Robotics
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Friday 6 November
We focus on the problem of high-speed multi-robot navigation in cluttered spaces with tight spatial constraints. A key challenge is balancing safety with agility, as existing methods often trade one for the other (slowing down near a bottleneck to avoid collisions versus failing to swerve in time at high speeds), especially for car-like systems. Our approach belongs to the family of formation control methods, in which Ackermann-steering agents coordinate while navigating tight spaces. However, current formation control techniques rely on rigid, pre-defined shapes and transitions that depend on initial conditions, the environment, and vehicle dynamics, and are impractical to maintain at high speeds. For example, agents passing through a small gap must first form a square, transform into a line using analytical equations, and then return to a square. In contrast, our approach is inspired by a droplet of water that continuously adapts its shape to the space it flows through without slowing down. We propose a hierarchical algorithm in which, at the high level, a reinforcement learning policy produces a dynamic, blob-like elliptical patch that morphs its shape in real time, squeezing, stretching, and orienting itself around obstacles and through tight spaces. At the low level, the robots are controlled via distributed model predictive control (DMPC) to stay inside the patch, which acts as a soft forcing function, while avoiding collisions with each other. We evaluate our approach, FastFunnels, on simulated F1TENTH racing cars across 14 scenarios with N=1 to N=6 robots, averaged over 20 trials each. FastFunnels achieves an 89% completion rate at 7.73 m/s, while all baselines either fail to complete the track or succeed only at very low speeds (0.07-1.79 m/s, about 40% completion rate).
Multiagent Systems
Sharing the Road: LLM-Based Bargaining in Mixed-Autonomy Traffic
Mixed-autonomy traffic involves heterogeneous driving populations competing for shared road space. Previous studies have shown the existence and emergence of collective rationality using game-theoretic and reinforcement learning approaches, which rely on explicitly specified behavioral assumptions or reward functions for driving agents. However, how bargaining behaviors between driving agents evolve and lead to collective outcomes based on richer agent characteristics and language-based reasoning remains less understood. To fill this gap, we develop a bargaining framework based on large language models (LLMs) in which population-level autonomous vehicle (AV) and human-driven vehicle (HV) agents negotiate road-space allocations under heterogeneous objectives and private information. Through controlled experiments varying traffic composition, agent objectives, and traffic-domain knowledge, we examine bargaining behavior, coordination, and resource utilization. Results show that negotiated allocations respond to relative traffic density without strictly following proportional sharing, while density-balanced states exhibit greater coordination difficulty. Fairness-oriented objectives substantially improve agreement rates but do not necessarily yield equally high road-space utilization. Furthermore, domain knowledge affects bargaining outcomes differently across objective configurations, shaping both coordination and the role-dependent effects of fairness. These findings demonstrate the potential of LLM-based bargaining for modeling and designing interactions among driving agents in mixed-autonomy traffic.
Read the full abstract
Signal Processing
FoundaRF: A Physics-Informed Foundation Model Framework for Radio-Frequency Passive Components
Designing RF/mmWave passive components is a time-consuming process that relies heavily on expert knowledge and iterative EM simulations. Although recent ML surrogate models have shown great promise in accelerating passive design, they remain task-specific and lack generalization. This paper presents FoundaRF, a foundation model framework that unifies passive layout modeling across diverse layout structures and technologies. It combines a self-supervised encoder with adaptive layout processing and physics-aware decoders to enable few-shot transfer. Experiments across four semiconductor technologies show strong generalization to unseen structures, outperforming baselines with up to a 3.3x reduction in prediction error, reducing labeled-data requirements by up to fourfold, and substantially shortening design turnaround. FoundaRF establishes a scalable and data-efficient solution for RF/mmWave passive design automation.
Read the full abstract
Medical Physics
Smart Spectral Doppler Ultrasound Imaging using Motion Tracking based on Wide-beam transmission and Neural Networks: Preliminary Results
Ultrasound spectral Doppler imaging provides quantitative blood flow information for vascular diagnosis, but its accuracy is limited by the one-dimensional (1D) sample volume (SV). Motion from respiration, patient movement, and probe manipulation can move the target vessel out of the SV, causing spectral broadening, signal reduction, or loss of the spectrum. A two-dimensional (2D) SV can address this, but sequential focused-beam transmission reduces the pulse-repetition frequency (PRF), and plane-wave transmission lowers the signal-to-noise ratio (SNR). In this study, we propose a wide-beam transmission method that acquires a 2D SV from a single transmit event while preserving the PRF. A Hamming network, which requires no prior training, tracks axial and lateral motion within the 2D SV and updates the position of an adaptive 1D SV for each frame. Field II simulations showed that the wide-beam transmission had a spectral SNR comparable to that of focused-beam transmission and approximately 10 dB higher than that of plane-wave transmission. The motion tracking correctly detected the target position in over 93% of frames for vessel radii greater than 1 mm, and in 100% of frames for radii greater than 2 mm. In vivo studies of a carotid artery (linear array) and a hepatic vein (convex array) showed that the proposed method recovered obscured spectra and increased the spectral SNR by 9.3 dB and 4.5 dB, respectively. These results demonstrate that the proposed wide-beam transmission with Hamming network-based tracking can improve the accuracy of spectral Doppler imaging under tissue motion. Moreover, we anticipate that the proposed approach can be implemented in real-time due to its single transmission and simple two-layer architecture with low computational complexity.
Read the full abstract
Systems and Control
Supporting Data Center Grid Integration via Co-Optimized Battery Energy Storage Sizing and Capacity Allocation
Interconnecting large data centers can be delayed by grid requirements on demand flexibility, ramping limits, and ride-through performance. This paper proposes an optimization-based framework to size a Battery Energy Storage System (BESS) that accelerates data center grid integration by providing BESS capacity required to meet selected flexibility and grid-support requirements. The model co-optimizes BESS power rating and energy capacity while allocating capability across three services: (i) demand response over a specified duration, (ii) mitigation of high-magnitude, high-frequency load fluctuations through reserved active power with zero net state-of-charge change, and (iii) voltage support via reactive power capability at the point of common coupling. Minute-level operational scheduling enforces state-of-charge feasibility and load ramp constraints, and incorporates a controllable queue of delay-insensitive computing tasks to enhance flexibility. A case study using a real data center load profile demonstrates the proposed method and quantifies the BESS power-energy requirements and operational trade-offs needed to satisfy the multi-service constraints, providing practical guidance for interconnection planning and deployment.
Read the full abstract
Combinatorics
A Trivalent Graph of Girth 14 and Order 372
The smallest known trivalent graph of girth 14 is constructed. The graph has 372 vertices, improving the old upper bound of 384 \cite{exoo14}. The construction uses methods that may be new to this problem.
Distributed Computing
Capital Efficiency, Isolation, Risk, and Ruin in Generative AI Inference Serving Infrastructure
Generative AI serving is an extremely capital-intensive endeavor that demands rapid, step-change improvements in capital efficiency to prevent industry wide increases in the cost of capital from limiting growth plans. Keeping pace with the economic necessities will increasingly require infrastructure providers to confront operationally-inconvenient realities of production serving, including the extreme heterogeneity of generative AI request costs. We have explored workload isolation as a foundational, fleet-level lever for increasing capital efficiency. Isolation is the practice of intelligently grouping "like" requests before execution. Rather than exposing every replica to the full heterogeneous workload distribution, isolation deliberately reshapes the workload presented to the hardware by assigning requests with different cost characteristics to different portions of the fleet. Workload isolation on its own does not make expensive requests cheaper; it prevents a costly portion of the workload from unnecessarily determining the operating point and capital efficiency of the entire fleet. It additionally lays the foundational groundwork that can serve as a launching point for compounding gains via narrow-band optimizations (tailoring) and custom silicon designs while also informing up on pricing policy. This paper uses representative production data and experience to convey the motivation behind isolation as a fleet-level lever for increased capital efficiency.
Read the full abstract
Computer Vision
Seeing vs. Counting: Complexity Twins Show Where Multimodal LLMs Fail on Diagrams
When a multimodal LLM miscounts in a diagram, does it fail at seeing or at counting? Existing visual graph and puzzle benchmarks vary visual density and algorithmic hardness together, confounding the two. We pair counting questions about the same drawing that sit in different counting-complexity classes, such as its spanning trees (a determinant) and its simple paths (#P-complete), and use a transcription probe and a text-only control to separate reading from computing. We test Gemini, Gemma and Qwen3-VL models; for the Gemini models, the answer depends on how the model computes. In chain of thought, most of them read small drawings correctly and then fail at counting, at a rate set far more by the size of the answer than by its complexity class (for the primary model, the odds of an exact answer fall 2.38-fold per decade of the answer). Given code and an exact transcription, the primary model counts correctly even on 10x10 grids, and reading becomes the bottleneck: a single view misses horizontal edges, and a second reading of a rotated copy with crop checks reads 86/96 large grids exactly, against 57/96 when the upright reading is simply repeated.
Read the full abstract
Machine Learning
Calibrating your confidence requires thinking of all the ways you could be wrong
A primary limitation of modern deep learning is the miscalibration of model confidence; traditional architectures fundamentally lack a mechanism to quantify their own uncertainty. Larger models and larger datasets mitigate this issue, but they do not resolve the underlying structural deficiency. Bayesian methods offer a principled uncertainty-aware learning framework through stochastic parameters. Yet, despite the established benefits, the inherent computational overhead often compromises either scalability or quality. This trade-off currently prevents the widespread adoption of Bayesian methods. This Thesis bridges the gap between high-quality and low-complexity by introducing a novel distribution on parameter space: a degenerate Gaussian whose covariance is a projection matrix aligned with the null space of the Fisher information matrix. We demonstrate the scalability of this Bayesian posterior through two numerically efficient algorithms: one for posterior sampling and another for the direct computation of predictive confidence. Empirically, we show State-Of-The-Art confidence calibration and out-of-distribution detection for large models. We provide theoretical grounding for this approach. We study the degeneracies inherent in Amari's information geometry and highlight its most relevant component: the split of parameter's perturbations into two orthogonal subspaces. From this perspective, our posterior naturally emerges as the simplest choice that captures this split, assigning unitary variance to one subspace and zero variance to the other. Moreover, through the tools of differential geometry, we utilize this split to identify and rectify a critical misalignment issue of the non-linearized Laplace Approximation. Finally, with a focus on PAC-Bayes generalization bounds, we pave the way for a deeper understanding of the learning capabilities of overparametrized models.
Read the full abstract
Multimedia
UniMore: Fine-Grained Supervision for Unified Understanding and Generation
Unified multimodal models combine image understanding and generation within one system, creating an opportunity to improve both capabilities through paired image–text supervision. We examine the outputs of a pretrained unified model in both directions and observe a loss of specificity: generated images can contain the requested objects with incorrect attributes, while captions identify a scene without expressing its distinguishing details. These failures call for supervision that specifies both the visual content to realize and the descriptive differences to distinguish. We propose UniMore, a post-training method that derives these complementary learning tasks from the fine-grained content of image–text pairs. The generation route predicts grounded crops from their corresponding phrases, matching the scope of the condition and visual target while using the native visual objective. The understanding route combines positive-caption fitting with ranking against captions that omit or alter details, conditioned on the full image. The two routes are optimized independently, and their learned updates are combined into a single unified model without changing its inference procedure. Experiments show gains across multiple understanding and generation benchmarks.
Read the full abstract
Computers and Society
Same Facts, Different Remedy: Linguistic Profiling and Arbitrariness in LLM and System One Decisions
Large language models are increasingly used to resolve consumer complaints, and the way complainants write varies with age, education and socioeconomic status (SES). We audit three remedy-decision systems using sociolinguistic rewrites of US consumer financial complaints that are instructed to preserve every fact: two generative LLMs that return a single label (GPT-4o-mini and GPT-5-mini) and a probabilistic "System One" model that returns a distribution over remedies (Jev). Using pre-registered hypotheses tested once on held-out complaints, we separate two phenomena that single-label audits conflate. At the group scale, older, college-educated and higher-SES styles receive slightly more generous remedies; these effects replicate but are small, with every monetary-remedy gap within 0.5 percentage points. At the individual scale, two style versions of the same complaint receive different monetary decisions 2.6-6.3% of the time. In a variance-component analysis of the probabilistic model, persona components shared across the two rewrite models account for about 1.6% of within-complaint variation. The persona gaps are associated mainly with delivery (vocabulary, grammar, politeness) rather than content. Positive controls show that all three systems respond 49-387 times more strongly to remedy-relevant facts than to style. Rewriting every complaint into one standard style reduced GPT-4o-mini's education and SES gaps by 94% and 87% while retaining fact sensitivity, but reversed two of the probabilistic model's gaps and did not make individual decisions more stable. Independent LLM auditors flagged material fact changes in 3.3% and 11.7% of rewrites, and instability persisted among unflagged rewrites. Fairness audits of LLM decision systems should report group disparities and individual instability as separate outcomes.
Read the full abstract
Human-Computer Interaction
The Drawing You Watch Is the Portrait You Keep: A Replayable Machine Portrait Studio
Ask an image model for your portrait and a finished picture arrives: no first line, no second thoughts, no record of how it got there, and a different picture if you ask again. I built a portrait studio that works the other way. A sitter chooses one of eight fictional artists, each written with a city, a medium and a tempo of its own, holds still for 8 to 12 seconds, and watches a drawing of themselves appear over 44 seconds. One still is sent after consent, a hosted image model returns a private study the sitter never sees, and a compiler turns that study into one drawing document. The live drawing, the saved sheet, the vector export and the film all read that document, and a replay makes no model call, so the drawing you watch is the portrait you keep. I describe the compiler in enough detail to rebuild it and say plainly what it is: an authored reconstruction at a compressed speed, not a recorded hand. On 16 generated test studies a sheet holds 8,784 to 16,378 timed contacts, about three quarters of them fixed-grid fills and nearly all shorter than one video frame, and recompiling one study gave byte-identical output four times on one machine. The paper also traces a bug that matters more than any of that. The earlier compiler estimated the paper colour tile by tile, read dark skin as paper, and in all eight hands kept less of a deep-complexioned face than a fair one (13 to 48 percent against 41 to 56). On 80 generated studies the rebuilt engine keeps the complexion difference the study already has (raw retention 89 to 120 percent), so the fairness question has moved upstream, to the study and the hosted model. The two test sitters are generated and differ in more than complexion, and no real-face or perceptual study exists yet. The studio is live at https://everyfaceproject.com.
Read the full abstract
Machine Learning
When2What: A Timing-Conditioned Framework for Long-Horizon Forecasting of Continuous Cardiovascular Waveforms
Long-horizon forecasting of quasi-periodic cardiovascular waveforms, such as the electrocardiogram (ECG), arterial blood pressure (ABP), and photoplethysmogram (PPG), requires predicting both when future cardiac events will occur (beat timing) and what the local structural pattern looks like around each event (beat morphology). However, standard sequence forecasters trained with pointwise mean squared error (MSE) average across uncertain future heartbeat phases, flattening sharp R peaks and pulse upstrokes into featureless baselines. Our analysis of intraoperative ECG reveals a clear temporal asymmetry: beat-time predictability from history decays within seconds due to stochastic heart-rate variability, whereas local beat morphology aligned at true beat times remains stable over several minutes. Guided by this insight, we introduce When2What, a timing-conditioned framework comprising three stages: Predict, Commit, and Generate. Given a 120-s history, When2What predicts a dense 60-s timing map, selects and sharpens discrete beat events into unit-height Gaussian impulses, and conditions a neural waveform generator on both these committed events and the observed history. Across intraoperative ECG, When2What substantially outperforms existing time-series and physiological forecasters, achieving a near-term waveform event skill of 0.644 and reproducing 95.62% of its predicted heartbeats in the generated waveform where direct long-horizon forecasters collapse into flat baselines. Furthermore, When2What transfers zero-shot to external hospital and atrial fibrillation cohorts, extends effectively to ABP and PPG pulse waveforms, and is supported by ablations and synthetic analyses that establish explicit timing conditioning as an effective paradigm for quasi-periodic long-horizon waveform forecasting.
Read the full abstract
Machine Learning
Enforcing Symmetries on Pre-trained ML Models for Physics
Post-hoc group averaging (or test-time augmentation) is a simple way to impose symmetry on a trained predictor without changing its architecture or retraining it. For finite and compact groups, averaging over a normalized Haar measure on the group gives the Reynolds operator: a canonical projection onto invariant or equivariant predictors with known non-increasing risk guarantees under certain assumptions. This paper provides numerical experiments on physics emulation models that demonstrate the effectiveness of the approach in this setting. We then extend this framework to non-compact group symmetries. For non-compact groups, the Haar measure cannot be normalized, so group averaging cannot be defined in the same way. We replace it with three alternatives: the Weyl unitary trick (which requires complexifying the space and extending the predictor), a data-dependent orthogonal projection (which involves a non-trivial optimization on the space of invariant or equivariant functions), and a frame averaging approach (which involves mapping each point to a set of orbit representatives). We illustrate the constructions theoretically and empirically on the unit-rescaling group and the Lorentz group. Our results show that in the non-compact case, the Weyl trick and frame averaging do not provide a consistent improvement over the baseline, except where the test data is out-of-distribution (an orders-of-magnitude improvement in this case) and, for the Weyl trick, in low-data regimes. The orthogonal projection can provide improvements in-distribution and out-of-distribution. We also observe that models trained directly on invariant features matched or beat the projection of an unconstrained one. We conclude that post-hoc symmetrization extends to non-compact groups, but it does not consistently outperform training a symmetric model.
Read the full abstract
Artificial Intelligence
Rethinking Optimization Methods for Hierarchical Feature Selection
Rare-event forecasting often depends on interpretable conjunctions: a single keyword, location, or temporal cue may be weak, while their joint occurrence can identify a meaningful precursor. Hierarchical-Task Numerical conjunctive feature Learning (HTNL) selects Numerical Conjunctive Features (NCFs) through a structured sparsity objective. However, the resulting optimization problem is challenging, making optimization design critical to its practical performance. We give a unified diagnosis of four HTNL optimization families: original majorization-minimization (MM), variational quadratic form (VQF), alternating direction method of multipliers (ADMM), and weighted group lasso (WGL). Our analysis distinguishes three key factors that jointly determine HTNL optimization behavior. First, VQF and MM are algebraically equivalent under matched variational updates and optimize a squared-penalty formulation. Second, ADMM optimizes the unsquared objective through consensus splitting; we replace its truncated FISTA proximal solver with a scalar bisection update (ADMM-Bisect) for more accurate treatment of near-threshold blocks. Third, WGL targets the same unsquared objective, while all methods face rapidly decaying active-set admission scores. Across two synthetic datasets and 11 civil-unrest and flu-surveillance tasks, VQF achieves the highest mean AUC with the most compact non-collapsing supports. ADMM-Bisect reduces measured runtime, while calibrated WGL avoids the empty-support failures of its original scale. Synthetic results further show how proximal accuracy, pruning, and search budget affect candidate recovery, especially for longer conjunctions. Code will be made publicly available upon publication.
Read the full abstract
Machine Learning
Action-Centered Causal Bandits for Budget-Constrained Promotion Targeting
Web platforms select among costly interventions for each user to maximize incremental outcomes under a global budget. We formulate this problem as a budget-constrained contextual bandit with heterogeneous treatment effects. Although an action-independent outcome baseline cancels from treatment comparisons, modeling a nonlinear baseline can increase statistical complexity. We propose DR-CATE-CBwK, an action-centered causal bandit that directly learns treatment contrasts. Its online update remains valid under baseline-model misspecification without inverse-propensity weighting. It combines an offline doubly robust learner to warm-start treatment-effect estimates, an online action-centered estimator for sequential contrast learning, and a primal-dual controller with an affordable-action rule to enforce the effective-cost budget pathwise. For linear conditional average treatment effects (CATEs) in a stationary i.i.d. setting, we prove a high-probability (\widetilde{O}(Kd\sqrt{T})) regret bound, where K is the number of priced actions, d the context dimension, and T the horizon. We benchmark against the fluid linear program (LP) under the coverage-guaranteed action rule. We further show that the action-centered estimating moment remains unbiased under any predictable baseline estimate, whereas a specified separate-arm reward regression can exhibit asymptotic contrast bias under action-dependent logging and shared-baseline misspecification. On a semi-synthetic job-marketplace benchmark with real logged features and simulated outcomes, our practical Thompson-restricted variant achieves the lowest regret among evaluated budget-feasible methods in the operating regime. It reduces regret by approximately 20% versus the strongest reward-based baseline, outperforming raw-reward contextual bandits, BwK baselines, and a static offline policy, while never exceeding the effective-cost budget.
Read the full abstract
Computer Vision
LEF-JEPA: Denoising-Trained Latent Energy Fields for Spatial Representation Learning
A joint-embedding predictive architecture (JEPA) that applies prediction supervision only at masked positions struggles to clearly capture object shapes and boundaries in its latent representations. We further observe that some existing methods improve spatial structure while also retaining rich variations in texture and color. However, preserving more low-level visual information does not necessarily make representations better suited to prediction and action. This raises a key question: how can we strengthen the spatial structure in latent representations that supports prediction and action without over-encoding these details? We introduce LEF-JEPA, which learns spatial structure through target-guided latent recovery and transfers this supervision to the visual encoder through a relative-energy constraint; the energy module is used only during pretraining and adds no inference overhead. To reduce the cost of energy optimization during pretraining, we further construct a conditional Ising Hamiltonian that allows low-energy-state search to be performed by specialized Ising solvers. On physical hardware, we validate the feasibility of achieving lower forward latency with fewer local updates. Under a unified CIFAR-10 pretraining and frozen-encoder evaluation protocol, both realizations outperform V-JEPA2.1 in image classification, foreground segmentation, and visual planning.
Read the full abstract
Functional Analysis
Finite-rank and compact Toeplitz commutators on the harmonic Bergman space
We completely characterize finite rank and compactness of commutators of Toeplitz operators with bounded harmonic symbols on the harmonic Bergman space of the unit disk. Apart from affine dependence, finite rank occurs precisely when both symbols are analytic with rational primitives, or both are coanalytic and their complex conjugates have rational primitives. For compactness, the corresponding condition is asymptotic affine dependence on hyperbolic disks of a fixed radius. This boundary behavior is characterized by three weighted vanishing conditions involving a mixed first-derivative expression and two Wronskians of first derivatives, and therefore requires derivatives of order at most two. Our proofs combine a block-matrix decomposition and explicit kernel identities with Luecking's finite-rank theorem and affine boundary rescalings to the right half-plane. Only the full harmonic symbols are assumed bounded; their analytic components need not be bounded separately. As consequences, every finite-rank commutator has square zero, and the compactness criterion yields characterizations of essential normality and compactness for analytic and mixed symbol pairs.
Read the full abstract
Robotics
VLA Handover: Failure-Prediction-Based Switching from On-Robot to Server VLA Models
Large vision-language-action (VLA) models generalize well but are difficult to run in real time on a robot, and small VLAs that run there fail more often. Remote inference over beyond-5G and 6G networks lets a robot use a large VLA on a server, but querying the server for every action chunk limits the number of robots per server. We propose VLA handover: the robot runs a small VLA (S-VLA) and switches to a large VLA (L-VLA) on a server only when the S-VLA is likely to fail. We formulate the switching as minimizing the request cost under a required success probability and, since this problem cannot be solved directly, propose a heuristic framework that retrains neither VLA. In simulation with SmolVLA and $\pi_{0.5}$ on three of the four LIBERO suites, VLA handover raised the success rate by 11-22 percentage points over the S-VLA alone and sent 49-77% fewer requests than server-only control at a 2-4 percentage-point lower success rate, and its completion time grew less with the network latency.
Read the full abstract
Materials Science
Native point defects in rock-salt and wurtzite MnSe: formation energies across the growth window and spin splitting in the A-type order
Manganese selenide grows both as rock-salt $\alpha$-MnSe, a type-II antiferromagnet, and as metastable wurtzite w-MnSe, whose magnetic order has not been measured. We compute the neutral Mn and Se vacancies and the two antisites of both polymorphs with density functional theory plus a linear-response Hubbard $U$, between the Se-rich limit bounded by MnSe$_2$ and the Mn-rich limit set by the measured formation enthalpy. In $\alpha$-MnSe the Mn vacancy is lowest over most of the window (0.92 eV under Se-rich conditions), consistent with its p-type conduction. In w-MnSe, the Se antisite is lowest under Se-rich conditions, narrowly in the A-type order and clearly in the stripe order, and the Se vacancy is lowest under Mn-rich conditions. The Se vacancy has nearly the same formation energy in both polymorphs, whereas defects that change the number of Mn atoms differ by 0.4-1.1 eV, although DFT+$U$ places wurtzite below rock salt, contrary to experiment. A spin-degenerate stripe order lies 14 meV per formula unit below the altermagnetic A-type order of w-MnSe, and the formation energies shift between these orders by up to 0.24 eV, most for the Mn vacancy. Three of the four defects leave an uncompensated moment of 3-5 $\mu_\mathrm{B}$. In the A-type order, band unfolding shows that the sign-reversing splitting of the most strongly split band, about 230 meV and 3 eV below the valence-band maximum, keeps its peak height within the 8% scatter of the analysis at 3% defects, while points on its shoulders move by up to about 70 meV. Band-edge states are not resolved.
Read the full abstract
Cryptography and Security
NotaryMark: Multi-Bit Watermarking for Black-Box DatasetUsage Auditing in Text-to-Image Diffusion Models
As generative artificial intelligence (AI) models increasingly require data, more publicly available yet licensed data is being harvested and used without authorization. Although several previous studies have proposed watermarking methods for dataset copyright auditing, their applicability to the text-to-image (T2I) setting remains notably limited. The core drawback is that watermarks embedded by existing techniques are often treated as noise and are gradually filtered out and attenuated during the latent noising-denoising process of diffusion model training. We propose NotaryMark, to our knowledge the first multi-bit watermarking method specifically designed for black-box auditing of T2I training dataset use. The core idea of NotaryMark is to align the watermark with the latent denoising of diffusion models, thereby overcoming the signal decay that plagues existing multi-bit methods during sampling. We evaluate NotaryMark across three datasets and three diffusion models spanning watermark capacities from 8 to 1024 bits. With a watermark capacity below 32 bits, it achieves a high average bit accuracy of 0.85-0.90. With a watermark capacity exceeding 32 bits, it maintains bit accuracy between 0.67 and 0.80, consistently outperforming existing multi-bit watermark methods whose bit accuracy is around 0.42-0.60. Furthermore, we design a voting-based aggregation scheme to boost the bit accuracy with a modest budget of 30 queries. Building on this, we incorporate error-correcting codes to enable reliable decoding and tracing of the embedded identifier information. Robustness evaluation against fifteen evasion attacks spanning seven categories shows that NotaryMark remains robust under most evaluated settings, while stronger attacks generally suppress watermark evidence only at the cost of substantial utility degradation, demonstrating its practical reliability against malicious tampering.
Read the full abstract
Analysis of PDEs
Sharp condition for Superharmonicity of positive singular solutions to the critical polyharmonic equation
We determine the sharp condition for local superharmonicity of positive singular solutions to the critical polyharmonic equation $(-\Delta)^m u=u^{\frac{n+2m}{n-2m}}$, where $m\geq2$ is an integer and $n>2m$. For a solution in a punctured ball with a nonremovable singularity at zero, local superharmonicity holds if and only if \(\liminf_{x\to0}u(x)>0\). The threshold is sharp. In the biharmonic case $m=2$, for every $n\geq5$, we construct smooth positive solutions of $\Delta^2u=u^{\frac{n+4}{n-4}}$ in a punctured ball with a nonremovable singularity at $0$, satisfying $\liminf_{x\to0}u(x)=0$ and $\limsup_{x\to0}u(x)=+\infty$, while $-\Delta u$ converges to a finite negative limit along a fixed ray. The construction relies on an infinite array of concentrating bubbles arranged on concentric polygons, with a Lyapunov–Schmidt reduction controlling their interactions against a vanishing background.
Read the full abstract
Astrophysics of Galaxies
Identifying Chemically Coherent Stellar Families in the Galactic High-{\alpha} Disk with Low-Resolution Spectroscopy
Identification of chemical substructures in the Galactic high-{\alpha} disk provides valuable insight into its formation history. In this work, we demonstrate the capability of low-resolution spectroscopy for this purpose by applying DBSCAN clustering to 87,275 red giants in the high-{\alpha} disk using 18 elemental abundances derived from LAMOST low-resolution spectra (R~1800). We assess the effectiveness of this approach by comparing the clustering results from the low-resolution data with those obtained by applying the same procedure to high-resolution survey abundances and to mock catalogs. These comparisons suggest that, despite the relatively large uncertainties in individual elemental abundances derived from low-resolution spectra, the astrophysical coherence among different elements provides sufficient information to identify chemically coherent families that may share similar enrichment histories without necessarily sharing a common birth environment. After scanning the parameter space of the DBSCAN procedure and merging the results from different parameter combinations, we identify 133 candidate chemical families containing 4,642 unique stars (5.3% of the high-{\alpha} disk sample). These families exhibit diverse multi-element abundance patterns, with particularly prominent differences in neutron-capture elements at nearly identical [Fe/H] values. These families therefore offer new constraints on the diversity of chemical enrichment pathways and on the assembly history of the high-{\alpha} disk.
Read the full abstract
Information Theory
Stabilizing Vector Approximate Message Passing with Circuit-Designed Lag-Lead Damping
Vector approximate message passing (VAMP) often fails to converge on ill-conditioned matrices, and its damping factor is tuned either manually or adaptively. Using the fact that VAMP iterations are the discretization of an electric circuit, we show that the denoiser and the likelihood act as nonlinear resistors, the Onsager correction is a mirror resistor and damping is a lag filter. This circuit gives a closed-form loop gain whose spectrum can predict when VAMP fails. Because a lag filter cannot isolate the oscillating modes, we add a lead capacitor whose design often results in a negative capacitance. Across a wide range of condition numbers, our proposed damping strategy achieves a higher convergence rate than adaptive damping and Anderson acceleration at a lower mean iteration cost. On ill-conditioned multipath channels, across ranks and sparsity levels, it also converges more reliably than existing damping baselines.
Read the full abstract
Functional Analysis
Power Mean, Karcher Mean, and Log-Euclidean Mean on Positive Constant Upper Triangular Matrices
We study weighted power, Karcher, and log-Euclidean means on the Lie group CP_n of real upper triangular matrices with a constant positive diagonal. Building on the general triangular-group theory, we use the central decomposition CP_n \cong \mathbb R_+ -by-UP_n to separate the scalar power mean from a unipotent mean with reweighted coefficients. We prove that the power mean extends real-analytically through order zero, where it equals the Karcher mean. We establish exterior-power covariance for the Karcher and log-Euclidean means, and tensor-product identities for all three means, distinguishing product weights from paired inputs. Finite nilpotent functional calculus yields concise formulas on CP_2 and CP_3, where the Karcher and log-Euclidean means coincide. In dimension four, their difference is exactly a central cubic commutator, giving a necessary and sufficient equality criterion and an explicit counterexample. In every dimension, the same commutator is the leading term of their small-dispersion difference.
Read the full abstract
Pattern Formation and Solitons
Breathers across the Focusing-to-Defocusing Transition: From Diffraction-Like Dynamics to Talbot Carpets
We experimentally investigate fundamental unstable breather wave groups propagating across the transition from focusing to defocusing regimes. In hydrodynamics, this transition is realized through a rapid reduction in water depth over a short spatial interval. We demonstrate that the time-periodic Akhmediev breather develops a Talbot-like recurrence pattern in the defocusing regime, although only its initial stages are experimentally accessible. In contrast, the Peregrine breather exhibits qualitatively different behavior, characterized by oscillatory envelope dynamics. These oscillations can be understood as a purely linear phenomenon arising within the nonlinear Schrödinger framework when dispersion dominates over nonlinearity, as further confirmed by numerical simulations over extended propagation distances. The experimental observations agree closely with finite-depth nonlinear Schrödinger predictions across regimes with substantially different relative strengths of dispersion and nonlinearity. Our findings reveal a temporal analogue of diffraction and provide a foundation for exploring focusing-to-defocusing transitions in other nonlinear dispersive media.
Read the full abstract
Algebraic Geometry
The splitting type of jet bundles on the projective line
Given a vector bundle $\mathcal{E}$ on a smooth variety X, one can associate to it its jet bundles, or bundles of principal parts. These bundles encode the Taylor expansions of local sections of $\mathcal{E}$ and arise naturally in the study of differential operators. When $X=\mathbb{P}^1$ is the projective line, jet bundles decompose as direct sums of line bundles, making it natural to study their splitting types. The splitting type is well understood in characteristic zero, but exhibits subtler behavior in positive characteristic. Our goal is to give a complete answer in every positive characteristic, and in particular to provide a characterization of semistability for jet bundles on $\mathbb{P}^1$. We also determine their extremal summands and give applications to differential operators and formulas for cohomology of line bundles on infinitesimal thickenings of the diagonal.
Read the full abstract
Data Structures and Algorithms
On the Communication Complexity of Shortcut Set Computation
A recent breakthrough of Kogan and Parter (SODA '22) provided the first polynomial improvement in decades on the size/diameter tradeoff available for shortcut sets. However, algorithmic applications of shortcut sets have mostly not taken advantage of these new bounds. A barrier is that many shortcut set applications occur in decentralized settings, and we generally lack decentralized algorithms implementing the Kogan-Parter bounds. To this end, we study the communication complexity of constructing Kogan-Parter shortcut sets. The edges of an $n$-node directed graph are split between two players, Alice and Bob, who are also given a target diameter $D$. Our main result is a protocol that, for $D \le n^{1/3}$, allows Alice and Bob to agree on a shortcut set matching the Kogan-Parter tradeoff between size and diameter, while exchanging only $\Oish(n^2 / D^2)$ bits of communication. We also provide (non-matching) lower bounds on the communication complexity required to find a correct $D$-diameter shortcut set of any size. The main technical ingredient for the upper bounds is a low-communication protocol for finding a path cover, a common subroutine in shortcut set constructions that has attracted some independent interest. Our protocol crucially avoids applications of Menger's theorem or antichain decompositions, as often done in prior work. On the lower bounds side, we relate the communication complexity of the shortcut set problem to the size of bridge-free path systems.
Read the full abstract
Machine Learning
Energy-Based Reasoners: Your Recurrent Reasoner Should Be an Energy-Based Model
Looping is a powerful inductive bias for reasoning. By repeatedly iterating a recurrent block, looped models can scale their effective depth and enable multi-step reasoning in latent space. However, existing looped models are not inherently guided by an explicit measure of progress, making it difficult to determine when to stop looping and how to improve upon the current state. To enable this, we introduce Energy-Based Reasoners (EBRs), which learn an energy function over latent reasoning states and formulate looped updates as gradient descent on this energy. The energy scores latent states, while its gradient guides refinement and provides a signal for stopping. To train EBRs effectively, we further introduce exploration-based training to discover additional modes and to better shape the energy landscape over a broader set of states. Across four algorithmic reasoning benchmarks, EBRs match or outperform prior recurrent reasoning models, achieving 99.9% accuracy on Sudoku-Extreme while using 345$\times$ less inference compute than the next-best model and maintaining comparable training compute. On all multi-solution benchmarks, EBRs achieve state-of-the-art coverage of distinct valid solutions while maintaining near-100% single-sample accuracy. These results suggest that energy-based models are a promising approach for strong reasoning performance.
Read the full abstract
Computer Vision
Refer-CityFlow and M3Track: A Benchmark Dataset and Reference Framework for Language-Driven Multi-Target Multi-Camera Tracking
Referring multi-object tracking enables users to specify targets through natural language, but existing Refer-MOT benchmarks primarily focus on single-camera settings. We formalize Referring Multi-Target Multi-Camera Tracking (Refer-MTMCT), where free-form queries specify identity-consistent trajectories to be recovered across a network of sequential, sparse-overlap cameras. We introduce Refer-CityFlow, a benchmark built on CityFlow's global cross-camera identity annotations and augmented with language queries requiring joint language grounding, cross-camera association, and spatio-temporal reasoning. Refer-CityFlow covers 96 videos from 40 roadside cameras with 660+ identity-consistent cross-camera trajectories, 4,362 trajectory-level descriptions, and 7,699 queries, including 1,699 real-user queries collected from 50 participants. Reference experiments on the four-camera Scenario 02 show that single-camera referring trackers reach 14.95-18.28 HOTA on real-user queries, while the evaluated cross-view method achieves about 0.05 HOTA in direct transfer. To establish reference performance, we provide M3Track, a modular multimodal reference framework in which a large language model (LLM)-based agent orchestrates interchangeable detection, tracking, and cross-camera association modules. M3Track reaches 44.10 HOTA and 65.40 IDF1 on the same real-user query set. The project repository is available at https://github.com/phuongnguyen194/Refer-CityFlow-and–M3Track.