Networking
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Monday 9 November
Cloud~\ac{VR} imposes stringent latency and reliability requirements due to the closed-loop interaction between \ac{DL} video streaming and \ac{UL} tracking updates. Although Wi-Fi~7 introduces \ac{MLO} to improve throughput and reduce latency, its adoption for latency-critical applications like Cloud~\ac{VR} is not straightforward. This paper investigates Cloud~\ac{VR} streaming over Wi-Fi~7 through a comprehensive cross-layer evaluation and proposes an application-aware \ac{TXS} scheduling policy that exploits the \ac{UL}/\ac{DL} coupling to jointly serve tracking and video traffic. To evaluate our mechanism, we introduce NeXus-VR, an open-source simulator that extends OMNeT++ to integrate the complete Cloud~\ac{VR} application pipeline and the Wi-Fi~7 MAC/PHY layers. Simulation results with multiple \ac{VR} users, \ac{OBSS} contention, and user mobility show that \ac{MLO} and the proposed \ac{TXS} scheduling policy provide complementary benefits. While \ac{MLO} improves transmission opportunities and robustness, the proposed Cloud \ac{VR}-aware \ac{TXS} scheduler preserves tracking freshness by reducing contention between \ac{UL} and \ac{DL} traffic. Together, they increase Cloud~\ac{VR} capacity by up to 2.25$\times$ over conventional single-link operation while consistently reducing latency and improving application-level quality of experience.
Optimization and Control
Multi-Target Tracking via Mean Field Feedback Control
Cooperative tracking of multiple moving targets requires a swarm to continually adjust its deployment. Existing methods often rely on repeated agent-target reassignment or updates to position-dependent partitions, which can complicate coordination as swarms grow. This motivates us to seek an approach with a lighter coordination burden. Specifically, we formulate multi-target tracking as a time-varying density-tracking problem, with the targets' tracking and sensing demands encoded in a task-demand density. From an Onsager-type variational formulation, a closed-form mean-field feedback law is derived to make the swarm density track this evolving density, and the resulting closed-loop system is shown to be well posed. Under the stated assumptions, a Kullback–Leibler (KL) divergence analysis establishes exponential KL convergence under ideal feedback and conditional robustness to feedback mismatch under estimated. Numerical experiments demonstrate effective tracking of multiple moving targets and show smaller density errors at comparable applied control energy compared with an existing density-tracking feedback method.
Read the full abstract
Quantum Physics
Testing non-classical gravitational phase dynamics under strong dephasing
Whether gravity must be quantum is often approached by asking whether two masses can become entangled through their mutual gravitational interaction. In practice, the weak interaction competes with environmental decoherence, so the final state may contain no entanglement even when the underlying gravitational phase remains coherent. We investigate what can still be learned from the dynamics in a dephased Bose–Marletto–Vedral setup. The dominant Newtonian effect for well-separated paths is an accumulated phase rather than transfer between interferometer arms, a regime motivated by matter-wave gravitational phase experiments and quantified here by wavepacket-overlap and deflection estimates. Keeping independently characterized local which-path dephasing fixed, we prove an exact finite-time distinction between the gravitational conditional-phase evolution and every passive stochastic local-phase model (induced by a classical mediator), including regimes where the chosen input remains separable at all times. The distinction survives, with a modified sensitivity, when the comparison is enlarged to population-preserving classical measurement and feedback through a diagonal PPT relaxation. These results identify an operational test of gravitational phase dynamics in the strong-dephasing regimes, conditional on explicit and physically motivated classical mediator models.
Read the full abstract
Quantitative Methods
stMIND: Learning multiscale interaction dynamics from spatial-temporal transcriptomics
Interactions within local cellular neighborhoods and across broader tissue scales jointly govern molecular state transitions and morphogenesis during development and regeneration. While time-resolved spatial transcriptomics simultaneously profiles gene expression and spatial coordinates, destructive sequencing samples distinct cells at each time point, leaving intermediate trajectories and multiscale intercellular interactions unobserved. Building upon our interacting mean-field neural stochastic differential equation framework, we present stMIND (Spatiotemporal Multiscale Interaction Neural Dynamics), a continuous-time generative framework that infers coupled molecular, spatial, and proliferation dynamics from unmatched spatiotemporal snapshots. stMIND decomposes cellular expression velocity vector/drift into three interpretable operators: a cell-intrinsic program, a local graph-difference attention operator that aggregates relative expression differences and spatial geometry within a co-evolving physical radius, and a nonlocal mean-field attention operator that captures long-range population context beyond the immediate neighborhood. By evolving dynamic spatial neighborhoods alongside cellular migration and enforcing orthogonality between intrinsic and interaction-driven drifts, stMIND resolves scale-specific intercellular influences through optimal transport in joint expression-position space. In axolotl and zebrafish benchmarks, stMIND leads in expression prediction, anatomical consistency and spatial coherence. Ablations confirm the need to combine disjoint local and nonlocal operators. It reveals a regenerative transition in axolotl telencephalon 10 to 15 days after injury, with a drift energy surge followed by nonlocal dominance and local counterbalance varying by cell type in reactive ependymoglial niches. It tracks developmental vector fields across zebrafish stages.
Read the full abstract
Mesoscale and Nanoscale Physics
Topological phases and scattering protected by projective chiral symmetry
The bulk-boundary correspondence states that topological properties of a bulk material guarantee the existence of protected edge modes, oftentimes associated with the attractive property of backscattering immunity. In this Letter, we consider a time reversal invariant classical system with both translation and chiral, or sublattice, symmetries. We show that when these two symmetries anti-commute, the material exhibits a topological phase protected by the same invariant as type D topological superconductors. We realize such a system experimentally and observe the existence of the associated edge modes in an acoustic network. Our measurements are in excellent agreement with theoretical predictions from discrete lattice models. Finally, we show that these edge modes are immune to backscattering for perturbations consisting in two subsequent identical defects.
Read the full abstract
Robotics
A Game-Theoretic MPC Planner for Cooperative and Adversarial Drone Coordination
We present and evaluate a game-theoretic model predictive control planner for cooperative goal reaching and adversarial pursuit-evasion with drones. The planner combines 3D prediction, individual objectives, optional shared separation constraints and trajectory-tracking interfaces. A comparison of cooperative game solvers assesses their ability to find feasible trajectories that resolve conflicts. A computational study quantifies how number of drones, prediction horizon and move blocking affect computation time and numerical acceptance. Move blocking holds control inputs constant over groups of prediction steps, reducing the number of independent control variables and lowering computation time. Timing measurements identify the practical limits of real-time planning in the tested configurations. PyBullet simulations demonstrate cooperative goal reaching and pursuit-evasion. A laboratory flight complements the simulations by validating the trajectory-tracking interfaces on physical drones. The planner can serve as a basis for further evaluation and controller development.
Read the full abstract
Software Engineering
A Cost That Practice Accuracy Hides: Worked Solutions in Self-Generated Training Data and the Compositional Use of New Rules
A worked solution can make self-generated training targets shorter and cheaper, but accuracy on the practiced operation may hide how well a model can reuse newly learned rules. We study this choice in a 2B-parameter reasoning model learning two invented quantitative rules. The model generates targets from a rule sheet alone or from the sheet plus a program-written worked solution; targets are accepted for final-answer correctness, and the support is removed for fine-tuning and evaluation. Training questions request forward use in one segment; composition requires sequential reuse of the same rule across two segments. After six passes, using a program-checked numerical signature, sheet-only targets lead by 24.3 percentage points (95% item-bootstrap CI 14.6-34.0) on composition in a prospective test. On unseen argument values, observed forward accuracy is identical (87.5% each), but composition accuracy is 58.3% versus 23.6%, a 34.7-point gap (25.7-43.8). An independently sampled second target-bank pair repeats the direction (+14.6 points, 2.1-26.0). Ten passes on the shorter worked-solution targets, approximately matching the domain tokens of six sheet-only passes while changing other training quantities, narrow the gap to 16.3 points (6.3-26.7) in the first bank pair and 6.2 points (-6.3 to 18.8) in the second, where the pre-specified test does not hold. At six passes each, the worked-solution policy uses fewer capture tokens and less update time, yet the trained models use 2.6 times as many generated tokens per composition success under the numerical signature (2.1 times in the second bank pair). Choosing a capture context therefore requires testing how the new rules are reused, alongside the inference cost of those uses, rather than relying on forward accuracy alone.
Read the full abstract
Optics
Tunable Near-Unity Light Absorption via Surface Exciton-Polaritons in Periodic Gratings of Anisotropic Polarizable Materials
Near-perfect light absorption in ultrathin films is hindered by two fundamental obstacles: the momentum mismatch that prevents resonant excitation of surface exciton-polaritons (SEPs) under ordinary illumination, and the 50% absorption ceiling imposed on free-standing conductive sheets by geometric symmetry. Here we show that both obstacles can be overcome simultaneously in a simple, planar geometry. A one-dimensional periodic grating of anisotropic polarizable ribbons composed of J-aggregates supporting bright molecular excitons provides the reciprocal lattice momentum needed to excite SEPs without employing an attenuated total reflection configuration, while a metallic mirror beneath a dielectric spacer folds the reflected field back onto the absorber, breaking the $50\%$ limit through interference between the incident and reflected waves. We develop a semi-analytical matrix formulation to compute the full optical response of such structure, and demonstrate that the absorptance at the first-order SEP resonance can be continuously tuned from near-unity to near-zero by varying the spacer thickness alone. The grating period and filling factor offer additional spectral control, and oblique incidence unlocks a second, out-of-plane polariton branch. These results establish periodic gratings of excitonic media as a versatile, practical and integrable platform for spectrally selective, near-perfect light absorption.
Read the full abstract
Information Theory
Mitigating the BICM Loss in MIMO Receivers with Correlation-Informed Logits
Bit-interleaved coded modulation (BICM) is widely used in multiple-input multiple-output (MIMO) wireless systems. However, separating MIMO data detection from channel decoding and only exchanging per-bit soft information discards correlation information among the bits within each transmit vector. We show that the resulting information loss limits the performance of practical one-shot receivers. Even though iterative detection and decoding (IDD) can recover most of this loss, IDD significantly increases receiver complexity and latency. In order to reduce this loss, we propose COrrelation-Informed Logits (COIL), a soft-output format that combines per-bit soft information with pairwise coupling logits for correlated bits. We show how to incorporate these coupling logits into a conventional belief-propagation-based channel decoder, which enables correlation-aware decoding in a one-shot receiver without IDD. We evaluate COIL in a MIMO-BICM system with a conventional low-density parity-check code. For a 4x4 MIMO orthogonal frequency-division multiplexing (OFDM) system, a one-shot linear minimum mean-square error (LMMSE) receiver with COIL outperforms a conventional one-shot unbiased LMMSE receiver by 1.3dB and operates within 0.8dB of four-iteration IDD, while requiring only a single pass through the data detector.
Read the full abstract
Artificial Intelligence
PARIS: A multi-agent LLM framework for end-to-end, PRISMA-guided systematic literature reviews with pre-synthesis corpus quality assessment
Systematic literature reviews (SRs) remain slow and labor-intensive, and most automation tools address single stages of the workflow. We present PARIS (PRISMA Agentic Research Intelligence System), an open-source framework in which a conductor agent running in Claude Code orchestrates 15 specialized large language model (LLM) agents, from scoping a research question to exporting a manuscript draft. Agents are specified declaratively, coordinate exclusively through persisted artifacts that the conductor validates before every transition, and delegate retrieval, bibliometric analysis and scoring to deterministic Python utilities. Three human checkpoints gate study inclusion, corpus adequacy and the synthesis outline. We also introduce QualityScore, a pre-synthesis index that aggregates normalized network, bibliometric, semantic and process sub-scores into a 0-100 grade. Two case studies illustrate both: AI chatbots and university students' learning (388 records, 100 included) and AI-assisted SR methods (131 records, 123 included, QualityScore 69.1 recomputed with a proxy network sub-score). They show end-to-end feasibility across domains and the diagnostic use of QualityScore: in the first run, corpus improvements chosen at the corpus-adequacy checkpoint raised the score from 74.6 to 81.9 and the number of pre-2023 studies from 1 to 19. Audits of both runs exposed failure modes: incomplete retrieval, counts that drifted between agents, unrecorded or delegated checkpoint decisions and a grade that depends on how citation data are handled. PARIS fully supports 8 of the 27 PRISMA 2020 items and partially supports 11. Output quality has so far been judged only by an internal reviewer agent; independent validation against human reviews is the next step.
Read the full abstract
Analysis of PDEs
Generation of semigroups for vector-valued pseudodifferential operators on the torus with space-dependent symbols
We consider toroidal pseudodifferential operators with spatially dependent, operator-valued symbols of positive order and finite regularity. For an arbitrary complex Banach space $E$, we prove that the negative of the closure of a parabolic operator on $C^\infty(\mathbb T^n,E)$ generates an analytic semigroup on $W_p^k(\mathbb T^n,E)$, where $1\le p<\infty$ and $k\in\mathbb N_0$, provided the symbol has sufficiently many spatial derivatives and discrete differences. The proof combines uniform estimates for parameter-dependent kernels with right and left parametrices. It does not require the UMD property of $E$. Real interpolation gives corresponding resolvent estimates on $B_{pq}^s(\mathbb T^n,E)$ for $0<s<k$ and $1\le q\le\infty$. The induced semigroups are strongly continuous for $q<\infty$ and on the little Besov spaces when $q=\infty$.
Read the full abstract
High Energy Astrophysical Phenomena
Measurement of the Neutrino Spectrum along the Galactic Plane with IceCube
We present the first measurement of the variation of the high-energy neutrino flux along the Galactic plane using 10 years of neutrino-induced particle-shower data from the IceCube Neutrino Observatory. In this Letter, we fit both the neutrino flux normalization and the spectral index, in multiple regions along the plane, independent of model expectations. The neutrino flux per solid angle in the inner Galactic region ($\vert l\vert< 30^{\circ}, \vert b\vert < 8^{\circ}$) is approximately five times higher than in the outer Galactic plane ($\vert l\vert>60^{\circ}, \vert b\vert < 8^{\circ}$) at $E_\nu=5\mathrm{TeV}$. The spectral indices in the inner and outer region are $\gamma=2.76^{+0.18}_{-0.15}$ and $\gamma = 3.13^{+0.57}_{-0.37}$, respectively. Additionally, we compare the same region of Galactic emission measured in $\mathrm{TeV}$ gamma rays by LHAASO ($15^{\circ}<l<125^{\circ}$) to neutrinos, and conclude that the measurements are compatible within uncertainties, under the assumption of hadronic gamma-ray production.
Read the full abstract
Machine Learning
An Empirical Evaluation of Quantum Kernel Stability in Network Intrusion Detection
Quantum kernel methods have been proposed for network intrusion detection, yet their performance limitations are often attributed to noise in current quantum hardware. This study evaluates that assumption through a controlled analysis of Pegasos-QSVC on the NSL-KDD and ToN-IoT datasets. We perform 56 quantum experiments using one exact noiseless statevector backend and three noise-simulated superconducting backends while keeping preprocessing, PCA features, data splits, and classical baselines identical. XGBoost achieves accuracies of 98.94% on NSL-KDD and 99.41% on ToN-IoT, compared with 91.80% and 82.98% for Pegasos-QSVC. The quantum model also exhibits substantially greater seed-dependent variability. Eight runs collapse into degenerate classifiers, including four recall-collapse cases and four trivial-positive classifiers. Notably, two failures occur on the noiseless backend, demonstrating that hardware noise is not a necessary condition for classi-fier collapse. Kernel Gram matrix analysis further reveals low effective rank and severe ill-conditioning even in the absence of noise, indicating intrinsic fragility in the learned quantum representation. These results show that average accuracy alone can conceal operationally unacceptable behavior and highlight the importance of reporting worst-case performance, seed-level variability, class-specific metrics, and explicit degeneracy tests in quantum intrusion detection studies.
Read the full abstract
Materials Science
Designed atomic displacements for finite-temperature phonons from few force calculations
Finite-temperature phonons are usually fitted to first-principles forces in randomly displaced supercells, and each force calculation can be expensive. We introduce special optimal thermal cells (SOTCs), in which the displacements are designed rather than drawn at random. Our design, SOTC-D, starts from a harmonic trial model. It gives every supercell phonon mode its quantum-thermal amplitude, chooses the sign of each mode so that the force constants are best determined, and pairs each configuration with its inverse so that cubic forces cancel. Updating the trial leads to the self-consistent harmonic solution. Cells that match displacement correlations instead give good thermodynamics but not better dispersions. For eleven crystals, eight force calculations reproduce the effective dispersions within 0.03-0.31 THz, and SOTC-D is the most accurate of five designs in 9 of them. After one trial update, it beats random sampling in all crystals at 4-16 calculations and two-iteration stochastic sampling (sTDEP) in all at eight (supercell basis). On model potentials, the self-consistent form needs 2-16 times fewer calculations than the stochastic self-consistent harmonic approximation in 13 of 14 cases. With the HSE06 hybrid functional, it is ahead of pooled sTDEP at 4-16 calculations in Si and NaCl, and it gives the thermal expansion of Al within 3% of experiment at 300K. The method has limits: in Al and Si cells of more than 100 atoms, sTDEP is more accurate beyond 4-16 calculations (not so in NaCl up to 16); for dynamically stabilized bcc Zr, the outcome depends on the seed; and above 400 K, the free-energy curvature is not resolved.
Read the full abstract
Mesoscale and Nanoscale Physics
Impact of multi-dot configurations on electron temperature and double quantum dot latching in gate-controlled qubit devices
Many currently envisaged scalable architectures for Si/SiGe spin qubit devices utilize gate-controlled multi-dot configurations characterized by their peculiar Coulomb blockade physics. One example is the concurrent use of a single electron transistor (SET) as a charge sensor and as a reservoir for adjacent qubit quantum dots (QD). We find that the SET energy spectrum impacts pulsed gate spectroscopy of the QD when the SET simultaneously acts as the reservoir and the sensor. Then, using the qubit QD as a probe, we quantify how strongly the electron temperature of the SET depends on the applied bias. It surpasses a reference experiment by 100 $mK$ even at 50 $\mu V$ bias. Our results provide a strong incentive for choosing spin qubit initialization schemes which do not rely on coupling to the reservoir in such devices. Switching to a double quantum dot connected to a single reservoir, a frequent configuration for qubit readout, we finally demonstrate the tunability of the characteristic latching behavior by systematically mapping it in the few-electron stability diagram.
Read the full abstract
Biological Physics
Fluctuation-Response Relations and Memory in Tissue Cell Number Dynamics
The longevity of multicellular systems requires finely tuned feedback mechanisms to control the balance of cellular proliferation and loss. To understand the microscopic details governing feedback during homeostasis, it is crucial to quantify the feedback using a robust, minimally invasive method. Existing methods typically depend on perturbing the system to quantify the response. Such perturbations may alter cell physiology and are often technically challenging to implement. We propose a steady-state fluctuation-analysis framework to measure the dynamic response coefficient of cellular tissues. Under the minimal assumption of crowding feedback and Markovian noise, we derive a corresponding theoretical description. We validate this framework against three models of increasing complexity: a simple stochastic simulation, an agent-based mechanical model of tissue growth, and a multi-scale model of membrane-potential-regulated growth. These tests reveal the suitability of using passive fluctuations to characterize the tissue response parameters. Furthermore, this approach can help discover novel phenomena; in particular, we uncover memory effects arising from the non-equilibrium nature of growth. This work opens new avenues to quantify growth parameters in living tissues.
Read the full abstract
Combinatorics
Globally Simple Heffter Arrays $H(n;k)$ with $k \equiv 2 \pmod{4}$
Heffter arrays are combinatorial structures that can be used to construct orthogonal cyclic cycle decompositions of complete graphs. A square Heffter array $H(n;k)$ is a partially filled $n \times n$ array with $k$ filled cells in each row and column. Its entries are distinct and nonzero, form a half-set of the group $\mathbb{Z}_{2nk+1}$, and sum to zero in every row and column. If the natural orderings of the rows and columns of an array are simple, the array is called globally simple. The globally simple property guarantees the orthogonal cycle decompositions constructed from the Heffter array consist of simple cycles. While existence results are known for most congruence classes of the parameters $n$ and $k$ modulo 4 for square globally simple Heffter arrays $H(n;k)$, the existence for $k \equiv 2 \pmod{4}$ has remained largely open. In this paper, we present a construction for globally simple integer Heffter arrays $H(n;k)$ covering all admissible values of $n$ when $k \equiv 2 \pmod{4}$. This settles the existence question across all congruence classes.
Read the full abstract
Combinatorics
A criterion for the set of subsums to be a Cantorval for a certain class of multigeometric series
The paper studies the topological structure of the set of subsums (incomplete sums) of multigeometric series of the form $\frac{k_1}{s} + \frac{k_2}{s} + \ldots + \frac{k_m}{s} + \frac{k_1}{s^2} + \frac{k_2}{s^2} + \ldots+ \frac{k_m}{s^2} + \frac{k_1}{s^3} + \frac{k_2}{s^3} +\ldots$, where $k_1, \ldots, k_m, s \in \mathbb{N}$. In terms of graph theory, necessary and sufficient conditions are established for the set of subsums of a given series not to be a finite union of closed intervals, as well as necessary and sufficient conditions for this set to contain an interval. As a consequence of these statements, a criterion for the set of subsums to be a Cantorval is obtained. An example of applying the criterion to a series whose Cantorval property was previously unknown is given. A more general object of study of topological properties is the set of numbers whose representations in a numerical system with a natural base and a redundant alphabet are subject to restrictions on the use of digits. It is for this object that the above statements are proved.
Read the full abstract
Complex Variables
Branch points and natural boundaries of Poincaré functions and of solutions of nonlinear functional differential equations
In this paper, we study the singularity structure of the analytic continuation of solutions to the equation $g(\lambda w)=R(g(w))$, where $|\lambda|<1$ and $R$ is a rational function. We show that the analytic solution \(g\) near zero is branched, and that its analytic continuation yields an infinite-sheeted Riemann surface with various patterns of branch points. In one specific pattern, some branch points are delayed: they appear only when the analytic continuation reaches a particular Riemann sheet. Using this functional-equation framework, we extend the study to first-order nonlinear functional differential equations. We show that the continuation of certain special solutions near an attracting fixed point yields either single-valued or multivalued functions, with the unit circle as a natural boundary on every sheet.
Read the full abstract
Mesoscale and Nanoscale Physics
Plasmon excitations in monolayer stanene
Monolayer stanene exhibits strong intrinsic spin-orbit coupling (SOC) and notably distinct low-energy band features involving both $\Gamma$ and K valleys together with strong electron-hole asymmetry among group-IV monolayers. Its plasmonic properties and external-field tunability remain largely unclarified. Here, we systematically investigate its high-energy plasmon structures and low-energy collective excitation modulated by temperature, electrostatic doping and perpendicular electric field. Calculations use a $sp^3$ tight-binding model with SOC and sublattice staggered potentials, combined with the finite-temperature multiband Lindhard function under the random-phase approximation. Besides the $\pi$ and $\pi+\sigma$ plasmon known in graphene/silicene, stanene supports two new high-energy plasmon branches. For low-energy plasmons, thermal excitation extends the accessible energy-momentum range while preserving high quality factors and collective coherence. Electrostatic doping yields distinct plasmon behaviors: moderate doping gives K-valley dominated intraband $\sqrt{q}$ Dirac plasmons; heavy electron doping retains $\sqrt{q}$ dispersion via Drude contributions from $\Gamma$ and K valleys; heavy hole doping triggers interband transitions near the $\Gamma$ point, producing new $\sigma$-$\sigma$ and $\sigma$-$\sigma^*$ interband plasmon modes and complex multi-branch spectra. Furthermore, the perpendicular electric field enables continuous tuning of the $\sqrt{q}$ plasmon via band reconstruction below 0.3 V/Å and induces extra plasmon modes near band-gap closure. These tunable low-energy plasmons span terahertz-to-mid-infrared frequencies, promising for nanophotonic applications.
Read the full abstract
Systems and Control
Environmental Assessment of GaN Current-Source and Voltage-Source AC-AC Variable Speed Drives
Three-phase ac-ac converters for variable-speed drives can be realized conventionally as voltage-source converters (VSCs) or, leveraging emerging monolithic bidirectional GaN power transistors (M-BDSs), as current-source converters (CSCs). This paper presents a comparative life-cycle assessment of a realized VSC and a CSC demonstrators including grid-side and motor-side EMI filters, and with identical specifications (1.4 kW rated power, 200 V line-to-line rms grid and motor voltage). The embodied carbon footprints are similar (44.8 kgCO2eq for the VSC and 47.4 kgCO2eq for the CSC) and dominated by auxiliary and control electronics, and the printed circuit boards. Over 10 years of continuous operation under the NEMA Power Index mission profile for variable-load centrifugal systems (pumps, fans), the CSC's higher part-load efficiency translates into a clearly lower life-cycle carbon footprint of around 900 kgCO2eq compared to the VSC's 1130 kgCO2eq when operating with a carbon-intense electricity mix (440 gCO2eq/kWh). However, this difference almost vanishes when considering a future renewable electricity mix (25 gCO2eq/kWh).
Read the full abstract
Data Structures and Algorithms
Tight Competitive Bounds for Online Matching on Forests
We study the online matching problem, its fractional relaxation, and the fractional online vertex cover problem on forests whose edges arrive one at a time in an arbitrary order. Matching decisions are irrevocable, and vertex-cover weights can only increase over time. For online matching, we present a randomized algorithm with expected competitive ratio $4/7$ against an oblivious adversary, improving the previous $5/9$ guarantee. We also present a deterministic algorithm that simultaneously maintains a $4/7$-competitive fractional matching and a $7/4$-competitive fractional vertex cover. This deterministic algorithm improves the previous $5/9$ guarantee for fractional matching and the $11/6$ guarantee for fractional vertex cover. We prove that $4/7$ is optimal for both randomized and fractional matching, even on growing trees of maximum degree four.
Read the full abstract
Plasma Physics
Heterogeneous Acceleration of TRIMEG++ code using OpenMP for Gyrokinetic Simulations
Turbulent transport largely determines the performance of magnetic-confinement fusion devices such as tokamaks, and gyrokinetic particle-in-cell (PIC) simulations are a principal tool to predict it. Advancing up to hundreds of millions of markers and solving large sparse field systems at every time step makes whole-device simulations among the most expensive in plasma physics. GPUs can accelerate this workload, but porting a complex code requires restructuring its parallelism and memory traffic, ideally without a separate version per vendor. In this work, the gyrokinetic code TRIMEG-C1, whose unstructured triangular mesh and high-order finite elements cover both the confined core and the open-field-line edge of the plasma, is rewritten in C++ as TRIMEG++ and fully offloaded to GPUs with OpenMP target directives, so that one portable code base runs on both AMD and NVIDIA hardware. The particle-to-grid projection is offloaded in its standard and field-aligned formulations, the PETSc-based field solver runs on the GPU through the library's backends, and, together with the previously offloaded particle pusher, every stage whose cost scales with the marker count runs on the GPU. All speedups are node-to-node comparisons against a 256-core CPU node (2x AMD EPYC 9754). On the TCV-X21 tokamak reference case, the offloaded kernels reproduce the CPU results to a relative accuracy of 1e-11, and the full simulation runs 4.1-6.1x faster on a node with two AMD MI300A and 2.1-3.0x faster on a node with four NVIDIA A100, with the speedup growing with problem size. The MI300A, with shared CPU-GPU memory, needs no explicit data transfers, whereas the A100 needs them managed by hand. These results show that directive-based offloading can accelerate a complex gyrokinetic code across GPU vendors, making whole-volume tokamak simulations with hundreds of millions of markers practical on a single GPU node.
Read the full abstract
Functional Analysis
On operators with many spectral components
For a complex Banach algebra $\mathscr A$ with unit and $n\in\mathbb N$, let $\mathcal C_n(\mathscr A)$ be the set of all elements whose spectrum has at least $n$ connected components, and let $d_\sigma(\mathscr A)=\sup\{n;\;\mathcal C_n(\mathscr A)\neq\varnothing\}$ be the spectral depth of $\mathscr A$. We characterize nonemptiness of $\mathcal C_n(\mathscr A)$ in terms of orthogonal idempotents, compute $d_\sigma(\mathscr A)$ for finite-dimensional algebras, and show that every class $\mathcal C_n(\mathscr A)$ is open. For operators on an infinite-dimensional complex Banach space $X$ these classes are dense in $\mathscr B(X)$, which implies that $\mathcal C_\infty(\mathscr B(X))$ is a dense $G_\delta$-subset of $\mathscr B(X)$. By contrast, we exhibit a commutative Banach algebra whose spectral classes are all nonempty and none of which is dense. For self-adjoint operators on a Hilbert space, we obtain sharper perturbation results, with positive semi-definite perturbations of prescribed rank producing spectra with exactly $n$ components. We show that a quasinilpotent weighted forward shift $S_w$ on $X$ admits a nilpotent rank-one operator $F$ such that the spectrum of $S_w+\alpha F$ has infinitely many components for every nonzero $\alpha$.
Read the full abstract
Differential Geometry
Geodesic connectedness of globally hyperbolic spacetimes with a causal Killing vector field
We characterize the existence of a geodesic joining two prescribed points of a globally hyperbolic spacetime $(M,g)$ admitting a complete causal Killing vector field $K$ and a complete smooth spacelike Cauchy hypersurface $S$. We prove that two points $p,q\in M$ are joined by a geodesic if and only if there exists an $H^1$ curve $\gamma$ from $p$ to $q$ such that $g(\dot\gamma,K)$ is constant almost everywhere. The statement is intrinsic and allows $K$ to be timelike at some points and lightlike at others. A global splitting induced by the flow of $K$ is used in the proof, but no global bounds on the metric coefficients are imposed. The proof relies on stationary perturbations of $(M,g)$, whose relevant compactness properties follow from global hyperbolicity of $g$, and on uniform estimates based on an integral admissibility condition and a Glaeser-type inequality. Two examples illustrate geodesic connectedness and the obstruction detected by the characterization.
Read the full abstract
Machine Learning
A Geometric Invariant of Attention Predicts Head Specialization from Weights Alone
Transformers are built with continuous symmetries in the attention mechanism. We derive $\rho$, a scalar invariant that measures each head's capacity to couple to rotary positional encodings. This framework allows us to analyze heads without forward passes, opening a new research direction into data-independent analysis of modern LLMs. Empirically, we find that $\rho$ predicts induction heads, the most-studied circuit type in mechanistic interpretability, with Spearman correlation between $-0.55$ and $-0.74$ across four bias-free RoPE models spanning $135$M to $7$B parameters. As a preliminary proof of concept, we use $\rho$ to apply YaRN to only the top-50% positional heads, which degrades perplexity less than uniform application. Code is available at https://github.com/MPedraBento/rho-invariant.
Read the full abstract
Robotics
Graph-Based RL for Sequential Rendezvous Planning with Multiple Patrolling USVs
In patrol missions involving multiple unmanned surface vehicles (USVs), a designated command USV can support fleet management by sequentially rendezvousing with each patrolling USV. These operations require motion-feasible route planning and priority-aware rendezvous sequencing, with the objective of minimizing mission completion time and total tardiness relative to priority-dependent due dates. We propose a framework combining grid-based path planning, surrogate travel-time prediction, and graph-based reinforcement learning. The path planner accounts for ocean currents and approximates turn-rate constraints through multi-stage maneuvers, with path feasibility empirically validated through trajectory-tracking simulations. For rendezvous sequencing, a Feasible Transition Graph (FTG) represents candidate rendezvous events and temporally feasible transitions, enabling a graph neural network to exploit spatiotemporal rendezvous relationships for policy learning. A surrogate predicts travel times from compact current and motion features to avoid repeated path searches during FTG construction and policy training. Evaluation using grid-based path planning shows statistically significant average objective reductions of 10.5%, 3.5%, and 1.5% compared with greedy, MLP-RL, and 2-step lookahead greedy, respectively. The policy achieves these improvements while requiring approximately 1/35 of the latter's computation time. It maintains its advantage over MLP-RL under unseen priority-tier settings, with larger gains as fleet size increases.
Read the full abstract
High Energy Physics (Theory)
Einstein–Yang–Mills Plane Waves: Uniqueness and an Exact Curvature Double Copy
Coleman found plane waves of Yang–Mills theory that are genuinely non-abelian in four dimensions and yet obey linear field equations, and Güven showed that they remain exact when gravity is included. We study these waves in $D$ dimensions. Coleman's profile, which is linear in the transverse coordinates, is singled out twice: up to gauge, it is the only profile with bounded energy density, and the only one for which the full Riemann tensor equals the color-contracted square of the Yang–Mills field strength that sources it. In anti-de Sitter space, we construct a closed-form family for every $D\geq4$; there, the full Riemann tensor cannot equal this color square. Waves whose transverse space is a sphere are always singular. In quadratic and Gauss–Bonnet gravity, a hyperbolic transverse space avoids this only if the waves defocus light or the higher-curvature terms become comparable to the Einstein term. All solutions extend, with changed coefficients, to generic gravity and gauge theories.
Read the full abstract
Statistics Theory
Minimum cumulative power divergence estimation with applications to copulas
Parametric inference based on distribution functions is useful when a model is naturally specified through its cumulative distribution function or when a tractable density is unavailable. We study minimum cumulative power divergence estimation, obtained by comparing a parametric distribution function with its empirical counterpart. For a fixed positive tuning parameter and a fixed finite integration measure, we establish the existence of a consistent sequence of local solutions to the estimating equations and derive its asymptotic normality. The resulting covariance has a sandwich form whose components involve derivatives of the distribution function and the covariance kernel of the multivariate empirical process. We specialize the construction to copulas using Lebesgue measure on the unit cube, and distinguish estimation with known margins from estimation based on ranks. In the latter case, the covariance incorporates the effect of estimating the marginal distributions. The relationship with the copula divergences of Eguchi and Kato is made explicit: the power loss is closely related, whereas the integration measure determines a different estimation criterion. A simulation design for Gumbel, Clayton, and Frank copulas examines the estimation of Kendall's tau under correct specification, family misspecification, and contamination by independence. The simulations show competitive clean-data accuracy but no general robustness advantage over the empirical-copula-weighted alternatives under independence contamination.
Read the full abstract
Human-Computer Interaction
Co-Creating Neurorehabilitation Technology Alongside Lived Experience Experts: A Participatory Design Workshop
Involving patients' lived experience (LE) early in neurorehabilitation technology design can improve the relevance and effectiveness of solutions, yet LE experts are often involved only when finished prototypes are tested. We present and evaluate a participatory design (PD) workshop for early LE involvement, examine how LE contributions influence design processes and outcomes, and provide recommendations for similar approaches. We conducted a PD workshop at the 6th International Conference on NeuroRehabilitation. Five people who have had a stroke (LE experts) and 18 workshop attendees (including the LE experts' caregivers) were organized into five groups. Following the Discover, Define, and Develop phases of the Double Diamond model, each group used Journey Maps and Data Collection Notebooks to co-create a solution for an objective specified by an LE expert. Journey Maps were thematically analyzed, and group audio transcripts underwent structured extraction using a large language model. LE input shifted discussions from discipline-based assumptions toward priorities grounded in lived experience, including fear, confidence, sensory discomfort, fatigue, aesthetics, and independence. Narrow objectives, such as opening a blister pack, produced task-specific solutions, whereas broader objectives generated device concepts, including sensory feedback for cycling and a stimulation-equipped cooking splint. Satisfaction was high among attendees and LE experts. All workshop tools and protocols are shared to support adoption. The workshop offers a practical foundation for involving LE experts early in neurorehabilitation technology development. Future iterations should better capture psychological and social dimensions, allow more reflection, and include the full range of relevant stakeholders through sustained transdisciplinary engagement between researchers and the LE community.