Papers

We present a simple model of a Bloch line in a thin-film ferromagnet with strong easy-axis and intermediate in-plane anisotropies. The meron is represented as a charged, massless particle that slides along the linear domain wall in a background magnetic field- a gyroscopic bead on a string. Its gyroscopic behavior dynamically isolates the wall segments on either side of the meron, motivating the model for a Bloch line as two strings joined at a moving boundary. The meron clamps the wall at this junction, completely reflecting low-amplitude spin waves that propagate along it. Out-of-plane and in-plane magnetic fields and spin-transfer torques then act in two ways: by exerting a force on the Bloch line itself, or by driving the wall as a whole. Predictions of this model are in agreement with numerical simulations.

As large language models (LLMs) become increasingly capable of code generation, adopting generated code in software development requires assessing not only its functional correctness but also its maintainability-related quality. If such quality could be estimated before generation, developers could avoid the cost of generating, reviewing, and discarding low-quality code. Although prior work has shown that the functional correctness of the LLM-generated code can be predicted in advance, it remains unclear whether maintainability-related quality is similarly predictable. We introduce Pre-Generation Maintainability-Related Quality Prediction (PreMaQ), which predicts the Code Smell Score (CSS) and Maintainability Index (MI) of generated code from the internal representations of LLMs before generation. Our evaluation covers four open-weight LLMs and four Python code generation benchmarks, comprising 2,695 tasks in total. Our results show that predicted CSS and MI consistently correlate with their observed values across all 16 model-benchmark combinations, achieving mean Spearman rank correlations of 0.57 and 0.65, respectively. When used for model selection, PreMaQ achieves 59.5% of the maintainability-related quality improvement attainable by an ideal maintainability-based selector over random selection on tasks for which multiple models generate functionally correct code. Combining predictions from PreMaQ and prompt embeddings increases this proportion to 60.7%, indicating that the two signals are complementary. These findings suggest that PreMaQ can be used to predict and improve the maintainability-related quality of LLM-generated code.

Stochasticity shapes evolutionary dynamics in finite populations beyond the reach of infinite-population theories. Using a stochastic birth–death framework on random regular networks, we establish evolutionary stability criteria for finite structured populations playing three-strategy games and identify critical degree thresholds governing invasion dynamics. Our framework goes beyond two-strategy evolutionary games to reveal that the initial composition of multi-type mutants, together with population structure and stochasticity, shapes the evolutionary fate of the resident population.

We study kernel functions induced by a linear time-invariant (LTI) stochastic differential equation (SDE), an approach that embeds prior knowledge about the function to be estimated directly into the state-space description of the SDE. This kernel-design perspective is relevant across the many settings where kernels play a central role, including kernel methods in statistics, Gaussian process (GP) regression in machine learning, and kernel-based regularization methods in system identification. The resulting class of SDE-induced kernels includes several kernels of practical interest, such as Matérn kernels with half-integer smoothness, spline kernels, and related kernels used in system identification. In contrast to existing linear-time GP regression methods for these kernels, which are typically based on Kalman filtering and Rauch–Tung–Striebel smoothing on the latent state, we work directly with the SSS representation of the observed covariance matrix. This yields linear-complexity recursive algorithms for the matrix operations required in GP regression, including factorization, linear solves, log determinants, posterior covariance, log-marginal-likelihood gradients, and posterior sampling; the same generator-level recursions extend, with only minor modifications, to the analogous computations in kernel-based regularization methods for system identification. Compared with previous uses of SSS representations in GP regression, we derive efficient recursive algorithms for analytic marginal-likelihood gradients in three application-relevant cases, and we develop new recursions for posterior computations and Matheron-type sampling. Numerical experiments verify the accuracy of the proposed recursions and illustrate posterior inference and sampling for representative SDE-induced kernels.

A bipedal robot cannot deviate from its path to avoid an obstacle without disturbing its balance, and this coupling is most severe on underactuated platforms such as the biped considered here, which has four actuated joints per leg and no hip or ankle roll. This paper presents a Hierarchical Reinforcement Learning (HRL) framework in which a High-Level (HL) policy observes the robot pose, 36 raycast proximity measurements, moving-obstacle states, and a receding-horizon local goal, and outputs a body-velocity command $(v_x, v_y, \omega_{yaw})$ every ten control steps, while a velocity-conditioned Low-Level (LL) policy tracks each command through PD-controlled joint targets. Both policies are trained jointly with Soft Actor-Critic (SAC). Because the converged gait is task-agnostic, it is frozen and driven by classical planners over the same command interface, yielding three controlled baselines: SAC+A*, SAC+RRT*, and SAC+APF. Across 100 evaluation trials per method in randomized PyBullet environments, the proposed method reaches the goal in 98.0% of static and 88.0% of dynamic trials, against at most 78.0% and 68.0% for the planner hybrids, with path lengths within 4% of the A* reference, and ablations confirm that each observation channel and reward term contributes materially to this performance.

Triangular attention is a core computation in AlphaFold3-style biomolecular models, with cubic cost in token count. Its shared pair bias adds a gradient reduction across attention slices to the usual reductions over queries and keys. The open-source backends we examine handle these reductions through repeated probability recomputation, floating-point atomics, or full score-gradient storage. Separately, computing the softmax backward correction from BF16-rounded forward outputs loses numerical precision. We present TurboPairFormer, a triangular attention implementation for NVIDIA Hopper GPUs that addresses both issues. Our key-tile-parallel backward algorithm recomputes each probability tile once for the query, key, value, and pair-bias gradients, using ordered partial reductions for deterministic accumulation without floating-point atomics or full score-gradient storage. Output-residual compensation retains a BF16 approximation of the output-rounding residual to compute the backward correction more accurately in FP32, without changing the BF16 output. With BF16 inputs at crop sizes 384, 640, and 768 and head dimensions 16 and 32, TurboPairFormer achieves the lowest mean query, key, and pair-bias gradient RMSE against an FP64 reference among the implementations compared in this paper. Residual compensation reduces these RMSE values by 28-47% in controlled ablations. All four gradients are bitwise identical across five repeated calls in all 600 input cases under fixed execution conditions. Integrated into OpenFold3 with our triangle multiplication kernels, TurboPairFormer achieves the lowest GPU computation time per optimizer step among the evaluated backend configurations on 16 H100 GPUs, with speedups of $1.73\times$ over OpenFold3's Triton backend and $1.13\times$ over cuEquivariance at crop size 768.

Monitoring concept drift from an adaptive classifier's error stream creates an operational conflict with the model's own update loop. When internal adaptation outpaces evidence accumulation, accuracy recovers before cumulative detectors (CUSUM, Page-Hinkley) can reach threshold. Instrumenting an Adaptive Random Forest (ARF) shows that surviving trees absorb 98.6% of the post-drift error transient through incremental leaf updates alone. The first background tree swap accounts for just 0.71% of this erased error volume, but drops external detection rates by 31 percentage points. We derive the finite-horizon boundary where cumulative evidence fails to cross threshold and measure a critical magnitude floor ($\Delta e_c = 0.120$) below which false-alarm budgets preclude detection. This failure manifests as missed shifts on stationary streams and false-alarm flooding triggered by internal tree swaps on noisy baselines. We validate on synthetic shifts, ARMA-GARCH series (ProteuS), and tabular benchmarks (BAF, INSECTS); on the synthetic sweep at a standard threshold, the blind spot appears at $\Delta e \approx 0.25$, showing why classical benchmarks like SEA ($\Delta e \le 0.21$) failed to reach it.

Frozen representations are widely reused for downstream classification, yet each new task typically requires fitting a new predictor. We ask whether the few-shot prediction procedure itself can instead be learned once and reused across datasets and representation spaces. To study this question, we introduce RepShiftBench, comprising 1,218 encoder–dataset tasks across text, image, and audio, with separate evaluation of generalization to unseen datasets, unseen encoders, jointly unseen datasets and encoders, and unseen modalities. The benchmark exposes a substantial gap: Logistic Regression fitted independently on each episode outperforms every evaluated in-context learner across all settings. We introduce RepICL, a meta-trained in-context learner that canonicalizes each episode through episodic whitening before prediction. Its inductive variant, RepICL-I, surpasses Logistic Regression in all 12 benchmark settings, while RepICL-T substantially outperforms existing transductive methods. Ablations identify episodic whitening as the primary source of these gains, while showing that it is not a universally beneficial preprocessing step. Across both variants, the gains concentrate on queries for which simple support prototypes favor the wrong class or provide little separation between the true class and competing classes. Transduction provides its largest additional gains when limited support coverage gives a misleading view of class separation. Together, these results demonstrate that a shared few-shot prediction procedure can generalize beyond the representation spaces observed during training.

In this work, we propose a new natural dark matter candidate: Kaluza-Klein towers of vectors from $d\geq 2$ extra-dimensional models. Their naturalness follows from two key features: (i) they arise unavoidably from the Kaluza-Klein decomposition of the higher-dimensional metric, and (ii) if the Standard Model is localised on a brane, they do not couple directly to it and interact only gravitationally through the other components of the higher-dimensional metric. Remarkably, the rich particle content comes with as few as a single extra parameter. We first explain the general idea for $d\geq 2$, then focus on the $d=2$ case as our benchmark scenario. We derive the field content and interactions among the model particle content, and show how the geometry of the extra dimension unitarises the relevant processes for cosmology. We then compute the multi-component relic abundances of the vectors from freeze-in and discuss the available parameter space.

Let $f$ be a regular endomorphism of $\mathbb{A}_{\mathbb{C}}^N$ of algebraic degree $d$ and let $f_{\infty}$ be the induced endomorphism of the infinity hyperplane $H_{\infty}$. Suppose that for every periodic point $x_0\in H_{\infty}(\mathbb{C})$ of period $n_0$, the geometric mean of the multiplicities $e_{f_{\infty}}(x_0),\dots,e_{f_{\infty}}(f_{\infty}^{n_0-1}(x_0))$ is strictly less than $d$. Then we show that the dynamical Mordell–Lang conjecture for $f$ holds for curves in $\mathbb{A}_{\mathbb{C}}^N$.

Radiation-induced composition redistribution at grain boundaries (GBs) in Ni-based structural alloys poses a critical challenge for the long-term deployment of advanced nuclear reactors. The mechanisms governing solute-specific segregation remain poorly understood because defect-mediated transport, thermodynamic driving forces, and competing microstructural sinks can all contribute under irradiation. We perform systematic Ni2+ ion irradiations on dilute Ni-Fe, Ni-Cr, and Ni-Mn binary alloys at 500, 800, and 1000 K to 2 dpa and characterize GB chemistry using STEM-EDS. The segregation behavior is interpreted using rate-theory phase-field modeling informed by atomistic transport data. We find that Fe depletes at GBs across all temperatures via preferential vacancy-solute exchange (inverse Kirkendall effect). Cr segregation reflects a temperature-dependent competition between self-interstitial atom (SIA)-mediated transport toward GBs and vacancy-mediated transport away from GBs, yielding near-zero net segregation at 500 K and depletion at higher temperatures. In contrast, Mn exhibits distinct behavior dominated by thermodynamic segregation that drives local enrichment at the GB. We leverage depth-dependent characterization along the same GB, spanning irradiated and unirradiated regions, to separate radiation-induced and thermodynamic contributions and validate the unirradiated region as an internal thermodynamic reference. Void formation at 800 K and irradiation-induced GB migration at 800 and 1000 K further add mechanistic complexity by suppressing and redistributing the solute. Our combined experimental and computational assessments show that interpreting GB chemistry evolution under irradiation requires concurrent treatment of non-equilibrium kinetic transport, thermodynamic driving forces, bulk sink effects, and GB migration.

We prove rapid mixing of a quantum Gibbs sampler for general one-dimensional non-commuting local spin systems at every fixed positive temperature. In particular, for a system of $n$ sites, we show that the Kubo–Martin–Schwinger (KMS) detailed-balanced Gibbs sampler of~\cite{CKG2023} converges from any initial state to trace-norm error $\epsilon$ in time $O(\log^4(n)+\log(1/\epsilon))$. This improves the previous $O(n+\log(1/\epsilon))$ bound and establishes rapid mixing beyond high-temperature and weak-interaction regimes. Our proof explains why the slowest decay rate, captured by the spectral gap, can substantially underestimate the speed of equilibration. The key insight is that a state far from equilibrium must have a substantial component outside the span of low-weight Pauli operators, where stronger dissipation estimates apply. This yields two stages of relaxation: a double-exponential convergence from a bad initial state, followed by exponential convergence near equilibrium. We turn this mechanism into a framework for proving rapid mixing through spectral analysis beyond the gap. Our decay estimate is equivalent to exponential convergence in sandwiched Rényi-2 divergence. Together with the spectral gap, it also implies a modified logarithmic Sobolev inequality with rate $\Omega(\log^{-4}(n))$.

We classify reductive monoids over arbitrary base schemes by descent from the split classification in arXiv:2607.00322. For a reductive group scheme $G$, we construct its sheaf of based root data, abstract Cartan torus, abstract Weyl group, and a scheme of Weyl cones. The last of these represents the functor of isomorphism classes of rigidified reductive monoids with unit group $G$. As applications, we deduce a criterion for the existence of monoids that are not groups, prove an extension theorem over normal integral bases, and construct the Vinberg monoids without a quasi-split hypothesis.

A Banach space has the ball-covering property (BCP) if its unit sphere can be covered by countably many balls that miss the origin. A Banach space has the renorming-stable ball-covering property (RBCP) if it has the BCP with respect to every equivalent norm. We investigate the RBCP from the perspective of $C(K)$-spaces. In particular, we show that there are many non-separable $C(K)$-spaces having the RBCP. This yields an emphatic negative answer to a question of Cheng, Kato and Zhang from 2020.

Diffusion models generate samples by learning to reverse a fixed corruption process, and classifier-free guidance (CFG) is the standard mechanism for conditioning this process on a desired class or prompt. CFG can be applied at varying guidance strengths, and while higher strengths improve image quality and conditional alignment, too high a guidance strength can degrade image quality and diversity. Furthermore, CFG violates principled diffusion sampling dynamics, and existing explanations for why it works despite the violation disagree on the underlying theory or do not extend to deterministic samplers used in practice. We address both these issues. We first frame CFG sampling as a continuous-time optimal control problem, treating the sampling trajectory as a sequence of controls chosen to maximise the probability of the desired condition. Solving the resulting Hamilton–Jacobi–Bellman equation shows that CFG is recovered under specific path costs when using an unconstrained control set. We argue this lack of constraint is responsible for CFG's failure at high guidance strengths, since it permits the sampling path to move arbitrarily far from the current image estimate. To fix this, we propose curvature-aware CFG (CACFG), which constrains the control set to a hypersphere informed by the Gaussian regularisation used when training variational autoencoders. We show that the control inputs produced by CFG sampling routinely violate this bound, and that across diffusion models, datasets, and guidance schedules, CACFG achieves superior generative quality at mid-to-high guidance strengths with a less severe quality-diversity tradeoff than regular CFG.

Recovering complete phase sets from powder X-ray diffraction (PXRD) is challenging when weak-phase peaks overlap stronger signals. A natural strategy is to identify phases iteratively, removing the contribution of each identified phase from the observed pattern before predicting the next. However, even after a phase is correctly identified, misestimating its contribution can distort the residual and cause subsequent errors. We introduce PhaseMatcher, an autoregressive framework for complete phase-set identification with physics-guided spectral decomposition. After each phase prediction, PhaseMatcher re-estimates the contributions of all selected phases and the residual from the original observation and all selected reference patterns, accounting for physically plausible variation between reference patterns and the corresponding phase contributions in the observation. The resulting residual guides subsequent phase identification, while a separate stopping module determines when the phase set is complete. On synthetic mixtures and controlled mixtures constructed from measured single-phase patterns, PhaseMatcher improves complete-set identification over the evaluated baselines. On PhaseMix-135K, it also estimates contributions and residuals more accurately than scalar subtraction.

Predicting metal-organic framework (MOF) structures from given building blocks requires recovering their positions and orientations in a periodic crystal. The spatial effects of rotation errors are geometry-dependent and anisotropic. The same angular error can produce different atomic displacements depending on block size, shape, and rotation axis. Angular error alone, without reference to the specific block geometry, therefore cannot fully describe the spatial consequences of a pose error. We introduce AnchorPose, a meso-grained pose generation framework that incorporates this geometric dependence into its generative representation. It represents each block through a small set of representative atoms, combines their local geometry with the current spatial state, and generates their coordinates with Bayesian Flow Networks. Known atom correspondences enable rigid alignment to recover complete building-block poses and return geometrically consistent points to the generation process. This design connects point-level spatial prediction with block-level structural constraints. Geometry participates in the pose state and its prediction, while rigid reconstruction preserves intra-block structure without treating all atomic coordinates as assembly variables. On the MOF benchmark, AnchorPose improves single-candidate match rates over the compared block-level and all-atom baselines.

Linear attention enables efficient long-context autoregressive decoding by compressing history into recurrent states, but this compression can make selective access to sparse and distant information difficult. Existing chunk-based extensions increase memory capacity, yet learned chunk-mixing coefficients may remain fixed with respect to input content and therefore cannot adapt historical access to each query. We introduce Hybrid Linear Attention (HLA), a query-dependent chunk-level attention mechanism for Gated DeltaNet (GDN). HLA represents each completed chunk as an exact affine state transition and computes content-dependent routing gates from compact, self-attentively pooled representatives. Each gate interpolates the corresponding historical transition with the identity map, controlling both the chunk's additive memory and its transformation of earlier states. Effective-support regularization further encourages concentrated routing for sparse inference. We evaluate HLA under both pretrained adaptation and from-scratch training. Across Qwen3.5 models from 0.8B to 9B, HLA consistently improves over native GDN and fixed chunk mixing, with gains of up to 5.57 percentage points on LongBench-V2 and 3.97 points on RULER. In a controlled from-scratch 1.3B setting trained for 100B tokens with a 4K context, HLA also improves RULER performance from 4K to 32K, with gains increasing from 0.83 points at 4K to 4.22 points at 32K. These results demonstrate that query-dependent composition of recurrent memory improves long-context modeling and remains effective beyond the training context while using compact per-chunk affine summaries. Project page: https://caesarhhh.github.io/hla/

In this paper, we investigate nonexistence results for stable solutions of the semilinear elliptic problem $$(-\Delta)^s u+\lambda|x|^\alpha u=|x|^\beta |u|^{p-1}u \; \mbox{in}\; \mathcal{H}, $$ where $\mathcal{H}=\mathbb{R}^n$ or $\mathcal{H}=\mathbb{R}^n_+=\{x=(x',x_n),\, x'\in\mathbb{R}^{n-1},x_n>0\}$ , with Dirichlet conditions in the half-space case. We assume that $n\geq2s$, $p>1$, $\lambda>0,$ $\alpha, \beta>-2s$ and $0<s\leq1$. Our aim is to analyze the influence of the Hardy-type power potential $\lambda|x|^\alpha u$ in the presence of the Hénon weight $|x|^\beta$, with particular emphasis on the existence and classification of weak stable solutions, including solutions that may be unbounded or sign-changing. We establish nonexistence results for nontrivial stable solutions and, more generally, for solutions that are stable outside a compact set, covering critical and supercritical regimes without additional assumptions, and the subcritical regime under a suitable $L ^\infty$-smallness condition on a weighted version of the solution. Our results apply to both the local case $s = 1$ and the nonlocal case $0 < s < 1 $, in the whole space and in the half-space. The proofs are based on energy estimates combined with a Pohozaev identity. In the supercritical regime, we further establish a monotonicity formula, following the approach of Dávila et al. (Trans. Amer. Math. Soc. 369, 6087-6104, 2017).

A tool-using model can follow a malicious instruction even when its credentials are valid. We study whether task-scoped authorization contains the resulting tool execution. Our paired-replay testbed samples a model request once and submits the same action, resource, and arguments to broad bearer, scoped JWT, sender-constrained, and Open Policy Agent conditions. The frozen main experiment uses 128 scenarios across four tool domains and five local model configurations. Among valid attacked post-exposure decisions, broad-bearer harmful execution ranges from 8.9% to 37.8% across models; all three scoped conditions record zero. The policy changes execution, not the frozen model decision. These results support containment of the tested cross-action and cross-resource consequences under researcher-supplied task authority, not prevention of prompt injection. A bounded six-task AgentDojo extension preserves native scoring and records 11/24 injected broad-policy attack flags versus 0/24 scoped flags, with utility of 5/24 and 6/24. Low exposure, invalid decisions, and purposive task selection limit that comparison. The primary experiment measures safe continuation rather than final-answer correctness. An empirical-bank fresh-sampling comparison finds no estimation advantage from pairing when scoped outcomes are constant zero. The contribution is a controlled measurement of compromise, enforcement, and continuation, with explicit limits on what each observation establishes.

Robotic ultrasound couples a high-rate contact-force loop with slower, delayed image feedback, so image-guided ultrasound probe rotation can perturb contact force before the resulting image response is observed. We derive a closed-form orientation-rate limit that bounds the modeled rotation-induced estimated-force excursion over a finite horizon while accounting for disturbance rejection by the fast force loop. The limit depends on local contact stiffness, force-loop gains, a conservative rotation-to-force gain bound, the excursion budget, and the prediction horizon. We implement this model in a dual-rate controller with timestamp-based delay reconstruction and joint-torque-based force estimation, and evaluate it on a curved gelatin phantom using paired controller comparisons and component ablations. Relative to unconstrained image guidance, the proposed rate-limited controller reduced first-second root-mean-square (RMS) estimated-force error by 0.40 N while increasing cue-convergence time by 0.94 s. A fixed rate cap near the analytically predicted ceiling produced no resolvable difference in force error and converged 0.32 s faster, indicating that the principal practical value of the model is the rate-design rule rather than online prediction. Delay reconstruction had no resolvable effect at the tested latency. A single-subject popliteal scan demonstrated feasibility, although the image cue was noise-limited on heterogeneous tissue.

A $k$-bend path is a non-self-intersecting polyline that lies on a grid and consists of at most $k+1$ axis-parallel line segments. A $B_k$-CPG graph is a graph whose vertices can be represented by pairwise interiorly disjoint $k$-bend paths on a grid such that two vertices are adjacent if and only if the corresponding grid paths touch at a grid point. We prove that recognizing planar $B_0$-CPG graphs of maximum degree 8 is NP-complete, and that recognizing planar $B_1$-CPG graphs of maximum degree 11 is NP-complete. These results settle two of the three planar recognition problems left open by Champseix, Galby, Munaro, and Ries.

We establish large and moderate deviation principles for a broad class of nonlinear stochastic partial differential equations driven by multiplicative Poisson random measures in the small-noise regime. Our abstract framework is formulated for locally monotone evolution equations over a Hilbert triple and encompasses several important models arising in fluid dynamics and turbulence theory, including the two-dimensional Navier–Stokes equations, magnetohydrodynamics (MHD), the GOY shell model of turbulence, the two-dimensional tidal equations, and Navier–Stokes equations with nonlinear viscosities. A key analytical ingredient is the well-posedness of the associated controlled equations, established using the method of local monotonicity. The large and moderate deviation principles are then proved via the weak convergence approach.

Biological membranes are highly deformable structures that undergo shape transformations during numerous cellular processes. These transformations are often driven by active agents, including molecular motors, cytoskeletal machinery, and motile microorganisms, which consume energy and maintain biological systems far from equilibrium. Here, we investigate how a single active particle confined within a lipid vesicle drives membrane deformation. Using a lipid-resolved coarse-grained model, we examine how the activity and shape of an encapsulated self-propelled particle influence vesicle dynamics and morphology. Active particles generate persistent propulsion that exerts mechanical stresses on the enclosing membrane, thereby driving its deformation. Using particles of different shapes but identical surface area, we systematically investigate membrane morphologies as functions of activity and bending rigidity. We quantify membrane deformation through asphericity and excess bending energy relative to the equilibrium spherical state. Both measures exhibit a non-monotonic dependence on activity, peaking at intermediate activity. At sufficiently high activity, membrane rupture occurs before significant large-scale deformation can develop, limiting further remodeling. We further demonstrate that membrane surface tension plays a significant role in regulating activity-induced deformation dynamics.

Tactile Internet (TI) security analytics must balance reliable thresholded decisions with constrained computational and measurement resources. We study this tension for finite-shot hybrid quantum anomaly inference and introduce the Adaptive-Shot Variational Quantum Circuit (AS-VQC) policy. This validation-calibrated policy begins each record at 128 shots and cumulatively escalates through 256, 512, and 1024 shots only when the finite-shot anomaly score remains close to a validation-selected security threshold. The quantum scorer is evaluated as an off-path security analytics component rather than part of the haptic critical path. Using a 4,875-record CESNET-TimeSeries24-derived aggregate-flow benchmark, leakage-safe random, entity-group-disjoint, and temporal holdouts, and five trained quantum neural network (QNN) checkpoints per holdout, the primary AS-VQC-95 (beta = 0.95) policy averages 129.2, 276.9, and 131.2 shots per record, saving 87.4%, 73.0%, and 87.2% of the uniform 1024-shot baseline (Fixed-1024), respectively. The decision disagreement with analytic (exact-expectation) inference is 0.771%, 0.409%, and 0.635%, lower than both the uniform 128-shot baseline (Fixed-128) and a matched-budget shuffled-allocation control. Fixed-1024 remains more decision-stable, establishing a measurable reliability-resource trade-off rather than cost-free equivalence. A more conservative AS-VQC-99 (beta = 0.99) further reduces disagreement while using fewer than 512 average shots across all holdouts. These results show that finite quantum measurements can be treated as an inference resource and concentrated on boundary-sensitive TI-security decisions while exposing checkpoint-dependent escalation under unseen-entity conditions.

Broad Learning System (BLS) supports analytical training and incremental expansion, but its growth needs guidance on which inputs new blocks should learn from. Weight errors pose a further challenge by displacing learned outputs across class boundaries. We propose HiER-BLS to couple hierarchy-guided representation growth with error-correcting learning. Successive blocks focus on inputs selected by feature importance and correlation while preserving earlier representations. The evolving branch guides encoded learners through subspace size and sample confidence, so its learning experience informs both their feature views and supervision. For finite broad readouts, we show how codeword correlations transform fitted class scores. Prediction preservation depends on the distance from the actual output to the nearest decoding boundary relative to the model's sensitivity to weight errors. Experiments on five image and five tabular datasets demonstrate improved classification performance over representative BLS variants. Component studies show that hierarchy guidance benefits the encoded branch even when the guiding branch has lower standalone accuracy, with further gains from combining their scores. Longer codes continue to improve accuracy under stronger Gaussian weight errors after clean accuracy has largely saturated.

Long-running agents repeatedly call an LLM while retaining most of their document window, evicting old documents, and appending new ones. These rolling updates break exact prefix caching and motivate non-prefix KV-cache reuse with selective recomputation. We show that persistent KV-cache reuse with selective recomputation can be history-dependent: in our rolling-agent workload, an unchanged prompt can produce different answers depending on the requests processed before it. At a matched 5% recomputation budget, document-aligned recomputation reduces answer variation across request orders from 69.0% with CacheBlend's token top-$k$ policy to 26.1%. When each prompt is evaluated after a different sequence of preceding requests, document-aligned recomputation improves fidelity to full prefill by 34.5-52.5 percentage points over token top-$k$, while both policies achieve approximately 5.7$\times$ median TTFT speedup. Our ablation study shows that, in our rolling-agent workload, contiguity is the main factor associated with robust selective recomputation.

In decentralised priority coordination, agents announce priority levels and a shared resource serves them in decreasing order, as at an unsignalised intersection; the levels form the decision layer of a hierarchical controller. Such interactions are routinely replaced by a potential game, i.e.\ by a common objective, for analysis and design. This paper determines what that surrogate misses, using the Hodge decomposition of the incentives into a potential component, which a common objective can represent, and a harmonic component, which it cannot. For the linear payoff, both components are obtained in closed form on every conflict graph and for every deterministic tie-breaking protocol: in common units, the harmonic energy is the number of conflicts and the potential energy adds the number of adjacent pairs of conflicts. Consequently, for every rationality parameter, the best common-objective model of the agents' choice log-odds, weighted uniformly over unilateral moves, has a relative squared error of at least $1/(d_{\max}+1)$, where $d_{\max}$ is the largest number of conflicts of one agent; for an eight-vehicle intersection it is exactly one fifth, for any number of priority levels. Invisible to strict-improvement dynamics, the missed component is, under low-rationality log-linear learning with uniform revision and to leading order, the stationary probability current, and its energy sets the entropy-production rate. Payoff design cannot remove it: on the complete conflict graph of $N$ agents, under a total-order protocol and with at least three priority levels, every nonconstant rank-based payoff leaves a relative error of at least $1/N$, with equality exactly for affine payoffs.

Terminal agents are commonly trained by imitating long teacher trajectories, yet how much of each trajectory to supervise remains unexplored. We study the supervision horizon, the number of trajectory tokens retained for training, and show that it is a key design axis for reliability and cost. Reliability improves with longer horizons but saturates: on Terminal-Bench, a 12K-token horizon solves more tasks than 16K ($29\pm0.7$ vs.\ $26\pm0.8$) while requiring 30% less training time. The horizon also shapes agent behavior: short horizons cause premature termination, intermediate horizons yield productive error recovery, and long horizons induce over-persistence. We analyze this saturation through a bias–complexity bound, in which longer supervision reduces temporal supervision bias but increases finite-sample estimation error from more heterogeneous late-stage histories. Guided by this analysis, we propose selective long-horizon refinement, which first trains on short prefixes and then refines only on continuations that are most likely under the warm-start model. It consistently outperforms full long-horizon training. At 16K, it raises successful attempts from $110\pm2.7$ to $126\pm2.1$ and tasks solved in at least six of eight attempts from $9\pm0.7$ to $14\pm0.6$; with half of the long-horizon data, it still reaches $122\pm2.4$ while cutting training time by 23%. The gains transfer across benchmarks, from $64\pm2.6$ to $73\pm2.1$ on Terminal-Bench v2.0 and from $137\pm2.7$ to $155\pm2.2$ on OpenThoughts-TBLite. For long-horizon supervision, selecting the right trajectories matters more than training on all of them.

Agentic leaderboards increasingly evaluate systems on public benchmarks whose task statements and solution-bearing artifacts can remain accessible. We propose a measurement-first audit framework that distinguishes contamination claims according to the evidence required to support them. It separates three channels that require different evidence: training-time exposure, evaluation-time retrieval, and pipeline/scaffold leakage. Each channel is coded as open, partial, closed, or unknown under a fail-closed rule. Across nine Holistic Agent Leaderboard (HAL) configurations, none of the 27 channel assessments was coded closed, but incidents were confirmed in four configurations. We then apply the behavioral component of the framework to a reported file-localization gap on SWE-bench Verified, using an outcome-blind, same-repository matched-control design with symmetric prompt-leakage screening, paired and repository-aware uncertainty analyses, and scorer validation, evaluated on GPT-4.1 and DeepSeek-V4-Flash. Among the 100 pairs retained after symmetric screening and the pair-integrity exclusion, GPT-4.1 showed a $+10.0$-point pair-weighted Top-3 benchmark-associated gap, but the 95% intervals from both the prespecified paired-bootstrap procedure and the post-hoc repository-balanced analysis included zero, leaving the benchmark-associated gap inconclusive. The reproduction scorer did not pass its validation gate: against consensus human labels, sufficient scorer sensitivity could not be established for either model, and both DeepSeek-V4-Flash firings on correct-gold comparisons were false positives. Without provenance evidence, appropriate controls, symmetric leakage screening, and validated scorers, stronger contamination claims are not warranted. The results do not establish training-data membership, contamination prevalence, or benchmark-induced score inflation.