In briefUsing topology to classify all mutually unbiased bases in six dimensions
SummaryThe paper studies mutually unbiased bases (MUBs), which are complementary quantum measurements in which certainty in one basis gives no information about the others. Using the topology of the space of complex Hadamard matrices, which describe transitions between MUBs, the authors classify all sets of MUBs in dimension six. They show that every mutually unbiased triple belongs to a two-parameter family or is a single isolated triple built from the Fourier matrix, which implies that no four MUBs exist.
WhyThe authors use topological connectedness to classify all mutually unbiased triples in dimension six and report that no fourth basis exists.
In briefExact Resource Depth for Single-Qubit Dyadic Quantum Programs
SummaryThis paper studies a single-qubit dyadic fragment of quantum programs, where the deepest exact resource layer can be determined exactly rather than bounded. The authors prove a descent theorem for a projective Clifford-cyclotomic tower and combine it with an arithmetic-closure invariant for adaptive stabilizer processing. They show that three depth measures coincide: the minimal real cyclotomic field, the shallowest gate alphabet, and the shallowest program-state layer. The result is packaged as an exact static resource effect with an outlined certificate-producing compiler.
In briefA 3D-printed one-piece optical mount base for an unbalanced interferometer
SummaryThis paper presents a single-piece optomechanical base built with laser powder bed fusion, a metal 3D printing method, from Invar36, a low-thermal-expansion alloy. Optical mounts, flexures, and alignment features are built directly into the base, which the authors say produces a smaller footprint than conventional assemblies. The authors suggest the approach could support lightweight, customized optical systems and next-generation quantum sensors.
In briefTesting gravity at megahertz frequencies with levitated rotors
SummaryThe paper proposes a near-field experiment in which a levitated microrotor modulates a mass quadrupole and a second rotor measures the resulting gravitational tidal field as a torque, reaching megahertz frequencies. At 2 MHz the estimated torque is 2×10⁻²⁵ Nm. The authors argue this could constrain any delay in gravitational coupling to the 10⁻⁸ s level, more than three orders of magnitude better than existing bounds.
In briefOptimal Holding Times for Entanglement Swapping: tuning how long quantum entanglement is stored to maximize swap throughput
SummaryThis paper studies how long entanglement should be held in memory-based quantum repeaters to maximize entanglement swapping throughput under strict, probabilistic, and mean fidelity constraints. It considers two heterogeneous quantum links that differ in decoherence rates, memory capacities, attempt rates, success probabilities, and classical communication latencies. The authors develop models to estimate expected throughput and the conditional wait-time distributions for swapping, then optimize the maximum holding times of both links to meet each fidelity requirement. The proposed model and approach are validated with a discrete-event quantum network simulator.
In briefQuantum algorithms for computing molecular response functions, compared across six routes with query lower bounds.
SummaryThe paper compares six quantum algorithm routes for molecular response functions within one resource-accounting framework, including resolvent-kernel and time-domain sampling approaches. It states conversions between the resource calibers of these routes and analyzes their error budgets and readout costs. It also proves query lower bounds for several tasks and presents an accounting map across the routes.
In briefGFlowQEC: generative networks that learn to sample quantum error-correcting codes
SummaryChoosing quantum error-correcting codes requires scoring candidates with circuit-level simulations that are slow, so only a few hundred evaluations are possible. The paper introduces GFlowQEC, described as the first generative flow network for quantum code design, which builds codes one group element at a time and proposes them in proportion to a power of their score. Trained once and frozen, it proposes lifted-product codes with 180 physical qubits. Per paid evaluation, it finds more than twice as many new top-scoring codes as tuned parallel tempering.
Nuno Batista, Gabriel Falcao, and Penousal Machado
In briefLinear-time recovery of erased bits in quantum expander codes
SummaryThis paper introduces Small-Set-Peeling, a quantum generalization of the classical Peeling decoder for erasure errors in quantum expander codes. The decoder can correct up to D−1 erasures, where D is the design distance of the code, and it is guaranteed to run in linear time. Combined with the erasure-conversion algorithm Small-Set-Find, the authors also obtain a graph-local linear-time decoder for Pauli errors.
WhyThe authors generalize the classical Peeling decoder to quantum expander codes, reporting the first guaranteed linear-time correction of D-1 erasures, combined with erasure conversion for Pauli errors
In briefFinite-temperature transition-window weights on a fault-tolerant quantum computer
SummaryThis paper extends the excite-filter-readout framework from zero-temperature pure ground states to Gibbs states, where filters act on absolute energies and thermal responses depend on transition frequencies. Using a Liouville-space and thermofield-double identity, it expresses the thermally weighted transition-window weight as a spectral projection and builds a routed block-encoding, a conjugate-pair readout, and a compatibility analysis of four routes. It gives a per-window query bound and reports that the preparation factor grows exponentially with system size at fixed temperature in positive-entropy-density systems.
In briefResource-efficient ways to formulate the Knapsack problem for quantum computers
SummaryThis paper examines how to express the Knapsack problem as a QUBO form that quantum computers can solve. It compares three resource-efficient formulations: a smaller penalty constant, fewer slack variables, and a method that avoids slack variables entirely. In simulations of the linear ramp QAOA and experiments on a D-Wave quantum annealer, all three approaches raised the chance of finding the optimal solution, with the slackless method performing best.
Pim van den Heuvel, Vrinda Mehta, Dennis Willsch et al.
In briefQuantum entanglement between two particles colliding at non-relativistic speeds
SummaryThis paper numerically studies how entanglement entropy builds up between two non-relativistic particles that collide. The particles follow the ordinary Schrödinger equation with local, finite-range potentials, starting from the product of two independent wave packets. It examines several scenarios, including tunneling and mirror effects at low energy in front of a barrier, higher energies, and attraction instead of repulsion. Although the wave functions and density matrices differ significantly across these situations, the time trends and values of entanglement entropy look strikingly alike.
In briefNon-compact Kogut-Susskind gauge theories with exact SL(2,R) gauge invariance
SummaryThis paper revisits the Kogut-Susskind construction of gauge theories on a one-dimensional spatial lattice for the non-compact gauge group SL(2,R), keeping gauge invariance exact. Compared with compact groups, the Casimir operators of the gauge fields have a continuous spectrum and matter fields carry infinite-dimensional representations. For the pure gauge theory, restricting to the positive semigroup SL+(2,R) reproduces known results from the quantization of Jackiw-Teitelboim gravity. The authors also add static charges and dynamical fermionic matter, obtaining an interacting one-dimensional SL(2,R) Kogut-Susskind gauge theory.
In briefNear-quantum-limited readout for an axion search using a sub-GHz amplifier
SummaryThis paper presents a kinetic inductance traveling-wave parametric amplifier operating over 300-700 MHz, with gain above 30 dB and pump powers of a few hundred nanowatts. Measured with a variable-temperature noise source, its system noise is 2.5-3.5 quanta. The authors characterize performance relevant to axion searches, including gain drift, compression point, noise at different temperatures, readout with an input circulator, and magnetic field insensitivity.
Joe Wiedemann, Roman Kolevatov, Jianyang Fu et al.
In briefA quantum lower bound for finding triangles in graphs
SummaryThis paper studies triangle finding, a basic graph problem where the input is an N-vertex graph given by its adjacency matrix. The best known quantum query upper bound was O(N^{5/4}), while the best lower bound was the basic Ω(N) from unstructured search. The authors prove that Ω(N^{13/12}) quantum queries are necessary, the first superlinear quantum query lower bound for this problem. The same bound also holds for triangle detection and for detecting any fixed connected non-bipartite graph.
WhyThe authors report the first superlinear quantum query lower bound Ω(N^{13/12}) for triangle finding, using compressed oracles and transfer methods.
In briefSolving a non-abelian hidden subgroup problem on a trapped-ion quantum computer, examining how the measurement basis matters.
SummaryThe paper experimentally studies the hidden subgroup problem over q-hedral groups on the IonQ Forte trapped-ion quantum computer. It compares weak Fourier sampling, strong Fourier sampling in an adapted basis, adjacent-pair sampling, and forgetful abelian sampling. The hardware results largely match the theoretical predictions of Moore, Rockmore, Russell and Schulman. For p=7 and q=6, the device exceeds the success probability of an exhaustively optimised classical algorithm for up to three oracle queries.
In briefPhase-locked quantum battery charging through dissipative mode selection
SummaryThis paper studies how dynamical instability and reservoir engineering can improve charging quality in a bosonic two-mode quantum battery. Dissipative coupling selectively amplifies a single unstable quasi-dark mode, giving phase-locked exponential charging at a lower gain threshold. Coherent coupling instead amplifies two frequency-split modes, producing persistent beating. The dissipative approach yields near-unit ergotropy fraction and a higher work-like signal-to-noise ratio.
Borhan Ahmadi, Lei Du, Andr\'e H. A. Malavazi et al.
In briefQASPA-MaxCut on Unweighted Graphs: Edge Encoding and QAOA Restarts
SummaryThis paper presents QASPA-MaxCut, a shift-phase quantum circuit for the Max-Cut problem that encodes graph edges through controlled Ry rotations. The authors compare one-layer QASPA-MaxCut with depth-one QAOA using exact state-vector simulation on random backbone-augmented graphs and nineteen custom unweighted topologies. QAOA improves with more classical restarts, though the reported mean summaries remain lower than the QASPA selection scores.
In briefExperimental revival of hidden metrological steerability in noisy quantum channels via stochastic filtering
SummaryThis paper studies steering-assisted quantum metrology, where a trusted party improves local phase estimation using classical information from an untrusted party. In phase-covariant amplitude-damping channels, stochastic pre- and post-filtering restores violation of the metrological steering inequality (MSI) that was initially absent. A photonic experiment with polarization qubits, a Sagnac interferometer, and Brewster-angle glass plates shows the critical damping parameter rising from about 0.5 to about 0.75. This comes at the cost of a reduced success probability.
In briefCalibrating flux crosstalk in tunable superconducting circuits, from zero prior knowledge to high precision
SummaryThis paper addresses flux crosstalk in superconducting processors built from frequency-tunable qubits and couplers, which undermines independent flux control. It presents two methods that characterize coupler crosstalk without directly driving or measuring the couplers: zero-knowledge characterization, which uses intrinsic coupler-resonator coupling for fast estimates, and symmetric branch cancellation, which isolates crosstalk for high-precision quantification. Combined with an improved iterative procedure, residual coupler crosstalk is suppressed below 5×10⁻⁵.
In briefEfficient quantum ultraspherical transform and a well-conditioned quantum spectral method: a transform and spectral solver for solving linear differential equations on quantum computers.
SummaryThe paper constructs a quantum ultraspherical transform (QUT) for orthonormal ultraspherical polynomials, converting between amplitudes of function values and expansion coefficients. For each fixed integer m≥2 and dimensions with N+m=2^n, the transform achieves error at most ε with a gate count polynomial in n and log(1/ε), built on the quantum Fourier transform. Using the QUT, the authors develop a quantum spectral solver for regular linear ordinary differential equations, where preconditioning keeps condition numbers bounded independently of basis size under suitable assumptions.
WhyThe authors construct a quantum ultraspherical transform for all integer m≥2 and a spectral solver for linear ODEs with bounded preconditioned condition numbers.
In briefPrecision limits and non-Gaussian advantages in optical phase estimation under preparation constraints
SummaryThis paper studies precision limits in optical quantum metrology for phase estimation. It introduces a constraint, an envelope on the photon-number distribution arising from physical limits of state-generation methods, that separates physically admissible probe states from inadmissible ones. Using this constraint for pure single-mode probes, the authors define a resource quantity N_RMS fixed by the preparation constraints and derive a finite bound on the quantum Fisher information whose Heisenberg scaling is attainable. They also identify a metrological advantage of non-Gaussian states over Gaussian probes under the same constraints.
In briefTraining quantum circuit Born machines with noise as regularization under limited measurements
SummaryThis paper studies how hardware noise affects the generalization of quantum circuit Born machines (QCBMs), which learn probability distributions from data. The authors show that the generalization gap, when trained with maximum mean discrepancy, is bounded by the complexity of the distributions the noisy model can reach. Residual noise acts as a regularizer, and its optimal strength can be intermediate. They propose GENIE, a framework that selects the noise level during training from held-out data within a fixed execution budget, with provably bounded regret. Simulations on synthetic and census data show GENIE stays close to the best of four noise-control strategies.
In briefA counterexample to two conjectures about weak Gibbs and regularity properties
SummaryThe paper constructs a counterexample that disproves the weak Gibbs conjecture and the regularity conjecture. The example is a spin-one system on the square lattice with a translation-invariant nearest-neighbor interaction. At high temperatures and high on-site field strengths, the system has a unique KMS state that lacks both the weak Gibbs and regularity properties.
In briefCooling the environment to curb correlated atom loss in Rydberg arrays
SummaryThis paper studies how blackbody radiation triggers avalanche-like correlated atom loss in Rydberg atom arrays. Using a 45 K radiation shield, the authors reduce ambient blackbody radiation, which extends nS lifetimes and suppresses decay rates by 30% to 50%. Avalanche rates in cryogenic arrays drop in step with this suppression, and a parameter-free rate model matches the measurements.
Zhenpu Zhang, Brendan P. Marsh, Ting You Tan et al.
