64 (2017). [79] Natalie Klco and Martin J. Savage, “Fixed-Point [94] András Gilyén, Yuan Su, Guang Hao Low, and Nathan Quantum Circuits for Quantum Field Theo- Wiebe, “Quantum singular value transformation and be- ries,” arXiv:2002.02018 [hep-th, physics:nucl-th, yond: Exponential improvements for quantum matrix physics:quant-ph] (2020), arXiv: 2002.02018. arithmetics,” in Proceedings of the 51st Annual ACM [80] Eric T. Holland, Kyle A. Wendt, Konstantinos Krav- SIGACT Symposium on Theory of Computing, STOC varis, Xian Wu, W. Erich Ormand, Jonathan L DuBois, 2019 (Association for Computing Machinery, New York, Sofia Quaglioni, and Francesco Pederiva, “Optimal con- NY, USA, 2019) p. 193–204. trol for the quantum simulation of nuclear dynamics,” [95] Juan Miguel Arrazola, Timjan Kalajdzievski, Christian Phys. Rev. A 101, 062307 (2020). Weedbrook, and Seth Lloyd, “Quantum algorithm for [81] A. Avkhadiev, P.E. Shanahan, and R.D. Young, “Accel- nonhomogeneous linear partial differential equations,” erating Lattice Quantum Field Theory Calculations via Phys. Rev. A 100, 032306 (2019). Interpolator Optimization Using Noisy Intermediate- [96] H. K. Lau, R. Pooser, G. Siopsis, and C. Weedbrook, Scale Quantum Computing,” Phys. Rev. Lett. 124, “Quantum machine learning over infinite dimensions,” 080501 (2020). Phys. Rev. Lett. 118, 080501 (2017). [82] Dmitri E. Kharzeev and Yuta Kikuchi, “Real-time chi- [97] Dan-Bo Zhang, Zheng-Yuan Xue, Shi-Liang Zhu, and ral dynamics from a digital quantum simulation,” Phys. Z. D. Wang, “Realizing quantum linear regression with Rev. Research 2, 023342 (2020). auxiliary qumodes,” Phys. Rev. A 99, 012331 (2019). [83] Alexander F. Shaw, Pavel Lougovski, Jesse R. Stryker, [98] Dan-Bo Zhang, Shi-Liang Zhu, and Z. D. Wang, “Pro- and Nathan Wiebe, “Quantum Algorithms for Simu- tocol for implementing quantum nonparametric learn- lating the Lattice Schwinger Model,” Quantum 4, 306 ing with trapped ions,” Phys. Rev. Lett. 124, 010506 (2020), arXiv: 2002.11146. (2020). [84] Junyu Liu and Yuan Xin, “Quantum simulation [99] Dan-Bo Zhang, Guo-Qing Zhang, Zheng-Yuan Xue, of quantum field theories as quantum chemistry,” Shi-Liang Zhu, and Z. D. Wang, “Continuous- arXiv:2004.13234 [hep-th, physics:quant-ph] (2020), variable assisted thermal quantum simulation,” (2020), arXiv: 2004.13234. arXiv:2006.00471 [quant-ph]. [85] Niklas Mueller, Andrey Tarasov, and Raju Venu- [100] Guang Hao Low and Isaac L. Chuang, “Optimal hamil- 14

tonian simulation by quantum signal processing,” Phys. mon C. Benjamin, and Xiao Yuan, “Variational ansatz- Rev. Lett. 118, 010501 (2017). based quantum simulation of imaginary time evolution,” [101] Earl Campbell, “Random compiler for fast hamiltonian npj Quantum Information 5, 75 (2019). simulation,” Phys. Rev. Lett. 123, 070503 (2019). [118] Suguru Endo, Jinzhao Sun, Ying Li, Simon C. Ben- [102] Andrew M. Childs, Yuan Su, Minh C. Tran, Nathan jamin, and Xiao Yuan, “Variational Quantum Simu- Wiebe, and Shuchen Zhu, “A theory of trotter error,” lation of General Processes,” Physical Review Letters (2019), arXiv:1912.08854 [quant-ph]. 125, 010501 (2020), publisher: American Physical So- [103] Minh C. Tran, Su-Kuan Chu, Yuan Su, Andrew M. ciety. Childs, and Alexey V. Gorshkov, “Destructive error in- [119] Leonard Susskind, “Lattice fermions,” Phys. Rev. D 16, terference in product-formula lattice simulation,” Phys. 3031–3039 (1977). Rev. Lett. 124, 220502 (2020). [120] E. F Dumitrescu, A. J McCaskey, G. Hagen, G. R [104] Tameem Albash and Daniel A. Lidar, “Adiabatic quan- Jansen, T. D Morris, T. Papenbrock, R. C Pooser, D. J tum computation,” Rev. Mod. Phys. 90, 015002 (2018). Dean, and P. Lougovski, “Cloud quantum computing [105] Yangchao Shen, Xiang Zhang, Shuaining Zhang, Jing- of an atomic nucleus,” Phys. Rev. Lett. 120, 210501 Ning Zhang, Man-Hong Yung, and Kihwan Kim, (2018). “Quantum implementation of the unitary coupled clus- [121] Xiangdong Ji, “Parton Physics on a Euclidean Lattice,” ter for simulating molecular electronic structure,” Phys. Phys. Rev. Lett. 110, 262002 (2013). Rev. A 95, 020501 (2017). [122] Zhong-Bo Kang, Enke Wang, Xin-Nian Wang, and [106] Cornelius Hempel, Christine Maier, Jonathan Romero, Hongxi Xing, “Next-to-Leading Order QCD Factor- Jarrod McClean, Thomas Monz, Heng Shen, Petar Ju- ization for Semi-Inclusive Deep Inelastic Scattering rcevic, Ben P. Lanyon, Peter Love, Ryan Babbush, at Twist 4,” Phys. Rev. Lett. 112, 102001 (2014), Alán Aspuru-Guzik, Rainer Blatt, and Christian F. arXiv:1310.6759 [hep-ph]. Roos, “Quantum chemistry calculations on a trapped- [123] Zhong-Bo Kang, Ivan Vitev, and Hongxi Xing, “Next- ion quantum simulator,” Physical Review X 8, 031022 to-leading order forward hadron production in the (2018). small-x regime: rapidity factorization,” Phys. Rev. Lett. [107] Dan-Bo Zhang and Tao Yin, “Collective optimization 113, 062002 (2014), arXiv:1403.5221 [hep-ph]. for variational quantum eigensolvers,” Phys. Rev. A [124] Huey-Wen Lin, Emanuele R. Nocera, Fred Olness, 101, 032311 (2020). Kostas Orginos, Juan Rojo, Alberto Accardi, Constan- [108] Jarrod R. McClean, Mollie E. Kimchi-Schwartz, tia Alexandrou, Alessandro Bacchetta, Giuseppe Bozzi, Jonathan Carter, and Wibe A. de Jong, “Hybrid Jiunn-Wei Chen, Sara Collins, Amanda Cooper-Sarkar, quantum-classical hierarchy for mitigation of decoher- Martha Constantinou, Luigi Del Debbio, Michael Engel- ence and determination of excited states,” Phys. Rev. A hardt, Jeremy Green, Rajan Gupta, Lucian A. Harland- 95, 042308 (2017). Lang, Tomomi Ishikawa, Aleksander Kusina, Keh-Fei [109] Ken M. Nakanishi, Kosuke Mitarai, and Keisuke Fu- Liu, Simonetta Liuti, Christopher Monahan, Pavel jii, “Subspace-search variational quantum eigensolver Nadolsky, Jian-Wei Qiu, Ingo Schienbein, Gerrit Schier- for excited states,” Physical Review Research 1, 033062 holz, Robert S. Thorne, Werner Vogelsang, Hartmut (2019). Wittig, C.-P. Yuan, and James Zanotti, “Parton dis- [110] Oscar Higgott, Daochen Wang, and Stephen Brierley, tributions and lattice QCD calculations: A community “Variational Quantum Computation of Excited States,” white paper,” Progress in Particle and Nuclear Physics Quantum 3, 156 (2019). 100, 107–160 (2018). [111] Dan-Bo Zhang, Zhan-Hao Yuan, and Tao Yin, “Vari- [125] Jarrod R McClean, Nicholas C Rubin, Kevin J Sung, ational quantum eigensolvers by variance minimiza- Ian D Kivlichan, Xavier Bonet-Monroig, Yudong Cao, tion,” arXiv:2006.15781 [quant-ph] (2020), arXiv: Chengyu Dai, E Schuyler Fried, Craig Gidney, Bren- 2006.15781. dan Gimby, Pranav Gokhale, Thomas Häner, Tarini [112] Jin-Guo Liu, Liang Mao, Pan Zhang, and Lei Wang, Hardikar, Vojtěch Havlíček, Oscar Higgott, Cupjin “Solving quantum statistical mechanics with variational Huang, Josh Izaac, Zhang Jiang, Xinle Liu, Sam autoregressive networks and quantum circuits,” (2019), McArdle, Matthew Neeley, Thomas O’Brien, Bryan arXiv:1912.11381 [quant-ph]. O’Gorman, Isil Ozfidan, Maxwell D Radin, Jhonathan [113] Jingxiang Wu and Timothy H. Hsieh, “Variational ther- Romero, Nicolas P D Sawaya, Bruno Senjean, Kanav Se- mal quantum simulation via thermofield double states,” tia, Sukin Sim, Damian S Steiger, Mark Steudtner, Qim- Phys. Rev. Lett. 123, 220502 (2019). ing Sun, Wei Sun, Daochen Wang, Fang Zhang, and [114] D. Zhu, S. Johri, N. M. Linke, K. A. Landsman, Ryan Babbush, “OpenFermion: the electronic structure N. H. Nguyen, C. H. Alderete, A. Y. Matsuura, T. H. package for quantum computers,” Quantum Science and Hsieh, and C. Monroe, “Generation of thermofield dou- Technology 5, 034014 (2020). ble states and critical ground states with a quantum [126] Ville Bergholm, Josh Izaac, Maria Schuld, Chris- computer,” (2019), arXiv:1906.02699 [quant-ph]. tian Gogolin, M. Sohaib Alam, Shahnawaz Ahmed, [115] Anirban N. Chowdhury, Guang Hao Low, and Nathan Juan Miguel Arrazola, Carsten Blank, Alain Delgado, Wiebe, “A variational quantum algorithm for prepar- Soran Jahangiri, Keri McKiernan, Johannes Jakob ing quantum gibbs states,” (2020), arXiv:2002.00055 Meyer, Zeyue Niu, Antal Száva, and Nathan Killoran, [quant-ph]. “PennyLane: Automatic differentiation of hybrid [116] Youle Wang, Guangxi Li, and Xin Wang, “Variational quantum-classical computations,” arXiv:1811.04968 quantum gibbs state preparation with a truncated tay- [physics, physics:quant-ph] (2020), arXiv: 1811.04968. lor series,” (2020), arXiv:2005.08797 [quant-ph]. [117] Sam McArdle, Tyson Jones, Suguru Endo, Ying Li, Si-