Peer Revenue Publish or profit

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

All of arXiv

Sunday 6 September

Nuclear Theory

The role of strangeness in baryon and electric charge stoppings

Recently it has been proposed that comparing the net-baryon ($B$) stopping with net-electric charge ($Q$) stopping can help studies of the baryon stopping mechanism in nuclear collisions. Here we find the $B/Q\times Z/A$ ratio to be very sensitive to the difference between $s$ and $\bar s$ quark rapidity distributions. For mid-rapidity of isobar collisions at 200A GeV, a multi-phase transport (AMPT) model gives slightly more $\bar s$ than $s$, which leads to $B/Q\times Z/A<1$, while the model without the $s-\bar s$ asymmetry gives $B/Q\times Z/A \geq 1$. Comparing Ru+Ru and Zr+Zr isobar collisions, the AMPT (and UrQMD) model gives $B/\Delta Q\times \Delta Z/A<1$ at mid-rapidity at all centralities, which strongly contradicts the recent STAR data. We also find that the $B/\Delta Q\times \Delta Z/A$ ratio is very sensitive to the net-light quark ($u,d$) stoppings, but it is less sensitive to the $s-\bar s$ asymmetry than the $B/Q\times Z/A$ ratio by a factor of 3. These results are expected to help us resolve the stopping puzzle and better understand the baryon stopping mechanism in the future.
Read the full abstract

High Energy Physics (Phenomenology)

Reggeization of quarks from next-to-eikonal high-energy QCD

We develop a comprehensive Wilson-line formulation of quark Reggeization in QCD. Starting from a next-to-eikonal operator built from a semi-infinite Wilson line and a background-quark insertion, we identify an interpolating operator for the Reggeized quark and derive its nonlinear rapidity evolution using the background-field method. In the dilute regime, its positive-signature component exhibits Regge-pole evolution governed by the quark Regge trajectory, whereas the negative-signature sector displays mixing between quark and gluon degrees of freedom, in accordance with its known Regge-cut structure. In the planar limit, this mixing is suppressed, and Reggeization emerges without an explicit signature projection, recovering signature degeneracy. We illustrate the universality of the construction by extracting the same Reggeized-quark operator from a more general Wilson-line operator describing a gluon-to-quark transition. Furthermore, we extend the formalism to massive quarks, deriving a coordinate-space evolution kernel whose Fourier transform reproduces the massive-quark Regge trajectory. Finally, from the leading operator-mixing structure of the evolution equation and signature arguments, we show that the Reggeized-quark interpolating operator is expected to remain an eigenstate of the rapidity evolution up to next-to-leading logarithmic accuracy.
Read the full abstract

All authors Tolga Altinoluk, Guillaume Beuf, Jules Favrel, Michael Fucilla

Probability

Uniqueness for DLR equations of $\mathsf{Sine}_\beta$

We prove that, for all $\beta>0$, the Dobrushin-Lanford-Ruelle (DLR) equations for the unit-intensity $\mathsf{Sine}_\beta$ point process have a unique stationary solution under a particular finite electric energy condition. Uniqueness fails if this condition is dropped. This answers a question of Dereudre-Hardy-Leblé-Maïda and gives a canonical statistical physics characterisation of $\mathsf{Sine}_\beta$. The main technical ingredient is a relative entropy estimate which allows us to compare the conditional distribution in a finite box of a DLR solution to a corresponding circular $\beta$ ensemble.
Read the full abstract

Artificial Intelligence

NormViz: A Benchmark and Framework for Grounding Multimodal Reasoning in Global Cultures

AI systems are used worldwide, but they struggle to serve the needs of culturally diverse populations. Prior work on cultural understanding evaluates AI systems on text-only settings or on visual artifact recognition (e.g. foods, clothing). The ability to reason about visually observable behaviors through local social norms, which we call visual norm understanding, remains unexamined. We introduce NormViz-Bench, a high quality, human-validated benchmark of 3,268 contrastive image pairs (6,536 images) spanning 16 countries. Each pair varies only in the culturally relevant behavior (e.g., objects, attributes, spatial relations, and actions) that alters how each image is interpreted. Each image is labeled as conforming to, violating, or irrelevant to local social norms, and pair-level evaluation requires both images to be correctly classified, thereby preventing reliance on superficial visual shortcuts. Even the strongest VLMs, Gemini 3.0 Flash and Qwen2.5 VL 7B, succeed on only 26.6% and 21.6% of pairs, struggling most with identifying violating and culturally benign visual behaviors. Towards bridging this, we introduce NormViz-Train, a training dataset of 64k images paired with explanations. Though absolute performance remains low (<30%), finetuning on NormViz-Train improves pair accuracy relatively by up to 125% and 36% Qwen3-VL 4B and 8B respectively, showing a path forward to teach models to connect visual perception to cultural significance. Together, NormViz-Bench and NormViz-Train establish visual norm understanding as a challenging and consequential frontier for multimodal AI.
Read the full abstract

All authors Akhila Yerukola, Fabrice Y Harel-Canada, Simran Khanuja, Abhinav Sukumar Rao, Ashima Suvarna, Nanyun Peng, Saadia Gabriel, Maarten Sap

Machine Learning

Constrained Bayesian Optimization for Hierarchical Federated Learning in IoT Networks for Plant Disease Classification

The deployment of Hierarchical Federated Learning (HFL) in resource-constrained Internet of Things (IoT) environments requires careful configuration to balance predictive performance with energy consumption and execution time. This challenge is particularly relevant to smart agriculture, where distributed IoT devices can support automated plant disease classification while operating under limited computational and communication resources. This paper presents a constrained Bayesian Optimization framework for the efficient configuration of HFL deployments. The proposed approach jointly explores the deep learning backbone architecture, aggregation strategy, and number of communication rounds, while the federation size is determined according to the spatial coverage requirements of the agricultural deployment. A weighted objective function captures user-defined trade-offs among energy consumption, execution time, and predictive performance, while explicit constraints ensure compliance with deployment-specific resource and accuracy requirements. The framework is evaluated on an IoT-based plant disease classification task considering multiple deep learning architectures, federated aggregation strategies, and communication-round settings. Experimental results across 30 independent optimization runs show that the proposed approach explores only 11.11% of the search space, while consistently identifying solutions within 1% of the exhaustive-search optimum, with a mean optimality gap of only 0.056%.
Read the full abstract

All authors Athanasios Papanikolaou, Athanasios Tziouvaras, Apostolos Xenakis, Periklis Chatzimisios, Shameem A. Puthiya Parambath, George Floros, Enrica Zereik, Ivan Petrovic, Fabio Bonsignorio

Algebraic Geometry

A unified family of counterexamples to Batyrev's non-negativity conjecture on stringy Hodge numbers

Recently, Huang and Satriano determined the sharp dimension threshold for Batyrev's non-negativity conjecture on stringy Hodge numbers: the conjecture holds in dimensions at most four and fails in every dimension at least five. We show that their counterexamples fit into a single broad three-parameter family $Y_{m,d}:=V(\sum_{i=0}^m s_i g_i(x)+h(x))\subset\mathbb{P}^{m+d+1}$, where each $g_i$ is homogeneous of degree $d-1$ and $h$ is homogeneous of degree $d$, whose singular locus is a linear space $\Lambda\cong\mathbb{P}^m$. Blowing up $\Lambda$ gives a log resolution with one smooth exceptional divisor of discrepancy one. We compute the Hodge-Deligne polynomials of both the resolution and the exceptional divisor from their projective-space fibrations over $\mathbb{P}^d$, derive a master formula for $E_{\mathrm{st}}(Y_{m,d}\times(\mathbb{P}^1)^n)$, and prove the structure formula $h_{\mathrm{st}}^{p,q}(Y_{m,d}\times(\mathbb{P}^1)^n)=A_{p,q}+B_{p,q}-C_{p,q}$, where the three terms are non-negative. Thus the only possible source of negativity is $-C_{p,q}$, which is governed by the primitive middle cohomology of $Z_{m,d}=V(g_0,\ldots,g_m)\subset\mathbb{P}^d$. Off the diagonal, every negative entry equals $-C_{p,q}$, a binomially weighted sum of shifted primitive Hodge numbers of $Z_{m,d}$. On the diagonal, the behaviour is governed by the parity of $m+d$. We also obtain a complete classification for $0\le m\le d-1$, $d\ge3$, and $n\ge1$. The only non-counterexamples are $(m,d)=(0,3)$ for every $n\ge1$, and $(m,d)=(2,3)$ for $n\ge4$.
Read the full abstract

Statistical Mechanics

Improved mean squared displacement analysis for anomalous single particle trajectories

The mean squared displacement (MSD) is a cornerstone in the analysis of diffusion processes in complex media. When the system is heterogeneous and, in particular, when single-particle trajectories are short, it is essential to extract maximal information from each measured trajectory. This is typically done by time-averaging squared increments and examining the scaling of the time-averaged MSD in log-log space. However, classical regression methods perform poorly in this setting because time-averaging introduces correlations aggravated by those inherent to anomalous diffusion. We tackle these limitations by applying a generalized least squares framework, which substantially reduces variance and bias in diffusion parameter estimates, especially for short (around 100 points) and ultra-short (around 10 points) trajectories. The method is fully automated and requires no supervision. Furthermore, it enables prediction of estimation error probability density, which is asymptotically Gaussian, for both classical and enhanced approaches. Leveraging this prediction, we introduce a specialized deconvolution algorithm that reconstructs the underlying particle ensemble structure from experimental data.
Read the full abstract

All authors Jakub Ślęzak, Joanna Janczura, Diego Krapf, Ralf Metzler

Econometrics

Beyond Aggregate VARs: A Bayesian Benchmark for HANK Models

Heterogeneous-agent New Keynesian (HANK) models characterize how entire cross-sectional distributions respond to structural shocks. Traditional representative-agent models are routinely disciplined by impulse responses from aggregate vector autoregressions (VARs). HANK models have no comparable established empirical benchmark because they make predictions not only about aggregates, but also about distributions of micro-level data. We propose a Bayesian benchmark that jointly models macroeconomic aggregates and several marginal distributions from repeated cross sections, including distributions observed in different surveys. Our approach can use both standard structural VAR identification approaches on macroeconomic aggregates and identification restrictions imposed on micro-level data. The model delivers a joint posterior of the distributional effects of shocks, without the need for household panel data or a separate first-stage density estimate.
Read the full abstract

Artificial Intelligence

Formation of structural attractors in neuromorphic systems

This paper examines the theory of Invariant Structural Learning (ISL), which proposes a non-optimization approach to concept formation. Learning is interpreted as convergence to structural attractors in a hypergraph space, rather than as the minimization of a global loss function. The paper presents the ISL model, including its mathematical formalization, computational verification, and a hypothetical neurobiological interpretation. The mathematical section introduces the formal apparatus of the structural reduction process and proves its finite convergence, the existence and uniqueness of class structural attractors, and the self-organization of attractor maps. The computational section demonstrates the feasibility of the proposed approach on classical image recognition tasks, utilizing the proposed learning mechanism without backpropagation and with extremely small training datasets. Finally, the neurobiological section formulates hypotheses regarding the possible implementation of structural attractors in dendritic trees, neural coding as a projection of internal attractor dynamics, and the development of neural architectures supporting the proposed learning concept. These hypotheses are discussed in the context of modern experimental data in the fields of dendritic computations, synaptic plasticity, and the structural organization of neural circuits. The proposed neurobiological mechanisms are presented as testable hypotheses rather than established biological facts. The results demonstrate the mathematical consistency and computational feasibility of the proposed model, while the neurobiological hypotheses outline potential directions for its experimental verification.
Read the full abstract

All authors Yurii Parzhyn, Alexander Schwarzmann, Mykyta Lapin, Kostiantyn Bokhan

High Energy Physics (Phenomenology)

Inelastic Self-interacting Dark Matter and LUX-ZEPLIN 248 keV Event in a Dirac Modular Inverse Seesaw

We propose a novel framework that simultaneously addresses the origin of Dirac neutrino masses and the nature of self-interacting dark matter (SIDM). The model is based on an $A_{4}$ modular symmetry to ensure the Diracness of neutrinos as well as the stability of the DM. The neutrino sector realizes a Dirac Inverse Seesaw mechanism where the smallness of the neutrino mass is governed by the vacuum expectation value (VEV) of a singlet scalar $\phi$. This same scalar couples to a vector-like fermion DM candidate, inducing a tiny Majorana mass splitting that renders the DM pseudo-Dirac and inelastic. Crucially, the scalar also acts as a light mediator for DM self-interactions, potentially solving the small-scale structure problems of Cold DM. In light of the recent 248 keV nuclear-recoil event, LZ230616, observed by LUX-ZEPLIN (LZ) DM direct detection experiment, we demonstrate that our inelastic SIDM parameter space naturally accommodates this signal via endothermic scattering kinematics. Furthermore, the spontaneous breaking of the dark parity required for this inelasticity produces a network of cosmological domain walls. We show that the explicit symmetry breaking needed to safely annihilate these walls generates a stochastic gravitational wave background. The non-holomorphic modular symmetry reduces the free parameters, correlating neutrino observables, addressing DM phenomenology, $\Delta N_{\rm eff}$, and gravitational-wave signatures.
Read the full abstract

All authors Pritam Das, Biswajit Karmakar, Satyabrata Mahapatra, Partha Kumar Paul

Algebraic Geometry

On the Brill–Noether Theory of Sheaves on Regular Surfaces

We study the Brill–Noether theory of higher rank sheaves on surfaces $X$ with $H^1(X,\mathcal{O}_X) = 0$ and its interactions with the classical Brill–Noether theory of line bundles on curves lying on $X$. Denoting by $\mathrm{Spl}(v)$ the moduli space of simple sheaves on $X$ with Chern character $v = (r,c_1,\mathrm{ch}_2)$, we introduce the Brill–Noether loci $BN^k(v)^\circ \subseteq \mathrm{Spl}(v)$ of \emph{generically $k$-generated} sheaves and show, under assumptions on $c_1$ and $K_X$, certain well-behavedness results. In particular, if $c_1$ is indecomposable and $X$ is a K3 surface, we generalize a Brill–Noether theorem for sheaves on K3s due to Yoshioka. This is achieved by employing He's deformation theory of the moduli space of coherent systems and a globalization of the Lazarsfeld–Mukai construction, providing a correspondence between the Brill–Noether theories of generically $r$-generated sheaves on $X$ and that of line bundles on curves $C \in |c_1|$. The same techniques yield upper bounds on the dimension of $\mathcal{W}^{r,\mathrm{bpf}}_d(|c_1|) \setminus \mathcal{W}^{r+1}_d(|c_1|)$ consisting of basepoint-free line bundles in terms of the number of endomorphisms of associated Lazarsfeld–Mukai bundles, for arbitrary $c_1$. This generalizes a result of Aprodu and Farkas in the context of Green's conjecture beyond K3 surfaces. In the case $r = 1$ of pencils, we obtain smoothness and expected-dimension results under an effectivity condition on $-K_X|_C$ for $C \in |c_1|$. As an application, we prove Green's conjecture for the general curve on a rational surface with a nef and effective anticanonical divisor.
Read the full abstract

Computer Vision

Generalist Open-World Temporal Perception

The next generation of artificial intelligence systems will likely be natively temporal and multimodal in both inputs and outputs: able to converse, perceive, predict, reason, and synthesize through a shared world representation. Realizing this requires a temporal perceptual substrate integrating sensory streams, language, and structured outputs within a multimodal world model. We seek a generalist open-world perceptual system that represents biological forms, natural physical structures, and artifacts, and their interactions, as a coherent, temporally persistent process. The model should infer geometry, articulation, semantics, interaction structure, and uncertainty from raw multimodal streams; maintain identity through occlusion and viewpoint change; generalize across species, forms, mechanisms, and materials; and abstain or expand its ontology when encountering the unknown. The objective is a structured world state supporting understanding, prediction, counterfactual reasoning, and controllable synthesis. Recent work suggests that some cross-modal and reasoning-like capabilities can emerge from large-scale generative video pretraining, reminiscent of language-model scaling. Yet these capabilities are often accessed through language probes or expressed through photorealistic video, leaving explicit semantic, geometric, or temporal structure largely unexposed. This paper articulates an alternative and complementary paradigm: perception and synthesis as distinct conditionings within a shared Generalist Open-World Temporal Perception Architecture (GOWTPA). Recognition, structured prediction, and simulation arise as different conditionings of the same generative substrate, while reasoning and embodiment-specific policies build upon the resulting world state. This positions generalist temporal perception as a potential foundation layer for broader multimodal intelligence and physical AI.
Read the full abstract

Number Theory

High dimensional Riemann–Roch spaces in linear spaces with small squares

Let $F$ be a function field over an algebraically closed field $K$ and $S$ a finite dimensional $K$-subspace of $F$. The square of $S$ is spanned by all products of pairs of elements in $S$. We conjecture that if $\dim S^2 \leq 3 \dim S - 4$, then $S^2$ must contain a Riemann–Roch space of dimension at least $2 \dim S - 1 + g$, where $g$ is the genus of $F$. This generalizes a theorem of Freiman from additive combinatorics, stating that small sumsets must contain long arithmetic progressions. We prove our conjecture in the case that $S$ is contained in a Riemann–Roch space of dimension at most $3/2\dim S + 1$. For the proof we study the annihilator of $S$ and introduce the notion of weight for linear forms.
Read the full abstract

Symplectic Geometry

Mirror Counts of Spectral Curves

Given a convex lattice polygon $\Delta\subset \mathbb R^2$, let $N_\Delta$ be the count of rational, nodal curves in the linear system of the ample line bundle $L_\Delta$ on the toric variety $\mathbb P_\Delta$ defined by $\Delta$, having fixed intersection with the toric boundary. We show that the relative Jacobian of the linear system defines an integrable system that has a mirror dual in the language of constructible sheaves on a two-torus microsupported on a Legendrian link. In this setting, we define the dual counting problem using an analogue of rulings of Legendrian links in three-space, then prove equivalence with $N_\Delta$. We perform calculations of $N_\Delta$ in several examples using constructible methods, tropical curve counting, and localization in logarithmic Gromov-Witten theory to demonstrate the equality of the different approaches. Moreover, the rulings give rise to a stratification of the moduli of constructible sheaves. We conjecture that this ruling decomposition recovers the refined tropical invariants of Block-Göttsche, and hence encodes higher-genus logarithmic Gromov–Witten invariants, by a theorem of Bousseau.
Read the full abstract

Robotics

Diagnosing and Dynamically Filtering Occupancy World Models for Active Mapping

Active mapping requires a robot to select camera viewpoints that efficiently reconstruct an unknown 3D scene. To reason about unobserved regions, recent systems use pretrained occupancy networks as world models that complete missing geometry. The predicted structure contributes to expected coverage gain and constrains feasible robot motion. Consequently, occupancy errors can change both what the robot chooses to explore and where it is able to move. We diagnose these effects by holding the planner fixed and varying only the occupancy representation provided to it. We consider planning without completion, with learned occupancy, with false positives removed by a ground truth oracle, with false negatives restored by an oracle, and with ground truth occupancy. Our experiments show that correcting false positives or false negatives alone does not consistently improve final coverage. This finding reveals a gap between occupancy accuracy and downstream planning performance. Ground truth occupancy provides a much larger improvement in coverage efficiency than in endpoint coverage, suggesting that planning and reachability remain important bottlenecks even when the geometric world model is accurate. Based on these findings, we introduce a dynamic filtering strategy that preserves predictions in unexplored space while suppressing repeatedly unsupported occupancy using online observations. Preliminary examples show that this strategy can redirect viewpoint selection toward reachable surfaces that would otherwise remain unobserved.
Read the full abstract

Image and Video Processing

Drones as Annotators: Amodal 3D Auto-Labeling for Ground LiDAR with Aerial Priors

Scaling perception data in autonomous driving is hindered by manual 3D bounding box annotation, a costly and labor intensive process requiring substantial domain expertise. Existing auto-labeling methods reduce this burden, but most of them rely on onboard sensors, where a single ground-level viewpoint yields occluded and sparse observations and inaccurate object geometry. We introduce DAA (Drones as Annotators), a drone-assisted training-free framework for amodal 3D auto-labeling. DAA augments ground LiDAR with an aerial agent that provides less occluded vehicle detections with continuous tracks. Ground LiDAR, in turn, provides precise metric 3D geometry unavailable from aerial imagery alone. DAA exploits the two complementary yet heterogeneous modalities through a two-stage framework: (i) air-ground coordinate alignment unifies aerial detections and ground LiDAR scans into a shared map frame and bootstraps coarse 3D bounding boxes from the aligned aerial detections; (ii) EM-like amodal box refinement alternates between identifying the vehicle's foreground points and updating the coarse box from the identified foreground and aerial priors. We evaluate DAA on an air-ground cooperative perception dataset, where it consistently outperforms existing auto-labeling baselines. LiDAR detectors trained on DAA-generated labels remain competitive with those trained on manual annotations at moderate IoU thresholds. Evaluation against a reference vehicle with known poses and dimensions further confirms its accuracy. We also demonstrate that DAA transfers, without parameter retuning, from onboard to stationary roadside LiDAR and to multi-agent fused point clouds. Code and dataset will be released upon publication.
Read the full abstract

Probability

The maximum relaxation time of a random walk on regular graphs

We establish sharp quadratic bounds on the relaxation time of simple random walk on connected regular graphs, with leading constants $3/(2\pi^2)$ and $1/\pi^2$ for even and odd orders, respectively. This resolves a longstanding conjecture known as Aldous–Fill spectral gap conjecture (2002). We also prove uniqueness and stability theorems for the corresponding cubic and quartic chains, and settle the quartic uniqueness conjecture posed by Abdi, Ghorbani, and Imrich (2021) and by Abdi and Ghorbani (2023). We establish sharp bounds on algebraic connectivity under minimum-degree and regularity constraints and prove the conjectured chain structure of the minimisers, identifying their repeating blocks. This resolves a longstanding conjecture of Guiduli and Mohar (1996) and a structural conjecture of Abdi and Ghorbani (2024). Our quantitative stability theorem shows that nearly minimum algebraic connectivity forces nearly maximum diameter, proving their diameter conjecture. We further obtain sharp relaxation-time bounds in terms of edge-connectivity. For nonregular graphs, we establish sharp bounds for the gap between maximum degree and adjacency spectral radius, confirming a conjecture of Liu (2024). As applications, we prove the sharp bounds on hitting and commute times conjectured by Aldous and Fill (2002) for regular graphs.
Read the full abstract

Analysis of PDEs

Sharp Hessian integrability for fully nonlinear elliptic supersolutions in low dimensions

We settle the planar sharp Hessian-integrability conjecture of Armstrong, Silvestre, and Smart [Comm. Pure Appl. Math. 65 (2012), 1169–1184] for viscosity supersolutions of fully nonlinear uniformly elliptic equations. If $\kappa=\Lambda/\lambda$, then the optimal exponent in the $W^{2,\epsilon}$ regularity theory in the plane is exactly $$ \epsilon_2(\kappa)=\frac{2}{\kappa+1}. $$ The same mechanism reaches the known upper obstruction in dimension three throughout the full range $1\le \kappa\le 4$ and therefore gives $$ \epsilon_3(\kappa)=\frac{3}{2\kappa+1} \qquad(1\le\kappa\le4). $$ Beyond this threshold, it yields a closed algebraic lower bound and identifies the precise spectral configuration responsible for the remaining gap. The proof introduces a spectrally resolved continuum-in-opening mechanism for contact geometry. Throughout the contact evolution, the vertex Jacobian retains the negative index of the Hessian and its mean negative curvature—data erased by the classical pointwise reduction—and these variables are optimized only after integration over all openings. This reduces the analysis to a finite family of explicit one-dimensional kernels whose critical exponents recover the sharp planar threshold, identify the exact three-dimensional regime above, and expose a broader principle for extracting second-order regularity from the spectral geometry of contact sets.
Read the full abstract

Artificial Intelligence

Unsound Search with Policy and Value Networks in Legends of Code and Magic

Decision-time search in perfect and imperfect information games with enumerable belief states are effective methods for game AI. Collectible card games are imperfect information games with large belief states. Legends of Code and Magic is a collectible card game competition where the belief states are $2^{101}$. The Legends of Code and Magic (LoCM) champion, ByteRL, plays with no search. Other works claim sound enumeration-based search is unusable in the genre due to the number of belief states. We measured three previously defined properties that predict where theoretically unsound perfect information Monte Carlo's defects are cheap and found LoCM sits in the favorable region. Starting with imitation learning of the runner-up policy, NeteaseOPD, we created a policy and value feed-forward network. Our agent searches over worlds sampled from a prior over the opponent's deck built from the runner-up's drafts. Using our strictest configuration in the battle phase we beat ByteRL with a win percentage of 51.35% 95% CI [50.37, 52.33], over 10,000 pre-registered games using the LoCM official referee and time limit. Search is not a minor factor on the matchup between our agent and ByteRL. Without search this agent scores 26.8% and adding search adds +24.6 points. Unsound search in imperfect information games could be exploitable. We replicate a published best-response attack against ByteRL. We then apply the same attack protocol to two search configurations of our agent, and each one resists it better than ByteRL at every iteration. In LoCM unsound search gives us a stronger and more resilient agent.
Read the full abstract

Computation and Language

Typed Federated Artifacts for the Agentic Web:Sharing Tool-Routing Knowledge Across Frozen,Heterogeneous LLM Agents

An open, networked web will allow agents to run frozen models from multiple vendors, keep their history private, and teach each other which tool to call and when. Flat text (prompts, example pools) makes it difficult for the protocol to distinguish between noise statistics, merging rules, and documentation. Weights and adapters cannot transfer that knowledge between platforms. We suggest sharing typed federated artifacts, schema-validated objects with well-defined fields for per-field privacy (described here, but measured), dispute resolution, and cross-model transfer, and instantiating them as SYNAPSE1, a common tool-routing knowledge. After deleting 192 garbage entries and 1,916 training items that duplicate or almost duplicate test queries, a federated compendium routes within 1.1 points of a centralized one at 20 MB of JSON per client each round on StableToolBench (3,180 tools). The same experience merged and shown to the router as typed fields rather than one flat string is worth 8.5 points on clean data and 7.4 under 60% injected contradiction. Crossing merge and rendering shows the halves are inseparable (the typed merge shown flat is the worst arm), while three conflict policies are indistinguishable, so the conflict log that motivated this work is not the On {\tau}-bench retail, each compendium arm improves GPT-4o agents' per-step tool-call accuracy by at least 6.7 points, attributed to format rather than federated experience. Two cautionary findings conclude the paper: on a topic-labeled math proxy and StableToolBench, a TF-IDF classifier over the same labeled experience beats every LLM routing arm (by 48 and 26 points, mostly retrieval recall) because the benchmark's pool holds labeled queries for every supposedly unseen tool and every test query verbatim before our filter. It cannot measure routing to tools without labels, which routing exists for.
Read the full abstract

Quantum Physics

Quantum-Enhanced Learning Framework for Intelligent and AI-Native 6G Wireless Networks

The recent convergence of 6G wireless systems and Tiny Machine Learning (TinyML) has driven the need for on-device intelligence in edge networks, where ultra-low latency, stringent energy budgets, and tight compute constraints demand novel architectures. Lightweight deep models efficiently extract local patterns but fail to capture global dependencies, while attention mechanisms do so at the expense of energy and computational cost. To bridge this gap, we introduce Quantumer, a hybrid TinyML–quantum framework that integrates multi-scale dilated convolutions and scaled dot-product attention within a lightweight transformer architecture, employing a two-stage transfer learning pipeline from Quantum Pre-Training (Quantumer-Q) to Classical Fine-Tuning (Quantumer-C). We also present QuantiblentLayer, a four-qubit variational circuit that maps compact traffic representations into measurement-based Hilbert-space features using trainable rotations and cyclic entangling operations. The circuit is used only during offline pre-training as a nonlinear embedding teacher and is removed before Quantumer-C deployment, leaving a fully classical inference model without runtime quantum execution. By transferring these quantum-assisted embeddings into an energy-efficient, lightweight transformer, Quantumer achieves strong detection performance with minimal compute and memory overhead on resource-constrained edge devices. The intrusion detection system (IDS) is used as a case study and evaluated on the Edge-IIoTset, TON IoT, and WUSTL-IIoT-2021 datasets. Quantumer-Q achieves competitive compact-model performance with 105.86K parameters, 0.4038 MB memory usage, 0.5525 MB model size, and 5.5646 MFLOPs; the INT8 Raspberry Pi 4 deployment obtains 16.8413 ms latency with a 0.6493 MB footprint. These results support training-time quantum-assisted representation learning for compact edge-deployable IDS.
Read the full abstract

All authors Muhammad Bilal Akram Dastagir, Rakesh Saini, Omer Tariq, Saif Al-Kuwari, Shahid Mumtaz, Anwer Al-Dulaimi, Ahmed Farouk

Robotics

RoboSense: Leveraging Robotaxi Fleets as Drive-by Sensors for Urban Traffic Monitoring

Urban traffic monitoring plays a critical role in safety analysis, congestion management, and incident response. The growing deployment of robotaxis creates a new opportunity for network-level traffic monitoring. Although robotaxis are primarily designed to serve passengers, they can also be leveraged as drive-by sensors to collect traffic data. Compared to conventional infrastructure sensors or probe vehicles, a fleet of robotaxis forms a cooperative perception environment, which can collectively gather spatially and temporally continuous traffic information. This paper proposes a novel dynamic robotaxi routing framework that explicitly incorporates traffic monitoring tasks as an objective. The framework introduces: (1) a cell-based network representation that aligns with sensing capabilities of robotaxis; (2) a cell-level monitoring metric to quantify spatiotemporal robotaxi coverage; and (3) a mixed-integer linear programming (MILP) formulation that jointly minimizes time-dependent travel time and maximizes traffic monitoring performance. A 5 by 5 urban grid network is built in SUMO to evaluate the framework under three robotaxi market penetration rates (2%, 5%, and 10%) with a range of objective weight combinations. Results show that incorporating spatiotemporal network coverage in the objective function can effectively improve the traffic monitoring performance. Interestingly, with appropriate weights between the two objectives, monitoring performance and robotaxi average speed can be improved simultaneously. This suggests better network monitoring leads to more accurate traffic state prediction and improved mobility. This win-win situation could incentivize robotaxi operators to contribute their vehicles as drive-by sensors for traffic monitoring.
Read the full abstract

Networking

An FPGA-in-the-Loop Testbed for MU-MIMO OFDM Beamforming over Ray-Traced Wireless Channels

Wireless networks face ever expanding throughput demands from heterogeneous, high-density user populations, requiring beamforming algorithms that adapt to channel conditions with low latency. Validating such algorithms requires either costly over-the-air testbeds or simulation environments that lack the timing and resource constraints of real hardware, leaving a gap between algorithm design and hardware-realizable deployment. This work presents a frame-based hardware-in-the-loop (HIL) testbed that closes that gap by coupling an FPGA-based implementation of OFDM Waveforms with MU-MIMO beamforming to NVIDIA Sionna's ray-tracing channel simulator, enabling a physical base-station architecture to transmit and receive against a Sionna-rendered digital-twin propagation environment. Unlike prior work that validates beamforming algorithms either purely in simulation or on full RF testbeds, this architecture allows beamforming logic running on actual FPGA fabric to be evaluated under realistic, controllable, and repeatable channel conditions, including UE mobility and site-specific multi-path, without requiring an anechoic chamber or live RF front end. We detail the FPGA OFDM transmit/receive pipeline, the synchronization and data interface between the FPGA and the Sionna environment, and validation of signal quality under AWGN and ray-traced channel conditions, establishing this testbed as a platform for hardware-validated beamforming research.
Read the full abstract

Information Retrieval

Measuring GEO Visibility: Prompt Corpora Define the Answer Market

GEO (generative engine optimization) visibility scores aggregate source appearances, citations, or brand mentions in generated answers. The prompt corpus selects the situations evaluated, while weights determine their relative importance. Together they define an "answer market" that need not represent actual user demand. Prompt wording can alter retrieval, competing sources, and generated answers. Scoring then requires identifying the appearances, citations, or mentions of interest. If a language model performs this task, its instruction can change the score assigned to an unchanged answer. Our critical survey examines how these choices help define what a GEO score measures. It draws on research into whether indicators measure the intended phenomenon, total survey error, and information retrieval evaluation. The framework specifies situation annotation, prompt formulations, execution conditions, weights, and scoring rules. When weights are unknown or remain to be chosen, the framework reports sets of admissible scores. It distinguishes values compatible with data and assumptions about a target population (partial identification) from variation across weighting conventions (normative sensitivity). A citation alone does not establish a source's contribution. The article defines a comparison of answers generated with and without a source in a controlled documentary context, distinct from an intervention on the full engine with competing sources. The framework is supported by reproducible calculations. No new experiments are reported; its general empirical validity remains to be assessed.
Read the full abstract

Robotics

OCTN: Neural OCT Representations for Robot-Guided Precision Intervention

Optical coherence tomography (OCT) offers compact, contactless, micron-scale imaging suitable for intraoperative guidance, but native OCT volumes are discretely sampled, anisotropic, and currently inefficient for downstream geometric reasoning and robot integration. We present OCTN (pronounced "octane"), an implicit neural representation framework that converts volumetric OCT scans into a continuous, differentiable, and spatially faithful tissue-intensity field. OCTN uses a two-stage hybrid training strategy that combines supervision from acquired voxels with inter-slice interpolations, preserving B-scan fidelity while improving continuity in sparsely sampled regions. For versatility, we first show that OCTN enables fast volumetric reasoning by storing the learned tissue representation natively on the GPU, supporting intensity-based spatial queries with up to 43x speedup over conventional CPU processing. We then demonstrate OCTN-enabled OCT-guided robotic laser surgery where the continuous tissue representation supports implicit surface discovery and surface-constrained path planning via multiple optimization strategies, including Newton- and SGD-based optimization. Next, OCTN enables reconstruction of dense volumetric structure from sparsely acquired B-scans, while reducing acquisition time by 4x and preserving clinically relevant structures. Across the newly generated Duke TissueOCT dataset and public OCT datasets, OCTN achieves robust, high-fidelity reconstruction with PSNR > 30 dB and training time < 10 s, while preserving surface consistency within 10 $\mu$m Chamfer distance relative to baseline reconstruction. The TissueOCT dataset and code are publicly available at raprakashvi.github.io/octn
Read the full abstract

All authors Ravi Prakash, Ryan P. McNabb, Patrick J. Codd, Shan Lin

Computer Vision

Disparity Has a Sign: Stereo Matching Beyond the Zero-Disparity Plane

Modern stereo matching models fail when disparity crosses zero, with end-point error (EPE) rising by 4.6-37$\times$. Yet stereoscopic content, from cinema 3D to VR, routinely contains objects behind the zero-disparity plane (ZDP), corresponding to negative disparities. The blind spot cascades through datasets, architectures, and evaluation protocols, all of which inherit the non-negative geometry. Rectified parallel cameras place ZDP at infinity, so every finite depth yields $d=fB/z \ge 0$ by construction, and nothing within the standard pipeline can violate, or even measure, a negative disparity. To measure it, we propose ZDPShift, a benchmark of $21{,}495$ stereo pairs from seven cinematographer-authored open movies, each frame rendered at five zero-disparity-plane positions with dense signed ground truth. Six state-of-the-art image and video stereo matching models collapse once the plane moves. On identical scene content, FoundationStereo goes from $2.24$ px EPE to $75.33$px, with every backbone leaving roughly half of all pixels exceeding a three-pixel disparity error. What is missing, however, is not the underlying matching capability. % The capability itself, however, is already present. Training on supervision synthesized from SceneFlow, which adds no new data or parameters, keeps the error flat across the signed range. Training only the decoder, with the pretrained matching features frozen, performs comparably across all six backbones, with EPE jittering within $0.2$px. Thus, the pretrained features already extend to the negative regime they were never trained on, and only the output convention discarded it. Meanwhile, positive-regime accuracy on KITTI, Middlebury, ETH3D, and Sintel is largely preserved.
Read the full abstract

All authors Jian Shi, Xinge Yang, Chaoyang Wang, Wolfgang Heidrich, Peter Wonka

Information Theory

Quantum Advantage in Multiple Access Wiretap Channels with Entangled Transmitters

We investigate secure communication over a classical multiple-access wiretap channel (MAC-WTC), specifically exploring the benefit of shared entanglement between the transmitters. Under the strict semantic security criterion, we derive an achievable rate region and a regularized expression for the secrecy capacity. We further establish a single-letter upper bound on the secure capacity of the MAC-WTC with entangled transmitters. Our results demonstrate that entanglement strictly enlarges the secrecy capacity of MAC-WTC compared to sharing only classical correlated randomness, a finding we illustrate using a pseudo-telepathy game example. Finally, we establish new strong soft-covering lemmas for the output statistics of multipleaccess channels (MACs) with entangled transmitters. Our results generalize existing results for non-entangled systems.
Read the full abstract

Computer Vision

Comparative Study of Anatomical and Learned Features in AI Models for Structural Brain MRI

In this work, we comprehensively evaluate three popular feature-extraction paradigms in AI-based neuroimaging modeling: (1) computation of anatomical surfaces and volumes, (2) supervised learning with convolutional neural networks (CNNs), and (3) unsupervised pretraining of vision transformer (ViT) foundation models, followed by supervised finetuning. Our study is based on 18 publicly available datasets containing 3D structural T1-weighted MRI scans from approximately 80,000 participants across seven distinct clinical tasks. We observe that a linear model based on anatomical features matches the diagnostic performance of complex nonlinear features learned by sophisticated AI frameworks, including foundation models trained on thousands of scans. Conversely, CNNs and pretrained ViTs learn features that implicitly capture relevant anatomical information, bypassing the need for explicit feature extraction. Building upon these insights, we propose Anatomy Segmentation Pretraining (ASP), a novel method to incorporate anatomical information during foundation-model pretraining, which outperforms existing models in biological age estimation.
Read the full abstract

All authors Boyang Yu, Miquel Lopez Escoriza, Long Chen, Arjun V. Masurkar, Narges Razavian, Carlos Fernandez-Granda

Machine Learning

Train Smarter, Not Harder: Switching Signal-Guided Training in Active Learning

Training strategy, namely whether to retrain from scratch or fine-tune from the previous checkpoint, is an overlooked decision variable in active learning. We show that this choice has exploitable structure: retraining is most useful in early rounds, when each batch can substantially reshape the labeled distribution, while fine-tuning becomes safer once the model trajectory stabilizes. We propose HybridAL, an adaptive training schedule that monitors an online stabilization signal and switches from retraining to fine-tuning after sustained stabilization. Two complementary signals, spectral exponent change $\Delta\alpha$ (weight-based) and accuracy change $\Delta$Acc (validation-based), span different points on the time-calibration trade-off. Across three encoder backbones and six text-classification tasks (five seeds each), HybridAL keeps endpoint macro-F1 non-inferior to retraining and fine-tuning at a 0.010 margin, saves up to 49% of retraining time, and recovers a substantial fraction of retraining's calibration advantage as measured by negative log-likelihood (NLL). Compared with schedules that switch at a pre-committed round, HybridAL obtains lower NLL at moderate additional cost, showing that trajectory-dependent switching provides a stronger time-calibration trade-off than fixed early switching.
Read the full abstract

All authors Nagham Omar, Maya Rozenshtein, Evgeny Mishlyakov, Avigdor Gal

Instrumentation and Methods for Astrophysics

Multimessenger Spectroscopy

A multimessenger method is developed for scheduling optical spectroscopy of ultracompact double white dwarf binaries from Laser Interferometer Space Antenna (LISA) data. The orbital parameters inferred from the gravitational wave (GW) signal are used to predict the argument of latitude $u(t)$, which sets the phase dependence of the radial velocity. With $u(t)$ known in advance, spectra can be taken at selected orbital phases. The GW and spectroscopic descriptions are linked by $u(t)=\Phi_{\rm GW}(t)/2+\delta_{e,\omega}(t)+\mathcal{O}(e^2)$, where $\Phi_{\rm GW}$ is the GW phase, $e$ the eccentricity, and $\delta_{e,\omega}$ the equation of the centre to $\mathcal{O}(e)$, which vanishes for a circular orbit. Radial velocities are regressed on $\cos u$; for double-lined binaries, the two slopes determine the semi-amplitudes $K_1$ and $K_2$. With the inclination and orbital period measured by LISA, both component masses follow from the semi-amplitudes through Kepler's third law, and the resulting chirp mass with the GW amplitude gives the source distance. We derive waveforms to first order in the Laplace–Lagrange parameters; the circular quadrupole waveform is recovered as $e\to0$. Using simulated LISA data for one circular and one eccentric system, we forecast how long $u(t)$ remains accurate enough for phase-resolved spectroscopy. Masses derived from the velocity amplitudes are free of the biases that tides, mass transfer, or acceleration along the line of sight introduce when $\dot f_{\rm GW}$ alone is used. For a four-year LISA observation, scheduling at quadrature remains viable for roughly a decade after the observation ends. Four years after the observation midpoint, the ephemeris predicts the orbital phase to $0.3$~s for the eccentric system and $0.9$~s for the circular one, compared with $1$–$30$~s for optical ephemerides propagated over the same baseline.
Read the full abstract