Understanding quantum phases of matter is a fundamental goal in physics. For pure states, the representatives of phases are the ground states of locally interacting Hamiltonians, which are also renormalization fixed points (RFPs). These RFP states are exactly described by tensor networks. Extending this framework to mixed states, matrix product density operators (MPDOs) that are RFPs are believed to encapsulate mixed-state phases of matter in one dimension, where nontrivial topological phases have already been shown to exist. However, to better motivate the physical relevance of those states, and in particular the physical relevance of the recently found non trivial phases, it remains an open question whether such MPDO RFPs can be realized as steady states of local Lindbladians. In this work, we resolve this question by analytically constructing parent Lindbladians for MPDO RFPs. These Lindbladians are local, frustration free, and exhibit minimal steady-state degeneracy. Interestingly, we find that parent Lindbladians possess a rich structure that distinguishes them from their Hamiltonian counterparts. In particular, we uncover an intriguing connection between the noncommutativity of the Lindbladian terms and the fact that the corresponding MPDO RFP belongs to a nontrivial phase.

Matrix Product Operators (MPOs) are tensor networks representing operators acting on 1D systems. They model a wide variety of situations, including communication channels with memory effects, quantum cellular automata, mixed states in 1D quantum systems, or holographic boundary models associated to 2D quantum systems. A scenario where MPOs have proven particularly useful is to represent algebras of non-trivial symmetries. Concretely, the boundary of both symmetry protected and topologically ordered phases in 2D quantum systems exhibit symmetries in the form of MPOs. In this paper, we develop a theory of MPOs as representations of algebraic structures. We establish a dictionary between algebra and MPO properties which allows to transfer results between both setups, covering the cases of pre-bialgebras, weak bialgebras, and weak Hopf algebras. We define the notion of pulling-through algebras, which abstracts the minimal requirements needed to define topologically ordered 2D tensor networks from MPO algebras. We show, as one of our main results, that any semisimple pivotal weak Hopf algebra is a pulling-trough algebra. We demonstrate the power of this framework by showing that they can be used to construct Kitaev’s quantum double models for Hopf algebras solely from an MPO representation of the Hopf algebra, in the exact same way as MPO symmetries obtained from fusion categories can be used to construct Levin-Wen string-net models, and to explain all their topological features; it thus allows to describe both Kitaev and string-net models on the same formal footing.

We investigate the performance of an adiabatic evolution protocol when initialized from a Gibbs state at finite temperature. Specifically, we identify the diagonality of the final state in the energy eigenbasis, as well as the difference in energy and in energy variance with respect to the ideal adiabatic limit as key benchmarks for success and introduce metrics to quantify the off-diagonal contributions. Provided these benchmarks converge to their ideal adiabatic values, we argue that thermal expectation values of observables can be recovered, in accordance with the eigenstate thermalization hypothesis. For the transverse-field Ising model, we analytically establish that these benchmarks converge polynomially in both the quasi-adiabatic evolution time T and system size. We perform numerical studies on non-integrable systems and find close quantitative agreement for the off-diagonality metrics, along with qualitatively similar behavior in the energy convergence.

We propose applying the adiabatic algorithm to prepare high-energy eigenstates of integrable models on a quantum computer. We first review the standard adiabatic algorithm to prepare ground states in each magnetization sector of the prototypical XXZ Heisenberg chain. Based on the thermodynamic Bethe ansatz, we show that the algorithm circuit depth is polynomial in the number of qubits N, outperforming previous methods explicitly relying on integrability. Next, we propose a protocol to prepare arbitrary eigenstates of integrable models that satisfy certain conditions. For a given target eigenstate, we construct a suitable parent Hamiltonian written in terms of a complete set of local conserved quantities. We propose using such Hamiltonian as an input for an adiabatic algorithm. After benchmarking this construction in the case of the non-interacting XY spin chain, where we can rigorously prove its efficiency, we apply it to prepare arbitrary eigenstates of the Richardson-Gaudin models. In this case, we provide numerical evidence that the circuit depth of our algorithm is polynomial in N for all eigenstates, despite the models being interacting.

We present and test a protocol to learn the matrix-product operator (MPO) representation of an experimentally prepared quantum state. The protocol takes as input classical shadows corresponding to local randomized measurements, and outputs the tensors of an MPO maximizing a suitably defined fidelity with the experimental state. The tensor optimization is carried out sequentially, similarly to the well-known density matrix renormalization group algorithm. Our approach is provably efficient under certain technical conditions expected to be met in short-range correlated states and in typical noisy experimental settings. Under the same conditions, we also provide an efficient scheme to estimate fidelities between the learned and the experimental states. We experimentally demonstrate our protocol by learning entangled quantum states of up to N=96 qubits in a superconducting quantum processor. Our method upgrades classical shadows to large-scale quantum computation and simulation experiments.

We introduce regular language states, a family of quantum many-body states. They are built from a special class of formal languages, called regular, which has been thoroughly studied in the field of computer science. They can be understood as the superposition of all the words in a regular language and encompass physically relevant states such as the GHZ-, W- or Dicke-states. By leveraging the theory of regular languages, we develop a theoretical framework to describe them. First, we express them in terms of matrix product states, providing efficient criteria to recognize them. We then develop a canonical form which allows us to formulate a fundamental theorem for the equivalence of regular language states, including under local unitary operations. We also exploit the theory of tensor networks to find an efficient criterion to determine when regular languages are shift-invariant.

Tensor network methods provide powerful analytical and numerical tools for characterizing quantum phases of matter. While the mathematical structure of matrix product states (MPS) is well understood through the MPS fundamental theorem, an analogous understanding for mixed-state tensor networks remains largely absent: if two purification tensors generate the same density matrix, how are they related? In this work, we initiate the study of a fundamental theorem for matrix product locally purifiable density operators (LPDOs) and focus on sequentially generated LPDOs (sLPDOs), a broad subclass admitting an interpretation in terms of successive applications of quantum channels on an initial state. We prove that, under suitable invertibility or cyclic conditions, two sLPDO representations generate the same density matrix for arbitrary system sizes if and only if they are related by a matrix product isometry acting on the purification bonds. Beyond the sLPDO setting, we provide a counterexample that suggests an obstruction to a general fundamental theorem for LPDOs with periodic boundary conditions. Finally, we discuss implications for mixed-state symmetry-protected topological phases, including the possibility of nontrivial phases protected only by weak symmetry conditions.

We propose a quantum algorithm to simulate the dynamics in quantum chemistry problems. It is based on adding fresh qubits at each Trotter step, which enables a simpler implementation of the dynamics in the extended system. After each step, the extra qubits are recycled, so that the whole process accurately approximates the correct unitary evolution. A key ingredient of the approach is an isometry that maps a simple diagonal Hamiltonian in the extended system to the original one, and we give a procedure to compute this isometry. We estimate the error at each time step, as well as the number of gates, which scales as O(N^{2}), where N is the number of orbitals. We illustrate our results with three examples: the hydrogen chain, small molecules, and the FeMoco (Fe_{7}MoS_{9}C) molecule. In the hydrogen chain and the hydrogen molecule, we observe that the error scales in the same way as the Trotter error. For FeMoco, we estimate the number of gates in a fault-tolerant setup.

We introduce a variational Monte Carlo algorithm for approximating finite-temperature quantum many-body systems, based on the minimization of a modified free energy. This approach directly approximates the state at a fixed temperature, allowing for systematic improvement of the expressiveness without accumulating errors from iterative imaginary-time evolution. We employ a variety of trial states—both tensor networks as well as neural networks—as variational for our numerical optimization. We benchmark and compare different constructions in the above classes, both for one- and two-dimensional problems, with systems made of up to N=100 spins. Our results demonstrate that while restricted Boltzmann machines show limitations, string bond tensor network states exhibit systematic improvements with increasing bond dimensions and the number of strings. Published by the American Physical Society 2025

Cavity quantum electrodynamics (cavity QED) enables the control of light-matter interactions at the single-photon level, rendering it a key component of many quantum technologies. Its practical realization, however, is complex since it involves placing individual quantum emitters close to mirror surfaces within a high-finesse cavity. In this work, we propose a cavity QED architecture fully based on atoms trapped in free space. In particular, we show that a pair of two-dimensional, ordered arrays of atoms can be described by conventional cavity QED parameters. Such an atom-array cavity exhibits the same cooperativity as a conventional counterpart with matching mirror specifications even though the cavity coupling strength and decay rate are modified by the narrow bandwidth of the atoms. We estimate that an array cavity composed of ^87Rb atoms in an optical lattice can reach a cooperativity of about 10. This value can be increased suppressing atomic motion with larger trap depths and may exceed 10^4 with an ideal placement of the atoms. To reduce the experimental complexity of our scheme, we propose a spatially dependent AC Stark shift as an alternative to curving the arrays, which may be of independent interest. In addition to presenting a promising platform for cavity QED, our work creates opportunities for exploring novel phenomena based on the intrinsic nonlinearity of atom arrays and the possibility to dynamically control them.

Simulating quantum dynamics on digital or analog quantum simulators often requires “problem-to-simulator" mappings such as trotterization, floquet-magnus expansion or perturbative expansions. When the simulator is noiseless, it is well understood that these problem-to-simulator mappings can be made as accurate as desired at the expense of simulator run-time. However, precisely because the simulator has to be run for a longer time to increase its accuracy, it is expected that noise in the quantum simulator catastrophically effects the simulator output. We show that, contrary to this expectation, these mappings remain stable to noise when considering the task of simulating dynamics of local observables in quantum lattice models. Specifically, we prove that in all of these mappings, local observables can be determined to a system-size independent, precision that scales sublinearly with the noise-rate in the simulator. Our results provide theoretical evidence that quantum simulators can be used for solving problems in many-body physics without or with modest error correction.

We present a numerical method to simulate a system of multiple emitters coupled to a noninteracting bath in any parameter regime. Our method relies on a Block Lanczos transformation that maps the whole system onto a strip geometry whose width is given by the number of emitters. Utilizing the spatial symmetries of the problem and identifying the relevant range of energies of the bath, we achieve a more efficient simulation, which we perform using tensor network techniques. As a demonstration, we study the collective emission from multiple emitters coupled to a square lattice of bosons and observe how the departure from Markovianity as coupling strength and emitter number is increased prevents collective radiation. We also simulate the dynamic preparation of an excitation in a bound state from a multiexcitation initial state. Our work opens possibilities for the systematic exploration of non-Markovian effects in the dynamics and equilibrium properties of multiemitter systems. Furthermore, it can easily be extended to other setups, including finite bath temperature or impurities coupled to fermionic environments.

Large Language Models (LLMs) are rapidly advancing across diverse domains, yet their application in theoretical physics remains inadequate. While current models show competence in mathematical reasoning and code generation, we identify critical gaps in physical intuition, constraint satisfaction, and reliable reasoning that cannot be addressed through prompting alone. Physics demands approximation judgment, symmetry exploitation, and physical grounding that require AI agents specifically trained on physics reasoning patterns and equipped with physics-aware verification tools. We argue that LLM would require such domain-specialized training and tooling to be useful in real-world for physics research. We envision physics-specialized AI agents that seamlessly handle multimodal data, propose physically consistent hypotheses, and autonomously verify theoretical results. Realizing this vision requires developing physics-specific training datasets, reward signals that capture physical reasoning quality, and verification frameworks encoding fundamental principles. We call for collaborative efforts between physics and AI communities to build the specialized infrastructure necessary for AI-driven scientific discovery.

We investigate theoretically equilibrium and dynamical properties of a Kondo impurity coupled to either 1D or 2D superconductors, modeled by the attractive Fermi-Hubbard model. By employing a non-Gaussian variational approach, we go beyond the approximation of a constant superconducting (SC) gap. We show that dynamical properties of the system can be modified qualitatively, when space and time dependent renormalization of the SC gap and electron-impurity hybridization are included. For the ground state, we find the singlet-doublet phase transition and π-phase shifts of the SC order parameter. For dynamics, first we consider spin dynamics following an abrupt connection of the polarized impurity to the 2D bath. We find rapid relaxation of impurity polarization and directional emission of a magnetization pulse, which becomes damped as it propagates into the bulk. Then we analyze transport between two SC leads coupled through the impurity at finite bias voltage. Here we go beyond analysis of the steady state to investigate full-time dynamics following an abrupt application of the bias voltage. We uncover four distinct regimes in the transient dynamics and transport properties: (I) the AC Josephson effect regime; (II) dynamical competition between charge-density-wave (CDW) and SC orders with transient Kondo correlations; (III) the coexistence of AC and DC currents facilitated by partial Kondo screening and dynamical stabilization of the SC order; (IV) DC Kondo transport regime modified by the SC order. Regime II exhibits a dynamical transition from SC to CDW order that locally restores the U(1) symmetry. We argue that our findings for regime IV provide a theoretical explanation for the experimentally observed anomalous enhancement of DC conductance and suppression of the AC Josephson current. Finally, we discuss the potential experimental realization with ultracold atoms.

We investigate the potential of tensor network based machine learning methods to scale to large image and text data sets. For that, we study how the mutual information between a subregion and its complement scales with the subsystem size L, similarly to how it is done in quantum many-body physics. We find that for text, the mutual information scales as a power law L^ν with a close to volume law exponent, indicating that text cannot be efficiently described by 1D tensor networks. For images, the scaling is close to an area law, hinting at 2D tensor networks such as PEPS could have an adequate expressibility. For the numerical analysis, we introduce a mutual information estimator based on autoregressive networks, and we also use convolutional neural networks in a neural estimator method.

Matrix-product unitaries (MPUs) are many-body unitary operators that, as a consequence of their tensor-network structure, preserve the entanglement area law in 1D systems. However, it is unknown how to implement an MPU as a quantum circuit since the individual tensors describing the MPU are not unitary. In this Letter, we show that a large class of MPUs can be implemented with a polynomial-depth quantum circuit. For an N -site MPU built from a repeated bulk tensor with open boundary, we explicitly construct a quantum circuit of polynomial depth T = O ( N α ) realizing the MPU, where the constant α depends only on the bulk and boundary tensor and not the system size N . We show that this class includes nontrivial unitaries that generate long-range entanglement and, in particular, contains a large class of unitaries constructed from representations of C * -weak Hopf algebras. Furthermore, we also adapt our construction to nonuniform translationally varying MPUs and show that they can be implemented by a circuit of depth O ( N β poly D ) where β ≤ 1 + log 2 D / s min , with D being the bond dimension and s min the smallest nonzero Schmidt value of the normalized Choi state corresponding to the MPU.

We characterize the dynamical state of many-body bosonic and fermionic many-body models with inter-site Gaussian couplings, on-site non-Gaussian interactions and local dissipation comprising incoherent particle loss, particle gain, and dephasing. We first establish that, for fermionic systems, if the dephasing noise is larger than the non-Gaussian interactions, irrespective of the Gaussian coupling strength, the system state is a convex combination of Gaussian states at all times. Furthermore, for bosonic systems, we show that if the particle loss and particle gain rates are larger than the Gaussian inter-site couplings, the system remains in a separable state at all times. Building on this characterization, we establish that at noise rates above a threshold, there exists a classical algorithm that can efficiently sample from the system state of both the fermionic and bosonic models. Finally, we show that, unlike fermionic systems, bosonic systems can evolve into states that are not convex-Gaussian even when the dissipation is much higher than the on-site non-Gaussianity. Similarly, unlike bosonic systems, fermionic systems can generate entanglement even with noise rates much larger than the inter-site couplings.

For random quantum circuits on n qubits of depth Θ(log⁡n) with depolarizing noise, the task of sampling from the output state can be efficiently performed classically using a Pauli path method \cite{Aharonov2023_paulipaths} . This paper aims to study the performance of this method beyond random circuits. We first consider the classical simulation of local observables in circuits composed of Clifford and T gates – going beyond the average case analysis, we derive sufficient conditions for simulatability in terms of the noise rate and the fraction of gates that are T gates, and show that if noise is introduced at a faster rate than T gates, the simulation becomes classically easy. As an application of this result, we study 2D QAOA circuits that attempt to find low-energy states of classical Ising models on general graphs. There, our results shows that for hard instances of the problem, which correspond to Ising model's graph being geometrically non-local, a QAOA algorithm mapped to a geometrically local circuit architecture using SWAP gates does not have any asymptotic advantage over classical algorithms if depolarized at a constant rate. Finally, we illustrate instances where the Pauli path method fails to give the correct result, and also initiate a study of the trade-off between fragility to noise and classical complexity of simulating a given quantum circuit.

Matrix-product unitaries (MPU) are 1D tensor networks describing time evolution and unitary symmetries of quantum systems, while their action on states by construction preserves the entanglement area law. MPU which are formed by a single repeated tensor are known to coincide with 1D quantum cellular automata (QCA), i.e., unitaries with an exact light cone. However, this correspondence breaks down for MPU with open boundary conditions, even if the resulting operator is translation-invariant. Such unitaries can turn short- to long-range correlations and thus alter the underlying phase of matter. Here we make the first steps towards a theory of MPU with uniform bulk but arbitrary boundary. In particular, we study the structure of a subclass with a direct-sum form which maximally violates the QCA property. We also consider the general case of MPU formed by site-dependent (nonuniform) tensors and show a correspondence between MPU and locally maximally entanglable states.

The computation of dynamical response functions is central to many problems in condensed matter physics. Owing to the rapid growth of quantum correlations following a quench, classical methods face significant challenges even if an efficient description of the equilibrium state is available. Quantum computing offers a promising alternative. However, existing approaches often assume access to the equilibrium state, which may be difficult to prepare in practice. In this work, we present a method that circumvents this by using energy filter techniques, enabling the computation of response functions and other dynamical properties in both microcanonical and canonical ensembles. Our approach only requires the preparation of states that have significant weight at the desired energy. The dynamical response functions are then reconstructed from measurements after quenches of varying duration by classical postprocessing. We illustrate the algorithm numerically by applying it to compute the dynamical conductivity of a free-fermion model, which unveils the energy-dependent localization properties of the model.