The pursuit of topological phenomena in non-Hermitian systems has unveiled new physics beyond the conventional Hermitian paradigm, yet their realization in interacting many-body platforms remains a critical challenge. Exploring this interplay is essential to understand how strong interactions and dissipation collectively shape topological phases in open quantum systems. Here, we experimentally demonstrate dynamical spectral topology in a dissipative Rydberg atomic gas and characterize parameter-dependent winding numbers, which quantify a geometric winding of the spectral topology. By increasing the interaction strength, the system evolves from Hermitian to non-Hermitian regime, accompanying emergence of trajectory loop in the complex energy plane. As the scanning time is varied, the spectral topology becomes twisted in the complex energy plane, forming enclosed sub-loops characterized by opposite winding numbers. Furthermore, by changing the scanning direction, we observe the differentiated spectral loops, revealing a signature of scan-direction-dependent nonreciprocity. This work establishes cold Rydberg gases as a versatile platform for exploring the rich interplay between non-Hermitian topology, strong interactions, and dissipative quantum dynamics.

Artificial intelligence in dynamic, real-world environments requires the capacity for continual learning. However, standard deep learning suffers from a fundamental issue: loss of plasticity, in which networks gradually lose their ability to learn from new data. Here we show that quantum learning models naturally overcome this limitation, preserving plasticity over long timescales. We demonstrate this advantage systematically across a broad spectrum of tasks from multiple learning paradigms, including supervised learning and reinforcement learning, and diverse data modalities, from classical high-dimensional images to quantum-native datasets. Although classical models exhibit performance degradation correlated with unbounded weight and gradient growth, quantum neural networks maintain consistent learning capabilities regardless of the data or task. We identify the origin of the advantage as the intrinsic physical constraints of quantum models. Unlike classical networks where unbounded weight growth leads to landscape ruggedness or saturation, the unitary constraints confine the optimization to a compact manifold. Our results suggest that the utility of quantum computing in machine learning extends beyond potential speedups, offering a robust pathway for building adaptive artificial intelligence and lifelong learners.

We propose a decoder for quantum low-density parity-check (LDPC) codes based on a beam search heuristic guided by belief propagation (BP). Our beam search decoder applies to all quantum LDPC codes and achieves different speed-accuracy tradeoffs by tuning its parameters such as the beam width. We perform numerical simulations under circuit level noise for the $[[144,12,12]]$ bivariate bicycle (BB) code at noise rate $p={10}^{-3}$ to estimate the logical error rate and the 99.9 percentile runtime and we compare with the BP-OSD decoder which has been the default quantum LDPC decoder for the past six years. A variant of our beam search decoder with a beam width of 64 achieves a $17\times$ reduction in logical error rate. With a beam width of 8, we reach the same logical error rate as BP-OSD with a $26.2\times$ reduction in the 99.9 percentile runtime. We identify the beam search decoder with beam width of 32 as a promising candidate for trapped ion architectures because it achieves a $5.6\times$ reduction in logical error rate with a 99.9 percentile runtime per syndrome extraction round below 1 ms at $p=5\times {10}^{-4}$. Remarkably, this is achieved in software on a single core, without any parallelization or specialized hardware (FPGA, ASIC), suggesting one might only need three 32-core CPUs to decode a trapped ion quantum computer with 1000 logical qubits.

We formulate a minimal model of a quantum particle detector as an autonomous quantum thermal machine. Our goal is to establish how entropy production, which is needed to maintain the detector out of equilibrium, is linked to the quality of the measurement process. Using our model, we perform a detailed investigation of the detector’s key performance characteristics: namely, detection efficiency, gain, jitter, dead time, and dark counts. We find that entropy production constrains both the efficiency and temporal precision of the detection process, in the sense that improved performance generally requires more dissipation. We also find that reducing either the detection jitter or dead time unavoidably increases the rate of dark counts. Our work establishes a quantitative connection between entropy production and the quality of the irreversible detection process, highlights fundamental tradeoffs in the performance of particle detectors, and provides a framework for further investigations of the non-equilibrium thermodynamics of quantum measurement and amplification.

Quantum field simulators provide unique opportunities for investigating the dynamics of quantum fields through tabletop experiments. A primary drawback of standard encoding schemes is their rigidity: altering the theory, its coupling geometry, metric structure, or simulation time typically requires redesigning the experimental setup, which imposes strong constraints on the types of dynamics and theories that can be simulated. Here, we introduce the Optical Time Algorithm (OTA) as a unifying framework, enabling the efficient simulation of large classes of free quantum field dynamics using a single optical circuit design that separates the time from the Hamiltonian’s structure. By modifying the parameters of the optical elements, our method allows us to engineer timescales, coupling graphs, spacetime metrics, and boundary conditions, thereby facilitating the implementation of relativistic and nonrelativistic, real- and complex-valued, short- and long-range quantum field theories on both flat and curved spacetimes. We exploit the OTA’s configurability to investigate the spreading of quantum correlations in space and time for theories with continuously varying coupling ranges. Relevant features predicted by quantum field theory can be observed in systems with 10 to 20 modes under realistic conditions, paving the way for experimental implementations.

Drive-induced unwanted state transitions (DUST) are limiting both for microwave readout and parametric operations of superconducting qubits. Among them, measurement-induced state transitions (MIST) are due to intrinsic resonances described by the readout Hamiltonian. They were previously studied with a qubit linearly coupled to its readout mode, which constitutes the usual readout Hamiltonian. Since MIST can appear even at moderate powers, they limit the readout signal-to-noise ratio and the quantum non-demolition readout fidelity. In this work, we study the high-power readout regime in a different transmon readout scheme, implementing a nonlinear coupling called the $\cos \phi$-coupling. This coupling stems from a transmon molecule circuit and has symmetry properties that suppress nonparity-conserving MIST. We succeed in performing multistate single-shot readout up to the fifth excited state of the transmon, which enables us to identify leakage pathways from the computational subspace. The measurements indicate that the system is free of MIST up to high powers, with more than 300 photons in the readout mode. The MIST can be controllably turned on by breaking the parity symmetry of the coupling using flux-tuning. These experimental results are corroborated by branch analysis and simulations of the classical chaotic dynamics, showing that the $\cos \phi$-coupling is very robust to readout photons compared to the usual transverse coupling.

Displacement sensing is a fundamental task in metrology. However, the development of quantum-enhanced sensors that fully utilize the available degrees of freedom in many-body quantum systems remains an outstanding challenge. We propose many-body displacement sensing schemes that use spin-dependent squeezed (SDS) states—hybrid spin-boson states whose bosonic squeezed quadrature is conditioned on an auxiliary spin. We prove that SDS states are optimal, i.e., their quantum Cramér-Rao bound saturates the Heisenberg limit. We propose explicit measurement sequences that can be readily implemented in systems such as trapped ions. We also introduce a scalable state-preparation protocol and numerically demonstrate the preparation of 8.7 dB of spin-dependent squeezing 15 times faster than the standard approach using second-order sidebands in trapped ions. The potential applications of our sensing protocols range from measuring single-photon scattering to searches for dark matter.

We introduce a family of scalable planar fault-tolerant circuits that implement logical non-Clifford operations on a 2D color code, such as a logical $T$ gate or a logical non-Pauli measurement that prepares a magic $|T⟩$ state. The circuits are relatively simple, consisting only of physical $T$ gates, $CX$ gates, and few-qubit measurements. They can be implemented with an array of qubits on a 2D chip with nearest-neighbor couplings and no wire crossings. The construction is based on a spacetime path integral representation of a non-Abelian 2+1D topological phase, which is related to the 3D color code. We turn the path integral into a circuit by expressing it as a spacetime $ZX$ tensor network and then traversing it in some chosen time direction. We describe in detail how fault tolerance is achieved using a “just-in-time” decoding strategy, for which we repurpose and extend state-of-the-art color-code matching decoders.

Non-Gaussianity is a key resource for achieving quantum advantages in bosonic platforms. Here, we investigate the symplectic rank: a non-Gaussianity monotone that satisfies remarkable operational and resource-theoretic properties. Mathematically, the symplectic rank of a pure state is the number of symplectic eigenvalues of the covariance matrix that are strictly larger than those of the vacuum. Operationally, it (i) is easy to compute, (ii) emerges as the smallest number of modes onto which all the non-Gaussianity can be compressed via Gaussian unitaries, (iii) lower bounds the non-Gaussian gate complexity of state preparation independently of the gate set, (iv) governs the sample complexity of quantum tomography, and (v) bounds the computational complexity of bosonic circuits. Crucially, the symplectic rank is nonincreasing under postselected Gaussian operations, leading to new no-go theorems for Gaussian conversion. Remarkably, this allows us to show that the resource theory of non-Gaussianity is irreversible under exact Gaussian operations. Finally, we show that the symplectic rank is a robust non-Gaussian measure, explaining how to witness it in experiments and how to exploit it to meaningfully benchmark different bosonic platforms. In doing so, we derive lower bounds on the trace distance (resp. total variation distance) between arbitrary states (resp. classical probability distributions) in terms of the norm distance between their covariance matrices, which may be of independent interest.

We present universal properties of anticoncentration in weakly noisy quantum circuits at finite depth. We develop a generic framework for single- and multi-qubit noise channels in the weak-noise limit and introduce an effective description in terms of a random matrix product operator. Within this weak-noise regime, we show that distinct noise mechanisms act in a quantitatively similar way, yielding a universal distribution of bit-string probabilities that is largely independent of the microscopic noise channel and of the circuit architecture. We identify three depth-dependent regimes, each characterized by a distinct scaling of cross-entropy benchmarking ($\mathrm{XEB}$) with rescaled depth. In the shallow-depth regime, noise effects are perturbatively small; in the intermediate regime, circuit-induced fluctuations and noise compete on equal footing; and in the deep-depth regime, the output distribution becomes effectively classical, up to corrections that are exponentially small in the noise strength. We provide quantitative predictions for anticoncentration in generic finite-depth circuits and benchmark them against numerical simulations, finding excellent agreement even at shallow depths. Moreover, we show that, contrary to previous expectations, the late-time value of $\mathrm{XEB}$ provides direct access to the global circuit fidelity, even at large noise strengths. Our results are directly applicable to current quantum processors and demonstrate universal behavior beyond the pure random-matrix-theory regime, which only emerges at asymptotically large depths.

We investigate the dividing line between classical and quantum computational power in estimating properties of matrix functions. More precisely, we study the computational complexity of two primitive problems: given a function $f$ and a Hermitian matrix $A$, we compute a matrix element of $f(A)$ or a local measurement on $f(A)|0{⟩}^{\otimes n}$, with $|0{⟩}^{\otimes n}$ being an $n$-qubit reference state vector, in both cases up to additive approximation error. We consider four functions—monomials, Chebyshev polynomials, the time evolution function, and the inverse function—and probe the complexity across a broad landscape covering different problem input regimes. Namely, we consider two types of matrix inputs (sparse and Pauli access), matrix properties (norm and sparsity), the approximation error, and function-specific parameters. We identify BQP-complete forms of both problems for each function and then toggle the problem parameters to easier regimes to see where hardness remains, or where the problem becomes classically easy. As part of our results, we make concrete a hierarchy of hardness across the functions; in parameter regimes where we have classically efficient algorithms for monomials, all three other functions remain robustly BQP-hard, or hard under usual computational complexity assumptions. In identifying classically easy regimes, among others, we show that for any polynomial of degree $\mathrm{poly}\left(n\right)$, both problems can be efficiently classically simulated when $A$ has $\mathcal{O}(\log n)$ nonzero coefficients in the Pauli basis. This contrasts with the fact that the problems are BQP-complete in the sparse access model even for constant row sparsity, whereas the stated Pauli access efficiently constructs sparse access with row sparsity $\mathcal{O}(\log n)$. Our work provides a catalog of efficient quantum and classical algorithms for fundamental linear-algebra tasks.

The unrivaled maturity of its nanofabrication makes silicon a promising hardware platform for quantum information processing. To this end, efficient single-photon sources and spin-photon interfaces have been implemented by integrating color centers or erbium dopants into nanophotonic resonators. However, the optical emission frequencies in this approach are subject to temporal fluctuations on both long and short timescales, which hinder the development of quantum applications. Here, we investigate this limitation and demonstrate that it can be alleviated by integrating the emitters into Fabry-Perot instead of nanophotonic resonators. Their larger optical mode volume enables both increasing the distance to crystal surfaces and operating at a lower dopant concentration, which reduces implantation-induced crystal damage and interactions between emitters. As a result, we observe a fivefold reduction of the spectral diffusion linewidth down to 4.0(2) MHz. Calculations and experimental investigations of isotopically purified ${}^{28}\mathrm{Si}$ crystals suggest that the remaining spectral instability is caused by laser-induced electric field fluctuations. In direct comparison with a nanophotonic device, the instability is significantly reduced at the same intracavity power, enabling a tenfold increase of the optical coherence time up to $T_2=20(1)\text{μ}\mathrm{s}$. These findings represent a key step toward spectrally stable spin-photon interfaces in silicon and their potential applications in quantum networking and distributed quantum information processing.

Atom-like emitters in solids have emerged as promising platforms for quantum sensing and information processing. Among the major challenges are inhomogeneities in emitter fine structure, which complicates quantum control. Here, we introduce a framework that leverages this emitter diversity to simplify the experimental resources needed to create optically heralded spin cluster states across $N_q$ emitters from the conventional order $O(N_q)$ to $O(1)$ within ensembles of $N_q\sim 10–100$. Specifically, the optimized pulse sequence simultaneously corrects parameter variations (pulse-length error and frequency detuning error), achieving single-qubit gate fidelities exceeding 99.99% for errors (normalized relative to Rabi-drive strength) up to 0.3, while maintaining fidelities above 99% even for errors as large as 0.4. Applying this optimized pulse sequence in the form of a Carr-Purcell-Meiboom-Gill (CPMG) based dynamical decoupling protocol to the dominant noise spectral density of silicon-vacancy centers in diamond, our approach enhances ensemble-average coherence times by more than a factor of $7$ relative to interleaved bang-bang-based CPMG. For state-of-the-art dilution refrigeration systems, we further estimate sharply reduced heating when driving a global resonant optimal dynamical decoupling across $N_q$ silicon-vacancy spins, potentially resolving the current trade-off between spin coherence and scaling to $N_q\gg 1$. We further introduce a modified single-photon entanglement protocol with an efficient algorithm for deterministic entanglement compilation. Depending on the decoupling window, our method yields order $O({10}^2–{10}^4)$ more entanglement links than bang-bang sequences, with theoretical guarantees of order $\mathrm{\Omega }(N_q)$ unique links—improvable via control-parameter tuning. Our approach thus offers enhanced fidelity, scalability, and robustness. Together, these techniques provide foundational tools—including global unitary control, phase denoising, remote entanglement, and compilation—for scalable quantum computing architectures based on heterogeneous spin ensembles.

Topological data analysis (TDA) aims to extract noise-robust features from a dataset by examining the number and persistence of holes in its topology. We provide an efficient quantum algorithm for a computational problem closely related to a core task in TDA—determining whether a given hole persists across different length scales. Further, we prove the problem itself is ${\mathsf{B}\mathsf{Q}\mathsf{P}}_1$-hard, implying that a classical solution is extremely unlikely; this stands in contrast to all previous quantum approaches to TDA, where the problems were also intractable for quantum computers, or where a rigorous proof of classical hardness still remains open. This result implies an exponential quantum speedup for this problem under standard complexity-theoretic assumptions. Our approach relies on encoding the persistence of a hole in a variant of the guided sparse Hamiltonian problem, where the guiding state is constructed from a harmonic representative of the hole.

The advancement and scaling of quantum technology has made the learning and identification of quantum systems and devices in highly multidimensional parameter spaces a pressing task for a variety of applications. In many cases, the integration of real-time feedback control and adaptive choice of measurement settings places strict demands on the speed of this task. Here we present a joint model selection and parameter estimation algorithm that is fast and operable on a large number of model parameters. The algorithm is based on variational Bayesian inference, which approximates the target posterior distribution by optimizing a tractable family of distributions, making it more scalable than exact inference methods relying on sampling and that generally suffer from high variance and computational cost in high-dimensional spaces. We show how a regularizing prior can be used to select between competing models, each comprising a different number of parameters, identifying the simplest model that explains the experimental data. The regularization can further separate the degrees of freedom, e.g., quantum systems in the environment or processes, which contribute to major features in the observed dynamics, with respect to others featuring small coupling, which only contribute to a background. As an application of the introduced framework, we consider the problem of the identification of multiple individual nuclear spins with a single electron spin quantum sensor, relevant for nanoscale nuclear magnetic resonance and for the implementation of multiqubit quantum networking nodes. With the number of environmental spins unknown a priori, our Bayesian approach is able to correctly identify the model, i.e., the number of spins and their couplings. We benchmark the algorithm on both simulated and experimental data, using standard figures of merit, and demonstrate that we can estimate dozens of parameters within minutes. Our methodology is compatible with the implementation of real-time adaptive choice of experimental settings in multiparameter estimation, which has the potential to greatly decrease the estimation time in a variety of metrology, tomography, and characterization tasks by prioritizing the measurements with the highest content of information.

This perspective outlines promising pathways and critical obstacles on the road to developing useful quantum computing applications, drawing on insights from the Google Quantum AI team. We propose a five-stage framework for this process, spanning from theoretical explorations of quantum advantage to the practicalities of compilation and resource estimation. For each stage, we discuss key trends, milestones, and inherent scientific and sociological impediments. We argue that two central stages—identifying concrete problem instances expected to exhibit quantum advantage, and connecting such problems to real-world use cases—represent essential and currently under-resourced challenges. Throughout, we touch upon related topics, including the promise of generative artificial intelligence for aspects of this research, criteria for compelling demonstrations of quantum advantage, and the future of compilation as we enter the era of early fault-tolerant quantum computing.

Understanding the capabilities of classical simulation methods is key to identifying where quantum computers are advantageous. Not only does this ensure that quantum computers are used only where necessary, but also one can potentially identify subroutines that can be offloaded onto a classical device. In this work, we show that it is always possible to generate a classical surrogate of a subregion (dubbed a “patch”) of an expectation landscape produced by a parameterized quantum circuit. That is, we provide a quantum-enhanced classical algorithm which, after simple measurements on a quantum device, allows one to classically simulate approximate expectation values of a subregion of a landscape. We provide time and sample complexity guarantees for a range of families of circuits of interest, and further numerically demonstrate our simulation algorithms on an exactly verifiable simulation of a Hamiltonian variational Ansatz and long-time dynamics simulation on a 127-qubit heavy-hex topology.

Arrays of neutral atoms present a promising system for quantum computing, quantum sensors, and other applications, several of which would profit from the ability to load, cool, and image the atoms in a finite magnetic field. In this work, we develop a technique to image and prepare ${}^{87}\mathrm{Rb}$ atom arrays in a finite magnetic field by combining electromagnetically induced transparency cooling with fluorescence imaging. We achieve an average readout fidelity of $99.7(1)\mathrm{\%}$ at $98.2(3)\mathrm{\%}$ survival probability and up to $68(2)\mathrm{\%}$ single-atom stochastic loading probability in a 2.3 G magnetic field, with performance validated at fields up to 10 G. We further develop a model to predict the survival probability, which also agrees well with several other atom-array experiments. Our technique cools both the axial and radial directions, and will enable future continuously operated neutral-atom quantum processors and quantum sensors.

Rydberg–cavity systems are emerging as promising platforms for quantum simulation and quantum information processing. These hybrid architectures combine two complementary interaction mechanisms: cavity photons mediate collective long-range couplings, while Rydberg excitations generate strong short-range interactions. Together, they offer a setting for engineering many-body phases characterized by a hierarchy of interactions across widely different length scales. In this work, we introduce a minimal and scalable model for such systems. Focusing on the strong Rydberg blockade regime, we restrict the Hilbert space to the subspace enforced by the blockade, yielding a kinetically constrained long-range model in one spatial dimension. This approach both captures the physics of Rydberg–cavity experiments in the regime of strong Rydberg interactions and provides a conceptually transparent framework for studying the interplay of long-range and short-range interactions. At equilibrium, in addition to paramagnetic and Néel-ordered phases, the system supports a blockaded ferromagnetic/superradiant phase, distinct from the conventional superradiant phase. Out of equilibrium, we identify long-range quantum many-body scars, which are atypical nonthermal eigenstates that evade the eigenstate thermalization hypothesis, and giving rise to slow entanglement growth. In contrast to the linear-in-time entanglement growth characteristic of short-range scarred models, these long-range scars exhibit logarithmic entanglement dynamics. Our results establish a minimal yet versatile framework for Rydberg–cavity systems, and provide a stepping stone for future theoretical and experimental studies of this frontier platform in quantum many-body physics.

Classical simulation of quantum circuits plays a crucial role in validating quantum hardware and delineating the boundaries of quantum advantage. Among the most effective simulation techniques are those based on the stabilizer extent, which quantifies the overhead of representing non-Clifford operations as linear combinations of Clifford unitaries. However, finding optimal decompositions rapidly becomes intractable as it constitutes a superexponentially large optimization problem. In this work, we exploit symmetries in the computation of the stabilizer extent, proving that for real, diagonal, and real-diagonal unitaries, the optimization can be restricted to the corresponding subgroups of the Clifford group without loss of optimality. This “strong symmetry reduction” drastically reduces computational cost, enabling optimal decompositions of unitaries on up to seven qubits using a standard laptop—far beyond previous two-qubit limits. Additionally, we employ a “weak symmetry reduction” method that leverages additional invariances to shrink the search space further. Applying these results, we demonstrate exponential runtime improvements in classical simulations of quantum Fourier transform circuits and measurement-based quantum computations on the Union Jack lattice, as well as insights into the nonstabilizer properties of multicontrolled phase gates and unitaries generating hypergraph states. Our findings establish symmetry exploitation as a powerful route to scale classical simulation techniques and deepen the resource-theoretic understanding of quantum advantage.