Papers

This paper focuses on relay-assisted topology optimization to accelerate the convergence of distributed consensus algorithms over weakly connected networks. Instead of permanently adding a fixed set of relay links, we introduce a time-sharing framework where multiple relay configurations are activated in a probabilistic manner. An ActiveSet-based algorithm incrementally constructs candidate relay edge sets and jointly optimizes the mixing matrices and the associated relay set activation probabilities, allowing adaptive relay selection in the optimization process. Simulation results demonstrate a substantially improved performance-cost trade-off compared with fixed-cardinality relay selection strategies.

This paper investigates whether regenerative impulsive braking can enhance passive energy harvesting in harmonically forced oscillators. An impulsive braking event directly extracts kinetic energy while simultaneously inducing a transient that alters subsequent passive harvesting. The analysis shows that suitably timed braking can produce a net energy gain over passive harvesting alone. The braking-induced transient is derived analytically and used to obtain an exact expression for the harvested-energy gain over an arbitrary finite post-braking horizon, together with conditions for positive gain and the optimal braking instant. The corresponding infinite-horizon analysis shows that optimal braking occurs at zero crossings of the harmonic excitation, with complete braking maximizing the gain. Numerical results further show that the finite-horizon harvested-energy gain can exceed its infinite-horizon value. Simulations using short-duration high-gain braking closely match analysis, supporting potential of the proposed approach for energy-harvesting applications.

Let $(M,g)$ be a compact $d$-dimensional Riemannian manifold without boundary, and let $\{e_j\}_{j=0}^\infty$ be an orthonormal basis of Laplace eigenfunctions with frequencies $\{\lambda_j\}_{j=0}^\infty$. For a domain $\Omega\subset M$, consider $$ N_\Omega(\lambda) = \sum_{\lambda_j\leq\lambda} \left|\int_\Omega e_j\,dV_g\right|^2. $$ Suppose that $\partial\Omega$ is $s$-Ahlfors regular, with $s\in{[d-1,d)}$, and that $\Omega$ satisfies a two-sided corkscrew condition. We prove that $$ N_\Omega(\lambda) = \operatorname{vol}(\Omega)+O\bigl(\lambda^{s-d}\bigr), $$ and that this remainder is sharp. We further identify the geometric quantity governing the second term: an exact asymptotic of order $\lambda^{s-d}$ holds precisely when the boundary admits a natural $s$-dimensional crossing volume $\mathcal V^s(\partial\Omega)$, in which case $$ N_\Omega(\lambda) = \operatorname{vol}(\Omega) - C_{d,s}\, \mathcal V^s(\partial\Omega)\, \lambda^{s-d} + o\bigl(\lambda^{s-d}\bigr). $$ For admissible self-similar boundaries, the crossing volume exists in the non-lattice case, and hence the above two-term asymptotic holds. In the lattice case, the second-order behavior is instead governed by a log-periodic profile.

Path following requires an agent to converge to and traverse a geometric path without a prescribed timing law. Guiding vector fields (GVFs) address this directly, but as guidance laws they offer no mechanism to compensate for the agent's own dynamics. Differential flatness-based control (DFBC), on the other hand, compensates for the agent's nonlinear dynamics but relies on a time-parameterized reference trajectory. This letter presents a unified algorithm, which we term flatness-based geometric control (FGC), that generates the DFBC reference hierarchy directly from a GVF and its higher-order time derivatives, eliminating the need for a pre-planned time-parameterized trajectory while retaining full dynamics-informed control. A cascade Lyapunov argument establishes exponential convergence of the tracking error and the physical trajectory to the desired path. The framework is further extended to singularity-free GVFs, which removes the topological obstruction and guarantees the condition for global exponential convergence on closed or self-intersecting paths. The proposed algorithm is validated in simulation.

We establish a sharp comparison principle for the sharp-interface Ohta–Kawasaki energy and its Yukawa-screened counterpart in dimensions $2\le d\le 7$, for $m=0$, and under periodic and zero-flux Neumann boundary conditions. Whenever the finite optimal stripe width is compatible with the domain, we prove global minimality of periodic lamellae and characterize all minimizers. The proof combines radial comparison kernels, a sharp Fourier bound, and rigidity for a directional translation inequality in $BV$. The kernels are positive measures up to dimension five; in dimensions six and seven, distributional lifting and monotonicity of the averaged translation defect yield the same sharp comparison. In the Appendix, we also prove that no finite positive measure with the prescribed Fourier multiplier exists in dimensions $d\ge6$, and explain why the argument does not directly extend to $d \geq 8$.

We introduce a continuous-time counterpart of the median filter that gradually denoises a given image. In the limit of vanishing stencil size, our results show that the evolution converges to level-set mean curvature flow. Surprisingly, and for the first time for any median-type filter scheme, we can prove the convergence of energies in this limit. Such strong convergence results are of major interest as they often have to be assumed in the literature to prove convergence of related schemes. Moreover, this result carries crucial geometric information, namely the unit multiplicity of interfaces. The proof relies on a compensated compactness argument inspired by the work of Evans and Spruck.

Address translation is a major bottleneck in data-intensive workloads. TLB prefetching can hide translation latency, but existing spatial prefetchers struggle with irregular accesses, while temporal prefetchers store address deltas in fixed-capacity hardware that cannot scale with application memory footprints. Our characterization of 200 translation-intensive workloads reveals that each virtual memory region has a small, recurring set of deltas, and over 94% of deltas fit in 18 signed bits. We introduce Trail, a temporal TLB prefetcher that stores deltas in unused bits of leaf page table entries (PTEs). When a region triggers a page table walk, Trail identifies the region that last triggered a walk from the same instruction and records their delta in that source region's PTE. When a later walk fetches the source region's PTE cache block, Trail retrieves its deltas without additional memory accesses and prefetches translations for likely destination regions into the TLBs and cache hierarchy. Storing multiple deltas per PTE cache block improves coverage, while using existing PTE bits allows metadata capacity to scale with the application's memory footprint without additional metadata storage. Across 200 workloads and 100 multiprogrammed mixes, Trail improves single-core (four-core) performance by 5.7% (11.5%) on average over a baseline without TLB prefetching, outperforming the best prior standalone TLB prefetcher by 1.7% (2.5%). Trail requires only a 64-entry hardware table to track per-instruction page table walk history. Trail is freely available at https://github.com/CMU-SAFARI/Virtuoso/tree/trail-artifact-release.

Vision-language models (VLMs) are increasingly being applied to three-dimensional medical imaging, but their application to knee MRI remains limited, particularly for interpreting the complementary sequences used in clinical practice. We introduce Knee3DVLM, a sequence-aware VLM that uses full-volume DESS and fluid-sensitive TSE MRI to predict 57 anatomically resolved binary diagnostic targets derived from the MRI Osteoarthritis Knee Score (MOAKS) for structured reporting. We evaluated DESS-only, TSE-only, and paired DESS-TSE configurations using subject-disjoint Osteoarthritis Initiative partitions. In a held-out cohort of 1,074 examinations, the fused model achieved 72.98% average accuracy, 71.17% balanced accuracy, 78.96% mean ROC-AUC, and 78.74% macro ROC-AUC, the highest values among the three configurations. In a secondary multiclass analysis aligned with the released 3DReasonKnee cohort, Knee3DVLM was numerically higher than the strongest reported 3DReasonKnee configuration across five pathology categories. These findings support dual-sequence full-volume modeling for comprehensive knee MRI assessment.

Statistical prediction methods typically require some form of fine-tuning of tuning parameter(s), with $K$-fold cross-validation as the canonical procedure. For ridge regression there exist numerous procedures, but common for all, including cross-validation, is that one single parameter is chosen for all future predictions. We propose instead to calculate a unique tuning parameter for each individual for which we wish to predict an outcome. This generates an individualized prediction by focusing on the vector of covariates of a specific individual. The focused ridge – fridge – procedure is introduced with a two-part contribution: 1) first we define an oracle tuning parameter minimizing the mean squared prediction error of a specific covariate vector, 2) then we propose to estimate this tuning parameter by using plug-in estimates of the regression coefficients and error variance parameter. The procedure is extended to logistic ridge regression by utilizing parametric bootstrap. For high-dimensional data, we propose to use ridge regression with cross-validation as the plug-in estimate, and simulations show that fridge gives smaller average prediction error than ridge with cross-validation for both simulated and real data. We illustrate the new concept for both linear and logistic regression models in two applications of personalized medicine: predicting individual risk and treatment response based on gene expression data. The method is implemented in the R package "fridge".

We investigate the linear codes derived from the zero-divisor graph $\Gamma(R)$ of the finite local ring $R = \mathbb{F}_p[x]/\langle x^4 \rangle$ for an odd prime $p$. Using an equitable partition of $Z^*(R)$, we determine the parameters of the binary incidence code of $\Gamma(R)$ and construct induced bipartite subgraph codes achieving dual minimum distance $4$. Over $\mathbb{F}_p$, we parameterize the adjacency and Laplacian codes, derive the closed-form weight distribution of the Laplacian code via the MacWilliams identity, and establish its LCD property. Additionally, we determine the automorphism groups of the graph and its matrix codes. Finally, from extremal subgraphs of $\Gamma(R)$, we construct two families of Griesmer-optimal linear codes over $\mathbb{F}_p$, comprising a constant-weight minimal code and an optimal two-weight code.

Few-shot meta-learning traditionally formulates task adaptation either as analytical gradient descent through unrolled computational graphs or as metric-based distance comparisons over flattened 1D fea- ture vectors, which either incur costly test-time backpropagation or discard native 2D spatial geometry. In this work, we propose METALEARNNCA, a decentralized framework that achieves few-shot adapta- tion through the dynamical interaction of coupled Neural Cellular Automata (NCAs) without computing analytical gradients during inference. MetaLearnNCA decomposes task adaptation into an Active- NCA, which executes task inference conditioned on a continuous 2D spatial memory grid termed the spatial program, and a learned Meta-NCA, which acts as a decentralized cellular optimizer by diffusing spatial error residuals across local neighborhoods to dynamically update this program. METALEARN- NCA is competitive against canonical meta-learners in-distribution (96.12% on Omniglot) with Out-Of- Distribution transfer gains on MNIST, KMNIST, and Fashion-MNIST transfer across 10 independent testing seeds across 1-, 5-, and 10-shot regimes (e.g., surpassing Prototypical Networks by +10.54% on 10-shot MNIST and a +3.87% gain on 10-shot Fashion-MNIST over FOMAML). Our results establish that robust, gradient-free learning-to-learn can emerge from decentralized cellular dynamics on non-von Neumann substrates.

For a vertex $v$ of a graph $G$, the triangle-degree $\operatorname{td}_G(v)$ is the number of triangles containing $v$. A graph is triangle-distinct, or $K_3$-irregular, if its vertex triangle-degrees are pairwise distinct. Chartrand, Erdős, and Oellermann asked whether a regular $K_3$-irregular graph exists. We prove that such graphs exist for every regularity $r\ge 9$. More precisely, for every integer $k\ge 15$ we construct a $2k$-regular triangle-distinct graph on $4k+2$ vertices; complementation gives a $(2k+1)$-regular example of the same order. The construction uses two antiregular threshold blocks joined by a zero-one matrix with prescribed margins, followed by matrix $2$-switches that preserve those margins. After a uniform reference perturbation, exactly three triangle-degree collisions remain. Switches chosen according to parity remove two of them, and the last possible collision is controlled by a short quadratic discriminant argument. Odd $k\ge 17$ and even $k\ge 62$ are handled symbolically, while the remaining 24 values are settled by exact finite verification. Together with the known examples for $9\le r\le 29$, this closes the positive existence problem for every $r\ge 9$.

Scientific productivity and collaboration are closely related, yet the causal direction between these two dimensions of academic careers remains unclear. Here we analyze the careers of 26,876 Brazilian researchers holding a nationally prestigious fellowship across multiple disciplines to test whether productivity shapes collaboration or the reverse. Focusing on total publications and distinct coauthors, we assess their directional relationship using a causal discovery framework based on asymmetries in conditional and distributional structure to determine the more plausible generative direction. Collaboration conditioned on productivity exhibits a highly regular mean-variance relation consistent with a geometric model, whereas the reverse conditioning yields a much less coherent pattern. Consistently, publication counts follow a lognormal distribution, whereas the coauthor distribution is reproduced by a geometric-lognormal mixture where productivity determines the expected number of collaborators. Surrogate tests indicate that the data are considerably more compatible with productivity shaping collaboration than with the reverse. The two quantities also accumulate asymmetrically: publication inequality rises steadily with career age, whereas coauthor inequality does not, indicating that advantage compounds more in output than in network breadth. Area-resolved analyses reveal a more nuanced picture, and although the direction where collaboration shapes productivity is never favored, the two directions are most clearly distinguished in areas more organized around laboratory and research-group models, where productivity inequality also grows most clearly. Among this cohort of prominent scientists, larger collaboration networks are thus more likely to reflect cumulative advantage in productivity than to independently generate greater output.

The conservative species weighting scheme enables the direct simulation Monte Carlo (DSMC) method to resolve trace species such as electrons and ions in weakly ionized flows. The present study extends this scheme to reacting flows, utilizing a split-merge concept for particles with species-dependent weights. Reactive collisions between differently weighted species are processed according to their respective weights, introducing the production of an integer number of product particles and a probabilistic approach applied to both reactant and product particles to satisfy mass conservation. The extended scheme was validated in single-cell dissociation and exchange reaction test cases, verifying mass and total energy conservation and the mole-fraction histories against conventional DSMC. This extension provides the electron number density required for predicting the onset of communication blackout at high altitudes during hypersonic reentry. The method was applied to the Orbital Re-entry Experiment (OREX) flight conditions from 105.0 km to 92.8 km altitude. Translational and internal temperatures showed good agreement with conventional DSMC results, validating the macroscopic flow-field computation. The scheme enabled continuous spatial distributions of trace electron number density, consistent with the communication blackout onset observed in the flight data. The computed ion current reproduced the overall trend of the in-flight electrostatic probe data. The influences of the ion recombination at the vehicle surface and of the uncertainty of the N + O associative-ionization rate were quantified, and the range of the computed currents encompassed the available measurements. The 105.0 km case extended the prediction beyond available measurements, consistent with the overall trend.

In many-query scenarios, data-driven surrogate models provide an efficient alternative to high-fidelity solvers for simulating physical systems governed by Partial Differential Equations (PDEs). In this context, the Latent Dynamics Network (LDNet) has recently demonstrated remarkable performance in predicting the response of spatio-temporal systems, combining Neural Ordinary Differential Equations with nonlinear dimensionality reduction. However, the original formulation assumes a fixed initial condition, limiting its applicability to many real-world applications where a system evolves from varying starting states. In this work, we overcome this limitation while keeping the end-to-end training procedure of the original LDNet and its encoder-free nature, which preserves its intrinsic independence from spatial resolution and grid topology. We infer the initial latent state directly from a small set of early-time observations, treating latent-state initialization as an adaptation problem, and investigate two strategies: an auto-decoding formulation and a meta-learning approach in which the initial latent state acts as a task-specific context variable. We demonstrate the accuracy of the proposed methods across diverse physical phenomena, spanning advection-diffusion, fluid dynamics, and solid mechanics. Meta-learning markedly accelerates latent-state inference and induces smoother, better-conditioned optimization landscapes, and spontaneously organizes the latent space into a structured representation that reflects physically meaningful features of the underlying dynamics. The coordinate-based decoder enables training from spatially subsampled data while recovering high-resolution solution fields at inference. The resulting approach provides an efficient and resolution-independent surrogate modeling framework for many-query simulations of time-dependent PDEs with varying initial conditions.

Recent observations of very-high-energy gamma rays from massive stellar clusters and microquasars have demonstrated that these systems can efficiently accelerate particles to very high energies, renewing the debate on the dominant sources of Galactic cosmic rays (CRs) and casting further doubt on the role of supernova remnants (SNRs), long considered the prime candidates. In these astrophysical environments, particle acceleration is generally thought to be driven by diffusive shock acceleration (DSA). The question of the duration of the phase in which DSA can efficiently take place is therefore essential to clarify the contribution of the various candidate CR sources. We aim at clarifying the contribution of SNRs to the Galactic CR spectrum, and the expected efficiency of DSA, especially accounting for possible particle acceleration during part of the radiative phase. We compute the total CR proton content at typical SNRs from DSA and reaccelerated pre-existing CRs, and discuss the importance of these two mechanisms throughout the adiabatic and early radiative phases. We find that if DSA and reacceleration take place during a fraction of the radiative phase, the total content of particles energized at SNRs can quickly overcome the typical budget required to account for the bulk of CRs. This suggests that the typical instantaneous efficiency, often defined as the fraction of fluid ram pressure converted into CRs would need to remain below $\lesssim 1$% for most of the SNR evolution. Moreover, the radiative phase can contribute to shaping the spectra injected from typical SNRs. We illustrate that the difference between SNRs from thermonuclear and core-collapse SNe could help account for the measured difference in the helium and proton of the local interstellar spectrum.

Small multiple time series visualizations are often used for real-time data monitoring tasks in high-impact domains such as healthcare and manufacturing. Effective design is critical because users rely on these visualizations to monitor data from many entities, such as patients or machines, often while distracted. Users may need to rapidly appraise current values for each entity, monitoring for those that go outside an acceptable range, while also watching temporal trends. However, no design guidelines currently exist for visually emphasizing current values in historical time series represented by small multiples. Via an iterative design process informed by a review of related literature and theory on visual channels and emphasis, we present a design space for glanceable time series small multiple displays. We evaluate this space through two online empirical studies, testing against non-threshold and threshold rapid appraisal tasks. Our results provide insights into merging current value and historical data visualizations for rapid appraisal tasks in time series monitoring. For non-threshold tasks, we found that size encodings on the current value, spatially integrated into the line chart, may provide a good compromise, with 28% response time improvement for tasks involving finding large current values and minimal interference with trend lookup tasks. More generally, integrated designs outperformed separated designs (in which the current value representation is spatially separated from the historical trend line). For threshold tasks, color threshold encodings significantly outperformed shaded band encodings. All supplemental materials are available at https://osf.io/wzf6a.

This series of papers studies the stability of accretion discs to the reflex motion of the central object under the gravitational pull of the disc itself. In this second paper, we use linear analysis to reproduce or discover seven distinct instabilities. We outline their defining characteristics and propose a classification. Of the four instabilities that had already been discovered, three are artificial: one does not survive the transition to modern surface density profiles such as power-laws, one is due to the location of the domain's boundaries, and one is due to the use of damping layers near the domain's boundaries. Only the SLING instability of Adams et al. (1989) and the three new 'cavity' instabilities seem capable of affecting accretion discs. All four can be described to a high degree of precision by WKB methods. They may drive eccentricity growth, affect planet migration, create spiral and crescent substructures in the absence of any planet, and boost accretion onto the central object. The issue is that the three artificial instabilities are extremely hard to avoid in simulations, and would bias the same observables. This suggests that our community may not have the tools to model the reflex motion of the central object yet.

If the mass distribution within an accretion disc is asymmetric, it will exert a gravitational pull on its central object. The central object must then move away from the origin of the reference frame, resulting in atypical gravitational pulls on the disc. The goal of the present series of papers is to investigate the stability of this feedback loop. In this first paper, we report a surprising finding: the choice of reference frame affects the number and nature of the waves and instabilities contained in the analytical equations. To show this, we use a simple example: the modes discovered by Sanchez et al. (2025, Appendix A.3). We provide the first analytical description of those modes, we explain their physical mechanism, and propose to call them 'bracket modes' because they arise from the bracket side of an integration by parts. We show that the same disc can be stable in the stellocentric reference frame but unstable in the barycentric reference frame. This is because the boundaries of the simulation domain are frame-dependent. To avoid this, one must model the entire disc all the way from the central cavity to the outer edge. But this is prohibitively expensive, and some discs do not have a central cavity anyway. Our results suggest that extreme caution is required when interpreting simulations of asymmetric discs, and that our community may not have the tools to model the reflex motion of the central object yet.

We exhibit a triangulation of an equilateral triangle for which the space of bivariate $C^1$ cubic splines has dimension larger than the dimension formula conjectured by Strang. Notably, the triangulation is such that no two edges sharing a vertex are collinear, and the triangulation is invariant under the action of the isometry group ($D_3$) of the equilateral triangle which is triangulated.

Endovascular catheter navigation relies on continuous fluoroscopy, exposing patients and clinical staff to ionizing radiation throughout the procedure. We investigate whether a physics-based world model can sustain the navigation task between deliberately sparse X-ray acquisitions, and whether the model can signal when its internal state estimate is no longer reliable. We formulate sparse-fluoroscopy navigation as a partially observable Markov decision process where a Cosserat rod simulator supplies transition dynamics, sparse noisy projections provide observations, and a particle filter maintains a belief over the device state, reduced in this implementation to a tip state with a geometric contact proxy. We evaluate this formulation in a deliberately simplified setting: a synthetic planar vessel phantom with a quasi-static rod model and simulated projections, without clinical or animal data. In this setting, the belief-state model tracks the simulated tip with a root mean square error of 1.30 mm while acquiring observations at one fifteenth of the continuous rate (0.97 mm at every frame, 1.51 mm at one thirtieth), reported 90 percent credible intervals achieve 0.92 empirical coverage, and the belief-derived contact risk estimate discriminates unsafe contact events with an AUROC of 0.68. These results constitute a proof of concept on a simplified simulation rather than a demonstration of clinical readiness; their purpose is to establish that calibrated belief, rather than point-estimate accuracy alone, is the essential property a sparse-imaging world model must deliver.

We derive the explicit probability distribution of the total avalanche size produced by m primary electron-hole pairs created at an arbitrary position in a uniform avalanche diode. We also derive an operator that generates all raw moments.

The answers to many economic questions depend on the slope and curvature of demand. Estimating demand for differentiated products entails a trade-off between flexibility and scalability. Existing nonparametric approaches face a curse of dimensionality in the number of products. I develop a scalable nonparametric approach using market-level data that overcomes this curse through economically motivated restrictions embedded in many standard models. Simulations and an application to the U.S. beer market demonstrate flexibility in demand slope and curvature, with consequential differences in counterfactual price responses. Estimation in the application takes seconds, versus several hours for a random coefficient nested logit model.

The maximum-entropy sampling problem (MESP) seeks, for an order-$n$ covariance matrix $C$, a principal submatrix of order $s$ with maximum log-determinant. Mostly for convenience, we assume that $C$ is nonsingular. Al-Thani and Lee (2023) solved MESP in $O(n^5)$ time when $C$ or $C^{-1}$ is tridiagonal. We show that the inner maximization of their recursion depends only on a prefix of the index set and that no piece of a solution is longer than $s$; this gives an $O(ns^2)$-time algorithm that returns the optimal value for every budget $t\le s$. When $C^{-1}$ is tridiagonal, $C$ is, up to scaling, the covariance matrix of an Ornstein–Uhlenbeck process observed at unevenly spaced times, and MESP becomes choosing points on a line under a concave gap function with the Monge property; this gives an $O(ns)$-time algorithm and, in the first-order autoregressive case, a closed-form solution. Given only $C$, we solve MESP in $O(n^2)$ time whenever $C$ or $C^{-1}$ is tridiagonal, up to a symmetric permutation, and we recognize these cases within the same bound. For spiders, we make explicit, and sharpen, the dependence on the number of legs, and, drawing on a hardness result of Ohsaka for stars, we observe that, unless $\mathrm{P}=\mathrm{NP}$, the exponent of the running time must grow with the number of legs.

In this note we present a dimension formula for certain randomized self-similar sets. The self-similar sets we are considering are $d$-dimensional generalizations of rational projections of inhomogeneous sponges: they have overlaps, but the overlapping structure is in some sense well-organized. The randomization happens according to a labelled Galton-Watson tree corresponding to the cylinders of the IFS, for example as in the case of the Mandelbrot percolation set. The dimension formula is given in terms of a pressure function corresponding to the expectation matrices, which describes the overlapping structure and the randomness of the system.

Learned surgical simulators and world models can roll out plausible procedural futures, but they carry no grounded estimate of how often interventional devices actually harm patients. We propose treating population-scale adverse-event surveillance as a distinct belief layer of the surgical digital twin, and we evaluate a federated Bayesian protocol for learning it under formal privacy guarantees. Each site holds per-class Gamma-Poisson posteriors over adverse-event rates and exchanges only Rényi-differentially-private natural-parameter updates. We benchmark on the complete FDA MAUDE cohort for thrombus-retrieval catheters (product code NRY): 8,617 reports, of which 6,491 are classified by transparent keyword rules into five thrombectomy complication classes and partitioned across $K=8$ manufacturer sites. At a matched privacy budget of $(\\varepsilon \\approx 2.09, \\delta = 10^{-5})$, the conjugate protocol attains a held-out Poisson score of -5.78 per test event versus -26.58 for FedAvg with differential privacy. The non-private federated model also outperforms centralized pooling (+3.19 vs +2.93), evidence that manufacturer-specific complication profiles are real and that federation preserves them. Because MAUDE lacks procedure denominators, outputs are relative rate orderings rather than absolute risks, and we report all privacy-utility operating points.

Long-context inference and Retrieval-Augmented Generation (RAG) handle evidence selection at vastly different scales, from a single long prompt to an entire corpus. We ask whether a single model-internal mechanism can select evidence across this range. We introduce UNifying REtrieval And Long-Context with a Single Model (UNREAL), a model-native evidence selection framework to span corpus retrieval and long-context inference. UNREAL encodes chunks and derives retrieval queries directly from the frozen LLM's internal representations. It adds fewer than 500K trainable parameters and leaves the backbone unchanged. On a 3B-token, 21M-chunk Wikipedia index, all four dense and hybrid UNREAL backbones outperform state-of-the-art retriever-reranker systems. The best model raises recall from 49.1% to 73.2% on HotpotQA, from 31.7% to 60.1% on 2WikiMultiHopQA, and from 8.8% to 14.4% on MuSiQue. Applied to long-context tasks, the same selection mechanism removes distractors before generation, raising NoLiMa accuracy from 1.0% to 24.83% at its maximum context length of 128K tokens, and LV-Eval's F1 score from 49.97% to 54.66% at 256K. UNREAL also reduces FLOPs and time-to-first-token relative to full-context inference from roughly 32K tokens onward, with larger gains as context grows. Together, these results establish model-internal evidence selection as a common foundation for corpus retrieval and evidence-sparse long-context inference.

Core-collapse supernovae (CCSNe) interacting with a dense circumstellar medium (CSM) are promising sources of high-energy neutrinos. Most neutrino predictions rely on specific assumptions about the CSM density profile and progenitor mass-loss history. We develop a data-driven framework that connects X-ray observations directly to the expected neutrino signal. Using the temporal evolution of the hydrogen-equivalent absorbing column density, N_H(t), together with the shock velocity inferred from the X-ray plasma temperature, we reconstruct the CSM density profile and shock velocity evolution. We estimate the shock-breakout (SBO) epoch and the resulting neutrino emission, and apply the method to the nearby interacting supernovae SN 2023ixf and SN 2024ggi. For both sources, we compute the expected neutrino event rates for IceCube and KM3NeT/ARCA. We find that neutrino production is concentrated near SBO, when hadronic interactions are most efficient. This naturally defines a short time window, specific to each source, that maximizes the signal-to-background ratio. Our results demonstrate that X-ray observations can constrain at once both the expected neutrino signal and the optimal time window for its detection, providing an observationally-driven strategy for multimessenger searches of these shock-powered transients.

This paper investigates the joint electromagnetic (EM)-domain and baseband (BB)-domain beamforming for integrated sensing and communication (ISAC) systems enabled by pattern- and polarization-reconfigurable pixel-antenna arrays. A multiport-network-based antenna model is developed to map each binary switch configuration to its radiation response using the impedance and open-circuit radiation data extracted from Computer Simulation Technology (CST) Studio Suite simulations. Based on this model, polarization-aware virtual channels are constructed for downlink communication and statistical target sensing, leading to a mixed-integer nonconvex problem that maximizes the average sensing signal-to-noise ratio (SNR) subject to user signal-to-interference-plus-noise ratio (SINR) and transmit-power constraints. To handle the coupled binary and continuous variables, we propose an alternating optimization algorithm, in which the EM-domain switch-state matrices are updated using successive exhaustive Boolean optimization (SEBO), the BB-domain precoder is optimized via a minorization-maximization (MM) and second-order cone programming (SOCP) procedure, and the sensing combiner is obtained from a Rayleigh-quotient update. Simulation results show that the proposed design achieves higher sensing SNR and stronger target-angle robustness than the fixed-patch benchmark.

We propose a simple framework for constructing differentially private confidence regions in one shot, i.e., by adding noise only to the final resampling quantile instead of privatizing the estimator computed on each resample. The cost of privacy of our procedure is only logarithmic in the number of resamples $B$ under with-replacement ($m$-out-of-$n$) sampling and independent of $B$ under without replacement sampling (subsampling), avoiding the $\sqrt{B}$ factor that arises in previous works. We provide nonasymptotic Gaussian Differential Privacy (GDP) and utility guarantees for both subsampling and $m$-out-of-$n$ resampling, covering mean-like estimators with small global sensitivity as well as estimators admitting efficiently computable smooth sensitivity bounds, including quantiles and degenerate U-statistics. This allows us to also obtain private confidence regions for degenerate U-statistics where the private error is much smaller than the non-private error. In all, we provide a toolbox for widely applicable DP uncertainty quantification procedures under popular resampling strategies while avoiding the computational and privacy costs of privatizing many intermediate resample statistics.