Papers

Optically addressable defect spins in semiconductors are promising building blocks for scalable quantum technologies. Yet, the consequences of conventional contact schemes used in semiconductor device integration for the quantum spin environment remain largely unexplored. Using the silicon vacancy (V$_\mathrm{Si}$) in silicon carbide (SiC) as a model system, we show that a widely used contact metal, nickel (Ni), intrinsically perturbs the defect spin state and quenches the characteristic emission from the spin-quartet channel of V$_\mathrm{Si}$. Low-energy muon spin rotation further reveals that Ni contacts create a magnetically contaminated region extending at least $\sim$110 nm into the SiC, in stark contrast to non-magnetic contacts such as Ti and Al. Moreover, cross-sectional cathodoluminescence measurements on samples with high defect density demonstrate a suppressed quartet-state luminescence from the V$_\mathrm{Si}$ up to 500 nm beneath the Ni contact. This drastic influence on the defect's magnetic environment is accompanied by the appearance of a signature consistent with photoluminescence from the spin-doublet state of V$_\mathrm{Si}$ under the Ni layer, which was absent in other material stacks. These results establish that standard device configurations can drive quantum defects into unwanted charge and spin configurations, underscoring the necessity of precise design to preserve quantum-grade spin environments in semiconductor devices.

Automated Program Repair (APR) increasingly relies on Large Language Models (LLMs). ChatGPT-enhanced APR uses techniques such as self-correction and autonomous agents to improve repair without modifying model parameters. Although these approaches report strong results on Defects4J and SWE-bench, the stability of enhancement gains across benchmarks remains under-explored. We evaluate three ChatGPT-enhanced APR methods on three representative, long-standing benchmarks. With GPT-3.5-Turbo, SRepair achieves a larger absolute gain on HumanEval-Java than on Defects4J, while SRepair and FixAgent without extrinsic information yield negative gains on BugsInPy. With GPT-5.4-mini, the evaluated methods achieve larger absolute gains on Defects4J than on HumanEval-Java, while gains on BugsInPy are non-negative but limited. We investigate benchmark-related factors through code transformations and benchmark-specific fine-tuning. Code transformations reduce enhancement gains on Defects4J, while benchmark-specific fine-tuning increases gains on BugsInPy. Directly supplying GPT-3.5-Turbo with error messages and triggering tests yields more correct repairs than the evaluated ChatGPT-enhanced APR methods on BugsInPy. These findings highlight the need to evaluate generalizability across benchmarks and models using multiple metrics, and suggest that directly providing repair-specific extrinsic information may be more effective than enhancement methods when their gains are limited.

Accurate and timely detection of degradation in lithium-ion batteries is crucial to ensure safety, reliability, and longevity in high-demand applications such as electric vehicles and energy storage systems. Traditional incremental capacity (IC) analysis methods require low-current cycling for discharge measurements, limiting their practical use in real-time battery management. This paper proposes a novel neural network-based framework that predicts discharge IC features directly from charging signals, eliminating the need for low-current discharge. Trained on a comprehensive dataset of 53 battery cells cycled under diverse fast-charging protocols, the model demonstrates robust generalization ability, effectively estimating degradation indicators on unseen battery data. Among several architectures evaluated, the LSTM model provides the best balance of prediction accuracy and computational efficiency. The proposed approach enables real-time integration into Battery Management Systems (BMS), enhancing degradation monitoring without disrupting normal battery operation. This paper presents a study to bridge IC analysis with practical, fast-charging scenarios, marking a significant step towards intelligent and scalable battery health monitoring.

We give a unified arithmetic and geometric treatment of the Hall-Janko simple group and the icosian Leech lattice. Let $\mathcal{I}$ be the standard icosian maximal order over $K=\mathbb{Q}(\sqrt5)$. Following the mass-formula and Venkov framework for the Niemeier and Leech lattices, we classify positive definite unimodular Hermitian $\mathcal{I}$-lattices of quaternionic rank $3$. Hashimoto's mass formula, the Siegel-Weil average, and Gundlach's structure theorem determine the scalar theta series of every nonsplit class. A two-place harmonic separation of the Leech minimal shell proves that its three Hermitian components are quaternionic projective $5$-designs. For the norm-$2$ shell, integrality and projective moments recover the complete angle scheme and its $525$ orthogonal frames. The finite code of one frame then saturates the residual mass, giving exactly the split and Tits classes and an independent computation of the order of $\operatorname{Aut}_{\mathcal{I}}(L_{\mathrm T})\cong2.J_2$. Finally, a quaternionic line system with the prescribed five projective inner products has at most $315$ points, with equality precisely for the Hall-Janko configuration up to $P\operatorname{Sp}(3)$. Equality in Hoggar's bound supplies the design property; Cohen-Tits uniqueness, golden-angle fission, and the centered cubic moment give geometric rigidity.

Microresonators based on diverse material platforms have driven the progress of integrated photonics by enhancing light-matter interactions in compact volumes. Extending this strategy to magneto-optical (MO) media is highly attractive for nonreciprocal photonics, magnonics, and quantum transduction, yet compact high-Q resonators based on single-crystalline yttrium iron garnet (YIG) have remained elusive because of the difficulty of its nanofabrication. Here we demonstrate high-Q, small-mode-volume MO microdisk resonators based on monocrystalline Bi-substituted YIG (Bi:YIG). Low-loss Bi:YIG thin films prepared by bonding and thinning, combined with optimized Ar-plasma etching, enable submicron-thick, air-suspended microdisks that support telecom-band whispering-gallery modes with $Q = 1.52\times 10^5$, a mode volume $V \approx 200$ cubic wavelengths, and $Q/V \approx 760$, representing an orders-of-magnitude improvement for integrated high-Q YIG MO cavities. The device exhibits clear magnetic-field-dependent spectral responses, including reciprocal resonance shifts and nonreciprocal frequency splitting of counter-propagating modes. A perturbative analytical model quantitatively reproduces these responses and indicates an effective MO coefficient about twice the bulk value, suggesting a possible miniaturization-induced enhancement. These results establish monocrystalline YIG microdisks as a compact high-Q MO platform for integrated isolators, optomechanics, microcombs, and photon-magnon quantum interfaces.

Force-conditioned vision-language-action (VLA) policies can respond to contact, but when trained solely on demonstrations, their recovery behavior may be limited by demonstration coverage, and they do not learn from deployment outcomes. Human corrective imitation provides additional recovery examples, but its objective matches local action targets without explicitly optimizing task return. We present ForceRFT, a force-guided residual reinforcement learning framework that learns contact-dependent corrections from human supervision and autonomous task outcomes. A frozen, demonstration-trained SmolVLA-based prior generates force-conditioned action chunks, while a lightweight residual actor refines individual end-effector pose commands using wrist feedback acquired during chunk execution. The decision-time wrist wrench, its temporal change, and the selected base motion condition both residual correction and value estimation. Human corrections supervise the residual actor, while verified autonomous transitions train the twin critics and support value-guided updates to the same actor. Bootstrapping is restricted to autonomous segments, preventing TD credit from crossing human-intervention boundaries. Real-robot experiments on plug insertion, ring-on-peg assembly, and whiteboard wiping show higher autonomous success rates than the evaluated demonstration-trained and residual-imitation baselines. Comparisons with residual imitation support value-guided residual optimization, while plug-insertion ablations indicate the benefit of direct execution-time wrist feedback.

Can simple learning rules keep their regret bounded in self-play? Recent work achieves constant regret bounds through modified regularization and higher-order prediction. Yet for Optimistic Hedge, arguably the most canonical method in games, the best known individual regret bound remains logarithmic. In this work, we prove that plain Optimistic Hedge with a constant step size can attain $O_{n,d}(1)$ individual regret in general-sum games with $n$ players and $d=(d_1,\ldots,d_n)$ actions, under expected loss-vector feedback. As a corollary, its time-averaged play enjoys an $O_{n,d}(1/T)$ coarse correlated equilibrium (CCE) gap. Our analysis represents Optimistic Hedge as a real-analytic recurrence on a compact space, which yields an exact finite-order difference relation that eliminates horizon dependence. Our proof hinges on nonconstructive Noetherianity argument of Frisch (1967), so the $(n,d)$-dependence remains implicit.

Neuromorphic engineering began with the idea that the physical behavior of a system could itself be used for computation, taking inspiration from the way nervous systems sense, adapt, and evolve in time. The field has since expanded far beyond its early analog circuits to include event-based sensors, spiking processors, emerging memory devices, mixed-signal systems, and large-scale neural accelerators. With this expansion, however, the meaning of neuromorphic has become increasingly broad. In this Perspective, we argue that neuromorphic engineering should be defined neither by resemblance to biological components nor by any particular device, signal representation, or substrate. Its potential lies in identifying computational principles in biological systems and translating them into the organization and dynamics of artificial machines. A system is truly neuromorphic when the invoked biological or physical principle plays a causal, design-relevant role in how information is represented, how state evolves, or what capability the complete system achieves. We develop this view across sensing, collective computation, memory, learning, and interaction. Adaptation, nonlinear dynamics, attractor structure, variability, and closed-loop action illustrate how computation can be embedded in a machine's evolving physical state. From these examples, we identify a set of design principles for neuromorphic systems. As neuromorphic engineering moves toward wider deployment, this perspective shifts the central question from how closely machines resemble nervous systems to what computation becomes possible when their principles are understood and deliberately translated into new machines.

This paper develops the covariant Boltzmann equation governing the coupled evolution of photons and massive particles in general curved spacetimes. We construct the collision integrals for photon emission-absorption, Compton scattering, and scattering between massive particles, and show the conservation laws. Requiring the H-theorem to hold determines the relation between the absorption and emission transition rates, which reduces to the Einstein relation in the non-relativistic limit. At equilibrium the entropy production vanishes, yielding the detailed balance condition, from which we obtain the covariant thermodynamic relations of the photon gas. We illustrate these results with the thermal radiation from an accretion disk in Kerr spacetime, first in global equilibrium and then in a Keplerian disk in local thermodynamic equilibrium. We further extend the collision integrals to arbitrary multi-species processes involving both null and massive particles. The formulation treats non-equilibrium radiative processes in strong gravitational fields from first principles.

Time series foundation models (TSFMs) have recently delivered impressive zero-shot performance across diverse forecasting tasks. However, real-world decision-making frequently relies on irregular multivariate time series (IMTS), where inconsistent inter-observation intervals and asynchronous sampling across variables coexist with informative missingness. Existing TSFMs handle such inputs either through imputation that injects spurious values or through index-based positional encodings that ignore continuous time. There is still a gap in the foundation model that follows the original IMTS patterns. In this paper, we propose a hybrid attention model that learns a unified time-aware patch representation for IMTS forecasting. We first design a time-aware patch encoding that maps a variable number of intra-patch timestamps into a fixed-size embedding, producing a uniform format for irregular patches without resorting to imputation. We then introduce a time bias attention mechanism that calibrates inter-patch temporal misalignment and asynchronous cross-channel dependencies as auxiliary attention offset. Finally, on top of a decoder-only Transformer backbone, we adopt a hybrid causal mask that preserves a bidirectional full view over the historical context while keeping the forecast horizon strictly autoregressive. To support large-scale pretraining under irregular settings, we also curate VersaTSA, an archive of $30$B observations that retains the native sampling sparsity of its sources. Experiments on three IMTS benchmarks and a standard regular-MTS benchmark show that our model achieves state-of-the-art zero-shot performance on IMTS and remains competitive when transferred to regular forecasting.

Machine learning methods for Android malware family classification have achieved high accuracy, but their application is hindered by two major challenges. First, the widely used code obfuscation severely disrupts the automated labeling process and introduces substantial label noise into training datasets. Second, training datasets often exhibit severe class imbalance, leading to poor performance of family classification models. Although existing studies have proposed various solutions to either label noise or class imbalance, they often overlook the interplay between these two factors. Under class imbalance, the presence of hard-to-learn minority-class samples can significantly impair the effectiveness of existing countermeasures for noisy samples. To jointly address label noise and class imbalance, we propose a robust Android malware family classification framework, RoMaC. It employs a self-training strategy to correct noisy labels and, more importantly, discriminately treats head-family and tail-family samples. This design effectively mitigates the adverse impact of class imbalance on noise-robust learning. Moreover, RoMaC integrates a class reweighting mechanism with multi-model ensemble learning, thereby enhancing both classification accuracy and noise robustness. We evaluate RoMaC on a combined dataset constructed from two public datasets. When 30% of the samples are obfuscated, RoMaC achieves an overall Macro-F1 score of 0.803 and an accuracy of 0.871, as well as a tail-class Macro-F1 score of 0.672 and an accuracy of 0.784. Compared with existing methods, RoMaC demonstrates performance improvements of 6%-20% across various obfuscation scenarios and noise levels.

Open-vocabulary semantic segmentation (OVS) of remote sensing imagery is a challenging pixel-level task requiring strong generalization and adaptation to the spatial characteristics of remote sensing data. Although existing vision-language foundation models perform well in general domains, their image-level classification design weakens the spatial priors needed for high-resolution remote sensing segmentation: structural spatial relations are degraded during deep feature transformation, and fine-grained spatial details are lost during downsampling. To address these complementary deficiencies, we propose SatOV, a training-free framework for open-vocabulary remote sensing segmentation that restores spatial priors at two stages of the representation pipeline. Specifically, Residual QQ Attention (ResQQ) extracts Query-Key self-attention from an intermediate CLIP layer and fuses it with final-layer Query-Query attention via a residual combination, restoring structural spatial priors suppressed by the final-layer representation. Spatially Modulated Upsampling (SatUp) uses the original high-resolution RGB image as spatial guidance, combining spatial feature modulation with guided cross-attention to reconstruct pixel-level textures and boundaries. Extensive experiments on DOTA, UDD, LoveDA, and Vaihingen show that SatOV consistently improves training-free OVS and achieves competitive quantitative and qualitative results against state-of-the-art methods. These results validate the effectiveness of restoring spatial priors at both the representation and spatial-resolution stages for remote sensing open-vocabulary segmentation.

We study personalized federated reinforcement learning, in which $n$ agents, each acting in its own Markov decision process, collaborate through a server to learn a shared MAML-style policy initialization that becomes effective for an individual agent once that agent adapts it with a single local policy-gradient step. We propose Per-FedAvg-PG, in which agents take $\tau$ local stochastic meta-policy-gradient steps between communication rounds, and prove that it reaches an $\varepsilon$-approximate first-order stationary point of the personalized objective in $K=\mathcal O(\varepsilon^{-3/2})$ rounds with $\tau=\Theta(\varepsilon^{-1/2})$ local steps. The analysis rests on a structural feature of the reinforcement learning setting: under standard policy-class regularity, the per-agent objectives have uniformly bounded gradients and Hessians with explicit constants, so the bounded-gradient and bounded-heterogeneity conditions imposed by the supervised theory hold automatically and no separate heterogeneity assumption is needed. The exact meta-gradient requires the inner-loop policy Hessian, which our experiments identify as the practical bottleneck. We therefore analyze the Hessian-free variant, bound its bias, and exhibit fixed points at which the meta-gradient is nonzero and of order $\alpha$, showing that the resulting stationarity floor is a property of the method rather than of the bound. Experiments on tabular and neural navigation confirm the predicted behavior and show transfer to unseen agents at an order of magnitude lower sample cost than independent training. Together these results identify the adaptation step size as a tunable personalization knob and the curvature estimate as the quantity that governs whether exact meta-gradients are affordable.

We classify compact Kähler–Einstein surfaces whose metric is induced by a holomorphic isometric immersion into a finite-dimensional complex projective space. No symmetry assumption and no bound on the codimension are imposed. We prove that the only such surfaces are \[ (\PP^2,m g_{\FS}) \quad\text{and}\quad \bigl(\PP^1\times\PP^1,m(g_{\FS}\oplus g_{\FS})\bigr), \qquad m\in\mathbb Z_{>0}, \] realized respectively by the Veronese and Segre–Veronese embeddings. The main new ingredient is a codimension-independent exclusion of the entire Fano-index-one branch, combining a common anticanonical root construction with Gram–Gauss rank estimates and, in degree five, an equivariant curvature argument. Consequently, every connected compact Kähler–Einstein surface whose metric is induced by a holomorphic isometric immersion into a finite-dimensional complex projective space is homogeneous.

Generally, the work output of stochastic heat engines is governed by the stochastic trajectory distribution and the energy spectrum. Because the trajectory distribution depends on the thermal reservoir temperatures, the work output is not only constrained by temperature but is also susceptible to thermal fluctuations. To address these limitations, we introduce two continuous Maxwell's demons into a quantum Otto cycle, forming an Otto information engine (OIE). We demonstrate that the work output of the OIE depends solely on the energy level gap, enabling arbitrary work extraction while eliminating work fluctuations, thereby ensuring cycle-to-cycle identical work output. Furthermore, we show that the engine's efficiency can surpass the standard Carnot efficiency even after accounting for the energy cost of demon's memory erasure. Finally, we show that the OIE can deliver superior output power even when the demon's measurement time is taken into account, and the corresponding Monte Carlo simulation has been executed.

Detecting human acoustic cues from UAV-mounted microphones could support acoustic search and rescue, but rotor ego-noise often masks speech, cries, coughs, and other human sounds at extremely low SNRs. We study UAV human-audible-presence detection under this real operating constraint. Models are trained on a reproducible synthetic mixture pipeline built from public audio, but selected and evaluated on real DroneAudioSet recordings using a metadata-defined audibility filter, recording-grouped Dev/Test splits, and group-bootstrap confidence intervals. Our results show that synthetic accuracy is a weak and non-monotonic proxy for real UAV transfer: a from-scratch SE-ResNet appears competitive on synthetic mixtures but collapses on real ego-noise, while frozen audio foundation models and lightweight adapters transfer more reliably. We further evaluate a BEATs adapter family with rotor-aware conditioning and domain regularization. The Real-Dev-selected EgoRAP-DA configuration achieves the best locked-test low-false-alarm recall among the candidates, but its advantage over a vanilla adapter is not statistically significant under paired group bootstrap. The main contribution is therefore a real-validated benchmark and evaluation protocol showing that honest progress in UAV audition requires real, group-level validation rather than synthetic scores alone.

Collecting whole-body demonstrations for humanoid manipulation mostly relies on teleoperation, which is costly and hard to scale up. The Universal Manipulation Interface (UMI) provides a scalable data collection paradigm, but end-effector trajectories alone underdetermine humanoid whole-body coordination, which is insufficient for whole-body demonstration collection. Therefore, we introduce Whole-Body UMI (WB-UMI), a task-agnostic, real-time and end-effector conditioned motion generator that decouples whole-body coordination learning from task semantics learning through a shared end-effector interface. A diffusion policy learns from native UMI demonstrations, while WB-UMI learns independently from retargeted motion capture, requiring no body trackers or paired image–whole-body demonstrations during task-specific data collection. In real deployment, an asynchronous hierarchy integrates the diffusion policy, motion generator, and a whole-body controller with latency compensation and measured-state feedback. Real-robot experiments on G1 support real-time closed-loop transfer across four tasks, achieving 90% success in drawer closing, 80% in shelf pick-and-place, 30% in ball toss, and 40% in Loco-PnP, which shows the effectiveness of this hierarchy in transferring native UMI skills to humanoid whole-body manipulation.

We study data-driven policy construction for dynamical systems using trajectories generated by surrogate finite-horizon Model Predictive Control (MPC), with the goal of approximating an infinite-horizon optimal control policy. We interpret these trajectories as non-expert demonstrations and propose a memory-based, nonparametric policy that constructs a Lipschitz-regularized upper envelope of the finite- horizon surrogate value function, which serves as an explicit upper bound on the value of the resulting policy, thereby providing a computable performance certificate while bypassing explicit optimization at runtime. We establish relative optimality guarantees with respect to the infinite-horizon optimal value function and derive sufficient MPC horizon conditions, together with sample complexity bounds, for achieving a prescribed relative error. Numerical experiments on a rocket landing task validate the theoretical predictions and demonstrate near-optimal performance.

The effectiveness of active rear steering (ARS) control for four-wheel steering (4WS) vehicles is widely recognized in the automotive industry. At low speeds, ARS control can enhance maneuverability by steering the rear wheels in the opposite direction to the front wheels, reducing the turning radius. In contrast, at high speeds, ARS control can improve stability by steering the rear wheels in the same direction as the front wheels, preventing oversteer behavior. However, the performance of ARS control is often limited by the reference model used to generate the desired yaw rate, which is typically derived from a steady state of front-wheel steering (FWS) vehicle model. In this paper, we conduct a numerical analysis to construct an optimal yaw rate reference map for ARS control by formulating a constrained optimization problem. In the optimization problem, constraints are imposed to ensure that the vehicle operates within safe limits at steady state. Numerical simulations demonstrate the effectiveness of the proposed method in providing an optimal yaw rate reference map for ARS control.

Standard stochastic heat engines operate blindly, enforcing work extraction protocols indiscriminately on microstates, and consequently suppressing work output, stability, efficiency. To overcome this, we propose an Otto information engine (OIE) that employs a Maxwell's demon to filter out detrimental stochastic trajectories, and demonstrate that the OIE can provide enhanced work output and stability compared to the corresponding standard Otto engine. Remarkably, even after accounting for the energetic costs of the demon, the OIE efficiency surpasses both the standard Otto limit and Carnot bound. Furthermore, by applying the fluctuation theorem of information dissipation, we derive upper and lower bounds on efficiency, confirming that our results adhere to the second law of thermodynamics. Finally, and counterintuitively, measurement errors and quantum inner friction, traditionally considered deleterious, can be harnessed as resources, which improves robustness and enables us to ignore the adiabatic strokes time.

Let $\phi$ be a quaternion of modulus $1$. In this article, we study some topics related to $\phi$-conjugation for quaternionic matrices, including $\phi$-Hermitian matrices, $\phi$-conjugate normal matrices, unitary $\phi$-congruence and $\phi$-HSH decomposition (decomposition of a $\phi$-Hermitian matrix and a skew $\phi$-Hermitian matrix). In particular, we generalize the Autonne-Takagi factorization of quaternion $\phi$-Hermitian matrices for all unit quaternion $\phi$. This gives an affirmative answer to a problem proposed by R. Horn and F. Zhang in the paper “A generalization of the complex Autonne-Takagi factorization to quaternion matrices, Linear Multilinear A. 60: 1239–1244, 2012”.

Most neural speech enhancement systems use a fixed processing depth for all inputs, which can introduce unnecessary computation when fewer refinement steps are sufficient. We propose Adaptive Depth and Expert Refinement (ADER), a parameter-shared progressive enhancement framework with input-dependent computation. ADER combines an Adaptive Depth Controller (ADC) for hard early termination with a Conditional Expert Router (CER) that selects one lightweight residual adapter at each executed refinement iteration. We further introduce Exit-aware Intermediate Supervision (EIS) to directly optimize candidate intermediate outputs for early exit. On VCTK-DEMAND, ADER reduces the parameter count and average computation of MP-SENet by 70.4% and 51.3%, respectively, while achieving a WB-PESQ of 3.37. Overall, ADER enables input-dependent refinement and reduces redundant computation during inference.

We construct an exact Kerr-like rotating extension of the zero-redshift Morris–Thorne wormhole in a Teo-type stationary, axisymmetric ansatz. The Einstein equations for an anisotropic source comoving with zero angular momentum observers fix $\mathcal{K}(r)=r^2+a^2$ and reduce frame dragging to a radial quadrature. For $b(r)=r_0^2/r$, this is evaluated using incomplete elliptic integrals and normalised by the ADM angular momentum. The spacetime is asymptotically flat with zero ADM mass, so $a=J_{\rm ADM}/M_{\rm ADM}$ does not apply. The oblateness length $a$ and angular momentum $J_{\rm ADM}$ remain independent, yielding a family $(r_0,a,J_{\rm ADM})$. The choice $J_{\rm ADM}=ar_0$ selects only a one-parameter slice for numerical ergoregion and equatorial-capture illustrations; it is not a field-equation constraint. We find $r_{\rm th}=\sqrt{r_0^2-a^2}$, with reality bound $|a|\leq r_0$ and regular range $0\leq |a|<r_0$. In the canonical stationary foliation, the throat is defined quasi-locally as the unique closed surface $S_0$ in the family $S_\ell$ with zero mean curvature and positive area second variation. Moreover, $r=r_{\rm th}$ and $\ell=0$ are coordinate representations. A signed throat-adapted coordinate reveals two asymptotically flat ends, excludes closed timelike curves, and establishes that the throat is timelike, not a horizon. These conclusions hold independently of $J_{\rm ADM}=ar_0$. The supporting stress tensor, reconstructed from the Einstein tensor, is an effective phenomenological source. We do not claim any microscopic matter Lagrangian or realistic equation of state. We also determine the ergoregion onset. Since the full Hamilton–Jacobi equation is nonseparable, we limit ourselves to derive the optical result for a one-dimensional equatorial capture interval.

We present Hybrid Quantum Root Cause Analysis (HQ-RCA), an industrially grounded workflow for root cause analysis in banking IT operations, built on a hybrid Quantum Graph Neural Network (QGNN): the classical backbone of DynEdge (the IceCube neutrino-reconstruction GNN, which we call standalone DynEdge), with its classification head replaced by a Variational Quantum Circuit (VQC). On 13 months of anonymised IT data (13k alarm clusters) from a major European bank, the hybrid QGNN matches standalone DynEdge – the strongest classical baseline – on $F_1$, while standalone DynEdge leads the ranking metrics. A readout-sensitivity and layout-robustness study, analysed via Dimensional Expressivity Analysis (DEA), shows that the effective parameter dimensionality (rank) of the quantum observable has no measurable correlation with $F_1$; we therefore keep the simplest readout $\langle Z_0\rangle$ (the Pauli-$Z$ expectation on the first qubit), which in the deployed layout is rank-1, collapsing optimisation to a 1-D problem solvable by a gradient-free grid scan. Execution on IBM Heron r2 (no error mitigation) shows this gradient-free readout is executable on NISQ hardware after threshold recalibration.

Generally, quantum heat engines driven by Gibbs reservoirs achieve their maximum work and efficiency only in quasi-static processes, a constraint that hampers practical applications due to the resulting vanishing power output. To address this fundamental limitation, we propose a finite-time quantum Otto information engine (OIE) driven by a Maxwell's demon paired with a single Gibbs reservoir. We demonstrate that the demon's measurement and feedback control can preserve quantum coherence to extract coherent work, while harnessing quantum internal friction as a work source. Consequently, the OIE can produce work by only modulating the eigenstates of the Hamiltonian, and surpass the quasi-static limits of its conventional counterpart in the work output and the efficiency, which accounts for the energetic cost of the demon. Notably, the OIE can achieve near-perfect efficiency alongside positive work extraction in the rapid-driving regime.

Average squared error cannot reveal whether forecasting performance degrades because the future becomes less predictable or because forecasts move farther from the conditional mean. We introduce paired, mechanism-controlled stress tests that decompose changes in expected squared error at each lead time into environmental risk and forecast-oracle distance, using an origin-conditioned predictive oracle unavailable to the evaluated models. Three end-to-end controls have known attribution. Specifically, the null, environmental-only, and information-gap controls verify that the pipeline assigns changes to the correct component. We then apply the benchmark to 24 deployable forecasters. Under frequent switching, 14 methods have higher realized MSE but lower oracle distance; under outlier-variance feedback, 19 have higher MSE but lower scale-standardized MSE. Short- and long-lead stress-response rankings have Spearman correlation 0.624, revealing substantial horizon-dependent reordering. We then study multivariate relation shifts. Across six models and three coupling severities, oracle distance accounts for only 0.7-3.9% of the decomposed expected-risk increase, and environmental-risk majority persists in an eight-channel system and a matched-difficulty audit of Ring, Block, and Hub relations. Finally, prespecified contrasts on independent data-generating process (DGP) realizations show that several visually compelling discovery profiles, including trend accumulation and the hypothesized switching reversal, do not replicate. The benchmark thus combines component-wise diagnosis with a held-out stability audit. It complements real-data out-of-distribution evaluation, which measures performance under realistic shifts when exact oracle attribution is unavailable.

Counterfactual tool evaluation must distinguish authority, historical support, and what a comparison actually estimates. We study these distinctions with eleven executable enterprise-inspired tools, exact-propensity logs, and real local Model Context Protocol transport. An initial 45-run synthetic study is retained, then challenged by 30 realized-return control runs and 15 experiments on 1,930 independently released Berkeley Function Calling Leaderboard (BFCL) tasks. Full-return direct regression reverses an initially favorable doubly robust (DR) evaluation result in the linear setting: mean absolute errors are 0.0139 for direct regression and 0.0272 for DR. Under a shifted environment, DR retains an advantage, with errors 0.0227 versus 0.0948. On function-name-group-disjoint BFCL-derived splits, direct and DR selectors obtain balanced accuracies of 81.85% and 79.83%. Two pinned local Qwen2.5 models are evaluated on the same 200 held-out tasks, exposing a strong failure to abstain under the fixed prompt. We further characterize policy differences under missing support: unsupported actions shared by two policies cancel, allowing point identification of an incremental change when neither absolute value is identifiable. A disagreement-preserving fallback achieves this property in all five support-gap runs, but conservative sampling bounds do not certify deployment improvement. The contribution is a falsifiable evaluation method and independent public evidence, not a new DR estimator, official BFCL leaderboard score, or production-agent safety claim.

Memory-poisoning defenses for LLM agents are usually judged by how well they stop attacks, yet almost all the traffic an agent sees is benign. We measure what these defenses cost when no attack is present. Our harness holds the backbone model, memory backend, retrieval, answer prompt, and judge fixed and changes only the defense. Both LLM-based defenses run on the agent's own backbone, and an integrity check verifies every run. We test four defenses on half of LoCoMo: five long conversations (750 turns and 255 questions per run), with each condition run three times. Sanitization and provenance checking never fired. LLM-based anomaly detection blocked 9 of 2,250 turns (0.4%). A read-time reranker dropped 2.2% of the memories it judged, withholding at least one memory from 14.9% of questions. A blind audit found content-based justification for only 2 of 50 sampled drops. No defense showed a detectable accuracy cost, at 95% intervals of $\pm$2.4 to $\pm$4.6 points; the reranker changed accuracy by +0.6 points (95% CI [-2.6, +3.8]). Anomaly detection and the stacked condition gained 4.7 points, but a blocked turn changes the memories extracted after it, so we do not count this as a benefit. Token overhead was 0% for sanitization and provenance, 1.5% for anomaly detection, 3.0% for the reranker, and 4.5% with all four stacked. An earlier version of this study ran the reranker condition on a different model than its baseline. A baseline on that same model shows the swap produced 2.9 of the 4.4 points of accuracy cost we then reported. Attack success rate reveals none of this, so we report benign cost alongside it. We release the harness, labels, and per-run outputs.

We investigate the coupling dependence of ballistic energy transport and the emergent spin dynamics in an integrable Hermitian three-state chain that connects a clock interaction to a highly degenerate flag limit. By constructing a regular $R$-matrix to establish a globally conserved energy current, we analytically evaluate its full variance to obtain the exact, strictly positive leading high-temperature coefficient of the thermal Drude weight and the ballistic growth rate of the energy-correlation second moment. In the strong-coupling limit, the degeneracy is lifted by virtual transitions of a delocalized third-color spectator state, which generates an effective spin-$1/2$ XXZ Hamiltonian with anisotropy $\Delta = -1/2$ and fundamentally selects the all-active two-color sector as the true ground state. For periodic boundaries, this virtual spectator motion introduces a positive length-changing XXZ supercharge squared that, for $L\ge4$, strictly annihilates all states within a finite, length-independent energy interval above the ground state. Consequently, we rigorously prove that the full periodic effective theory perfectly replicates the exact low-energy XXZ spectrum, including all state multiplicities, as well as its macroscopic bulk free-energy density.

Reward-free latent world models can learn from offline videos and solve new image–goal tasks by optimizing actions through predicted latent futures. This setting places two demands on the planning state: its coordinates must be comparable with a goal image. Moreover, its dynamics must retain velocity, motion trend, contact, and other history–dependent information beyond those goal coordinates. Offline training creates a second mismatch: each recorded trajectory reveals one factual future, whereas a sampling–based planner compares many actions that were not taken from the same state. We introduce FIRM-WM (Factual–Interventional Recurrent World Model), a compact pixel world model designed around these two gaps. Its recurrent state separates a typed, goal–comparable configuration from a 128-dimensional dynamic fiber used for prediction but excluded from the terminal goal cost. Broad factual trajectories provide state coverage, while common–reset intervention branches provide observed outcomes for alternative action sequences. Before executing each branch, we reset the environment and restore the same recorded values exposed by the environment's state–setting interface. Under matched CEM planning and three independent full-pipeline seeds, FIRM-WM reaches 99.0$\pm$1.0% on TwoRoom, 92.7$\pm$2.1% on Reacher, and 88.0$\pm$3.0% on OGBench-Cube, compared with 89.0%, 88.0%, and 70.0% for LeWM. The deployed model uses 2.98–3.42M parameters and records 2.13–11.60$\times$ lower planning time on these tasks.