Search arXiv⌕ Search

arXiv subjects

Search papers

Search indexed arXiv papers on artificial intelligence, large language models, computer vision and robotics. Read source abstracts and follow links to arXiv.

At least 1,513 records · Page 84Linked to original sources
arXiv2026

Electromagnetic Modeling of Deformation-Based Pinching Antennas

Pinching-antenna systems (PAS) rely on local perturbations of a guiding structure to radiate part of the guided electromagnetic energy, making the physical response of the resulting pinching antenna (PA) fundamental to both guided-wave propagation and free-space radiation. To capture this coupled behavior, we develop an electromagnetic model of a finite deformation-based PA that directly relates the applied waveguide perturbation to its complex reflection and transmission responses and three-dimensional radiation characteristics. Full-wave MEEP simulations validate the model and demonstrate how the pinching ratio and pinched-section length jointly affect the transmitted power and phase, radiation power-fraction, and radiation pattern, thereby establishing PA geometry as a degree of freedom for controlling the radiation response. Importantly, this geometric dependence extends to communication performance, as the received signal-to-noise ratio (SNR) results reveal that neither minimizing the PA--user distance nor maximizing the radiation power-fraction necessarily yields the highest SNR, since the applied perturbation also determines the realized gain toward the user.

eess.SP↗
arXiv2026

Near-Quadratic Time Multivariate Polynomial Interpolation

Let $\F$ be a finite field. Given $N$ distinct points $\bmα_1,\ldots,\bmα_N \in \F^m$ and associated values $β_1,\ldots,β_N \in \F$, we study the problem of computing an $m$-variate polynomial $f(\vy) \in \F[\vy]$ of individual degree less than $d$, with $d^m \ge N$, that satisfies $f(\bmα_i)=β_i$ for all $i\in[N]$, whenever such an $f$ exists. This is known as the\textit{ multivariate polynomial interpolation} problem. In this work, given $N=d^m$, we present a deterministic algorithm for multivariate polynomial interpolation over an arbitrary finite field $\F$ that runs in near-quadratic time $N^{2+o(1)}\poly(m,d,\log|\F|)$, assuming $d$ is $ω(1)$. This bound holds whenever the associated \emph{multivariate Vandermonde} matrix is nonsingular, has constant co-rank, or admits some special structure. In the general case, we give a randomized Las Vegas algorithm attaining the same near-quadratic running time, provided the finite field is sufficiently large and $d$ grows similar to the previous case. Unlike other multivariate interpolation strategies, we do not assume black-box access to the polynomial. In particular, we are not permitted to query arbitrary input points to obtain their evaluations. Our interpolation result follows by combining Wiedemann's classical approach~\cite{w86} for solving sparse linear systems with known bounds for fast \emph{multivariate multipoint evaluation} over finite fields~\cite{umans08, KU11, BGKM23, BGGKU24}. Specifically, our deterministic interpolation strategy relies on a deterministic variant of Wiedemann's approach, which is more commonly used in the randomized setting.

cs.DS↗
arXiv2026

Query-Time Calibration of CLIP Features for Open-Vocabulary 3D Segmentation

Lifting vision-language features such as CLIP into 3D Gaussian reconstructions enables open-vocabulary segmentation through natural-language queries. Although CLIP represents a broad range of semantic concepts, the lifted features of individual scenes can exhibit strong directional concentration, while raw cosine similarity does not account for the scene-specific feature distribution. We show that substantial segmentation gains can be obtained by changing how a fixed lifted feature field is compared with text. Our method partially subtracts a shared scene direction from visual features and text prototypes and whitens text prototypes using a fixed reference vocabulary of generic terms. It then uses soft clustering for multiclass prediction and CSLS with a relevance margin against the reference vocabulary for single-query prediction. Beyond reducing directional concentration, scene centering improves class-relative feature separation in every evaluated scene. Across seven lifting methods and four datasets, the resulting query-time calibration consistently improves mean intersection over union in both multiclass and single-query segmentation. These gains require no retraining, source-image re-encoding, or feature re-lifting, demonstrating the value of scene-conditioned comparison for making better use of existing lifted representations.

cs.CV↗
arXiv2026

TwinStore: A Synthetic Consumer Platform Using Vision--Language Models

We present TwinStore, a source-available synthetic consumer platform for physical retail built on vision--language models. TwinStore turns a store into a digital twin---from 82 ready-made formats, a floor-plan image or phone photos of its shelves---and fills it with synthetic shoppers who plan a trip, walk the aisles and decide what to buy by looking at rendered shelves with packs, price tags and promotion signs. Every trip ends in a receipt, so a layout, price, pack or promotion can be tried on the same shoppers before it reaches the store. Around this core, TwinStore provides 3D playback, paired comparisons and layout optimization, shopper interviews, a trade-area model, counterfactual edits of real shelf photos, a marketing studio with eleven tools, and an assistant that turns a plain sentence into a drafted study or an answer grounded in a run's own numbers. In a held-out validation on eleven public datasets, simulated receipts approach the households' own natural variation (category-mix divergence 0.017--0.018 against a real drift of 0.012), price elasticities fall inside the range measured in the field, and three independent decision models agree. TwinStore runs on cloud or on-device models and is distributed with installers for Windows, macOS and Linux.

cs.AI↗
arXiv2026

H-2DGS: Height-Field Gaussian Splatting for Sharp and Deformable Neurosurgical Surface Reconstruction

Neurosurgical navigation commonly presents magnetic resonance imaging (MRI) as 2D slices, making it difficult to relate the anatomy in those slices to the cortical surface observed through the surgical microscope. We seek a common 3D visualization that combines a photorealistic reconstruction of the exposed cortex during tumor surgery with whole-brain MRI context and diffusion tractography pathways. This requires the reconstructed cortex to form a clear boundary around the brain anatomy rather than a diffuse volume that obscures internal structures. Conventional 3D Gaussian Splatting (3DGS) can reproduce observed views but, under limited surgical view coverage, may distribute opacity through a thick and geometrically ambiguous band. We introduce height-field 2D Gaussian Splatting (H-2DGS), a three-stage pipeline that concentrates a robust 3DGS reconstruction, fuses its depth into a shared truncated signed distance field (TSDF) surface, and represents that surface with curved, surface-oriented Gaussian primitives. Separate height, color, and opacity bases provide local geometry and appearance, while a neighbor-transition objective coordinates overlapping patches. The explicit surface further supports postoperative alignment: MRI-derived depth deforms the reconstruction, and tracked-color supervision with alternating optimization preserves its appearance. Across five real neurosurgical cases, H-2DGS gives low multi-view depth inconsistency, a compact 7.97MB representation, and interactive rendering at 29--42ms per frame. The resulting sharp, deformable surface integrates clearly with tractography and MRI anatomy, including a postoperative resection cavity.

eess.IV↗
arXiv2026

Energy-Efficient Risk-Constrained Adaptive Semantic Communications Under Uncertain Adversarial Threats

Semantic communication systems based on deep joint source-channel coding can provide task-oriented wireless transmission, but adversarial robustness may require substantial radio and receiver-side defensive-computation energy. This paper develops an energy-efficient, risk-constrained semantic communication framework under source-domain, over-the-air (OTA), and potentially composite adversarial conditions. Long-term total energy, including communication and diffusion-purification energy, is minimized subject to average semantic-loss, trajectory-level tail-risk, and end-to-end latency constraints. A recurrent primal-dual proximal policy optimization controller jointly adapts transmit power, semantic/channel bandwidth ratio, and purification complexity using causal channel observations, semantic confidence, imperfect attack-belief estimates, and recurrent history. A conditional value-at-risk (CVaR) constraint explicitly limits severe semantic failures. On CIFAR-100, at the unseen source-domain level $ε=12/255$, the proposed controller achieves approximately $61\%$ robust Top-1 accuracy with $\mathrm{CVaR}_{0.9}\approx0.36$ while using about $9\%$ less total energy than diffusion-only defense. At the unseen OTA level $Γ_A=-5$ dB, it achieves approximately $52\%$ robust Top-1 accuracy, reduces semantic tail risk by $40\%$, and consumes about $8\%$ less total energy than diffusion-only defense. These results demonstrate energy-efficient semantic defense under partially observable adversarial conditions.

eess.SP↗
arXiv2026

Hierarchical Fine-Grained Cross-Modal Alignment via Dynamic Granularity Discovery and Optimal Transport

Fine-grained image-text retrieval requires a model not only to assess global semantic consistency between an image and a text, but also to model visual evidence for local textual semantics, including objects, attributes, quantities, actions, and relations, so as to distinguish semantically similar candidates. Existing methods commonly use fixed image patches, detected regions, or text tokens as the basic matching units and then aggregate local interactions through max pooling, cross attention, or similarity accumulation. Such strategies remain limited in semantic unit construction and global matching consistency. This paper proposes a hierarchical fine-grained cross-modal alignment framework, termed HFCMA. The framework progressively transforms low-level patch and token representations into adaptive visual regions and textual phrases, and then performs structured region-phrase matching within a global retrieval scoring process. Specifically, HFCMA first adopts dynamic granularity discovery (DGD) to aggregate raw patches and tokens into semantically compact visual regions and textual phrases. It then formulates region-phrase matching as an entropy-regularized optimal transport problem, thereby estimating a structured alignment matrix under global marginal constraints. To reduce background noise and local semantic drift, HFCMA further introduces background debiasing, textual importance weighting, and prototype consistency regularization. Experiments on Flickr30K and MS-COCO show that HFCMA effectively combines dynamic semantic unit construction with OT-based structured matching under a dual-encoder retrieval framework. The resulting fine-grained score provides structured local alignment information for candidate ranking and improves the discrimination of subtle semantic differences.

cs.CV↗
arXiv2026

Watch and Learn: Teaching Visual In-Context Learning by Removing Textual Shortcuts

Vision-language models (VLMs) struggle with in context learning compared to their textual counterparts. Our preliminary study shows that VLMs rely on textual shortcuts and underutilize the visual demonstrations. We therefore hypothesize that VLMs are insufficiently trained to perform visual in-context learning (vICL). To this end, we construct MIME, a task-agnostic vICL training set without textual shortcuts. We ensure the instruction does not state the rule linking the demonstration images to their answers, so the model must infer it from the visual demonstrations. Fine-tuning on MIME improves vICL accuracy by 8.1 to 18.0 points across three VLMs, preserves single-image performance, and improves multi-image and text only in-context learning. In contrast, stating the rule in the training data sharply reduces this gain. We further propose DTAD (Demonstration Text Attention Dropout), which randomly withholds demonstration labels during training so that the VLM must infer the rule from the demonstration images. While achieving comparable vICL accuracy to fine-tuning on MIME, DTAD yields more consistent attention to the demonstration images. On TrueMICL, Qwen3-VL-8B trained on MIME matches frontier models such as Claude Opus 5, and surpasses them with DTAD. Finally, adding MIME to the instruction-tuning mix of LLaVA 1.5 13B more than doubles its vICL accuracy without hurting single-image abilities, whereas a same-size multi-image set improves it only slightly. These results suggest that vICL data should be a standard component of VLM training.

cs.CV↗
arXiv2026

Quasinormal modes, complex null geodesics and singularity resolution in regular AdS black holes

The black hole singularity is hidden behind its horizon, and it is difficult to determine whether it is resolved by quantum gravity or stringy effects. For an AdS-Schwarzschild black hole, it is known that the two-point function of the dual conformal field theory has a singularity associated with a null geodesic that bounces off the black hole singularity. In this paper, we investigate how the resolution of the black hole singularity is encoded in the dual conformal field theory by considering regular AdS black holes as regularized geometries. Using WKB analysis, we compute the quasinormal modes of regular AdS black holes and read off the dual two-point functions from the modes. We find that the singularity corresponding to the bouncing null geodesic for singular black holes is replaced by a smooth bump associated with a complex, rather than real, null geodesic. The location of the complex geodesic singularity is supported by analyzing the structure of the operator product expansion for the two-point function. Our results provide a holographic signature of black hole singularity resolution and clarify how information about the regular interior is encoded in boundary correlation functions.

hep-th↗
arXiv2026

When Code Gets Cheap, Evaluation Gets Scarce: An Experience Report on AI-Assisted Agile Software Engineering

Coding assistants are widely reported to speed up code production; most accounts stop there. We report the end-to-end construction and deployment of a compliance-critical, human-in-the-loop exam-grading system, built by a single developer who wrote no code by hand, protected by a gated process from stakeholder persona to audited behavior. The initial system was built in three working days and could grade a real cohort. We draw one claim from this experience: AI changes software engineering less by making programming faster than by changing what is scarce. Writing code becomes abundant. Reliable evaluation-knowing whether a result is correct and meets the actual need-remains necessary at every iteration and becomes the binding constraint. Engineering judgment remains a second scarcity. Several defects passed the test suite: they sat at the boundaries the tests had mocked, the browser and the language-model API, or appeared only in real use. Deployment exposed assumptions that local tests had missed. Moreover, an AI assistant acting across many steps bypassed process checkpoints until we made them enforceable gates. All of this was observed in a triple union: the developer was also the end user and domain expert, making expert judgment immediately available and evaluation as cheap as it can be. This favors the iteration gains, which may shrink where evaluation requires additional people or expertise, but makes the case a hard test for evaluation scarcity: evaluation still constrained progress. We draw five lessons: run a gated, traceable construction process; specify what is knowable and discover the rest against a running prototype; validate in the real environment; gate the assistant, not only the work; record decisions with their assumptions and the process state the assistant forgets. A second project showed similar patterns; an independent single-expert case corroborates the shift.

cs.SE↗
arXiv2026

Harnessing Language Models as Optimization Modeling Agents

Large language models (LLMs) can translate natural-language optimization problems into solver programs, but reliably producing correct formulations and implementations remains challenging. Existing approaches improve optimization modeling through agentic workflows or task-specific post-training. This raises a practical question: can a lightweight harness using a single model role achieve competitive performance without additional training? We introduce OPTICYCLE and OPTISTATE, two lightweight inference harnesses for optimization modeling, and evaluate them on nine benchmarks comprising 2,634 problems using 11 Qwen variants spanning 0.6B-27B parameters. Our results show that execution-guided revision consistently improves macro accuracy over direct generation for all evaluated models, with gains of up to 49.46 percentage points. Despite their simplicity, both harnesses outperform six adapted agentic workflows in same-model comparisons and four evaluated post-trained optimization models at comparable model sizes, without additional training. Explicit formulation state provides further but model-dependent improvements, suggesting that additional structure is not universally beneficial.

cs.AI↗
arXiv2026

Thermal expansion mechanisms, heat transport, and band-gap renormalization in strongly anharmonic Rb$_2$SO$_4$, Rb$_2$Ca$_2$(SO$_4$)$_3$, and Ba$_3$Lu(BO$_3$)$_3$ from machine-learned interatomic potentials

Rb$_2$SO$_4$, Rb$_2$Ca$_2$(SO$_4$)$_3$, and Ba$_3$Lu(BO$_3$)$_3$ are structurally complex materials whose thermal expansion has been interpreted, on the basis of high-temperature x-ray diffraction, in terms of specific polyhedral rearrangements. We study their lattice dynamics and its consequences for thermal and electronic properties using molecular dynamics with actively learned moment tensor potentials trained on density-functional theory data. The anharmonicity measure $σ^A$ (compares the anharmonic contribution to the interatomic forces with the total forces acting on atoms) reaches $\approx$0.3-0.5 for $β$-Rb$_2$SO$_4$, Rb$_2$Ca$_2$(SO$_4$)$_3$, and Ba$_3$Lu(BO$_3$)$_3$ and $\approx$0.75 for orientationally disordered $α$-Rb$_2$SO$_4$, and the imaginary harmonic modes of Rb$_2$Ca$_2$(SO$_4$)$_3$ and $α$-Rb$_2$SO$_4$ are stabilized at finite temperature. The simulations reproduce the measured thermal expansion and support the mechanisms proposed from diffraction: expansion along $c$ in $β$-Rb$_2$SO$_4$ arises from both elongation and straightening of corrugated Rb chains, BO$_3$ reorientation accompanies the anisotropic expansion of Ba$_3$Lu(BO$_3$)$_3$, and bending of polyhedral linkages limits expansion along the weakly expanding directions and in cubic Rb$_2$Ca$_2$(SO$_4$)$_3$. Green-Kubo lattice thermal conductivities differ by more than an order of magnitude between compounds with comparable $σ^A$, showing that the anharmonicity measure alone does not determine heat transport. Phonon-induced band-gap renormalization, including lattice thermal expansion, reaches 2.7 and 1.7 eV at 1250 K in Rb$_2$Ca$_2$(SO$_4$)$_3$ and Ba$_3$Lu(BO$_3$)$_3$ and 1.4 eV at 650 K in $β$-Rb$_2$SO$_4$, with zero-point renormalizations of 0.64, 0.70, and 0.31 eV for $β$-Rb$_2$SO$_4$, Rb$_2$Ca$_2$(SO$_4$)$_3$, and Ba$_3$Lu(BO$_3$)$_3$, respectively.

cond-mat.mtrl-sci↗
arXiv2026

Learning Cole-Hopf Coordinates for Stochastic Growth Dynamics: the KPZ class

Inferring effective equations from observations of nonlinear stochastic processes is challenging because individual trajectories are noisy realizations, while the relevant physical information is encoded in ensemble statistics. We introduce a data-driven framework that learns a change of variables in which nonlinear stochastic dynamics become linear. The model combines a neural change of representation with a linear stochastic latent evolution and is trained through a short-time moment-matching objective, thereby avoiding trajectory \rev{simulation} during optimization. This construction yields an interpretable generative model: both the learned transformation and the latent evolution operators can be directly inspected. We validate the method on the one-dimensional Kardar--Parisi--Zhang (KPZ) equation, a paradigmatic model of stochastic growth for which the Cole--Hopf transformation maps the nonlinear dynamics to the stochastic heat equation. From simulated height trajectories alone, the model recovers the local exponential Cole--Hopf map and the associated linear stochastic operator. Generated trajectories reproduce the main statistical signatures of the KPZ universality class, including the growth law $\mathrm{Var}(h^{t})\sim t^{2β}$ with $β=1/3$ and the GOE Tracy--Widom height distribution for flat initial conditions. These results show that interpretable stochastic equations can be inferred from data by learning representations in which the dynamics closes in a simple form. Furthermore the development of the physical informed learning permits to get insights on the physics of the dynamical process.

cond-mat.stat-mech↗
arXiv2026

DConstrain: Compiling Context-Parameterized Transitions for Diffusion Language Model Constrained Decoding

Diffusion language models (DLMs) decode multiple tokens in parallel, so individually valid tokens can be incompatible together under a schema. Exact joint inference over a grammar graph enforces compatibility, but it requires token-labeled successors across many hypothetical parser configurations. Encoding stacks and field histories in compiled states can cause exponential growth, which forces prior exact decoders to fix property orders and exclude valid outputs. We present DConstrain, which compiles schemas into context-parameterized transition programs. Cold compilation specializes token consumption to local grammar states while retaining context-dependent stack operations, environment guards, and caller-dependent deferred parsing symbolically. Tracking field histories as runtime values supports arbitrary property order without enumerating field subsets offline. At runtime, warm compilation binds reached configurations to these programs to construct exact successors, overlapping with the model forward pass because graph construction requires no model scores. Furthermore, token-effect grouping merges tokens with identical compiled effects before runtime contexts are bound, so groups are reused across stacks and environments even when concrete successors differ. Objective-specific score aggregation and token restoration provably preserve both the Viterbi optimum and the token-labeled path distribution of ungrouped inference. Across three benchmarks and two DLM families, DConstrain compiles 100% of BFCL schemas under arbitrary property orders, whereas an optimized finite-state baseline compiles only 71.30-85.81%. Under arbitrary property order, it reduces BFCL end-to-end decoding latency by 6.65-46.26 times with 100% schema validity and raises DiffusionGemma's BFCL accuracy from 37.27% under fixed order to 81.56%.

cs.AI↗
arXiv2026

A parametrization of nonassociative Petit algebras $K[t;σ]/K[t:σ](t^m-a)$ up to equivalence

Let $K/F$ be a cyclic Galois field extension of degree $n$ with $\mathrm{Gal}(K/F)=\langleσ\rangle$. The nonassociative Petit algebras $K[t;σ]/K[t;σ](t^m-a)$ have recently emerged as important tools in coding theory and cryptography, where the applications only require distinct algebras up to isomorphism. Algebras that are isomorphic via a monomial isomorphism of degree one are called equivalent. We present a novel and precise parametrization of the $F$-equivalence classes of these algebras in the key case that $F$ contains a primitive $n$th root of unity and $n\mid m$. We evaluate the parametrization explicitly when $F$ is a nonarchimedean local field, which has applications to the number field case recently used to construct a nonassociative variant of Learning with Errors.

math.RA↗
arXiv2026

Stability of rotation axis of galactic disks in HESTIA suite

We have found that in approximately 1/4 of Milky Way or M31 analogues in the HESTIA suite of simulations there is a significant, $>45^\circ$, change in the axis of rotation orientation of the gaseous disk during the last 6 Gyr. In 6 out of 7 such cases the change is likely caused by the accretion of individual satellites, while in one case such a connection is not obvious. All galaxies that underwent a significant change in rotation axis direction also had angular momentum amplitudes below the sample median 6 Gyr ago. Our findings support the idea that last 6~Gyr was a quiet phase of evolution in the LG. On the other hand, the mechanism we detect, that satellites can flip galaxy orientation, can explain the difficulties in identifying correlations between the rotation axis and the orientation of large-scale structure elements.

astro-ph.GA↗
arXiv2026

Low-Latency MoE Inference in Space Data Centers: A Wireless-Power-Transfer Perspective

Space data centers (SpaceDCs) are emerging as platforms for large language models (LLMs), supported by continuous solar-energy harvesting and high-speed interconnections. The mixture-of-experts (MoE) architecture supports distributed LLM inference by deploying experts across satellites. However, skewed expert activation creates uneven computation power demands despite uniform solar-power supply at the dawn-dusk sun-synchronous orbit (SSO). Consequently, active experts remain power-constrained while harvested power at idle experts is underused. These reduce inference efficiency and increase latency. To address this supply-demand mismatch, we propose the SpaceWPT framework, which uses inter-satellite laser-based wireless power transfer (WPT) to redistribute harvested power. The key design challenge is to balance the computation demands of power-donating and power-receiving satellites under WPT losses: transferring more power accelerates computation at the receiver but leaves less power for the donor. After deriving design principles that simplify WPT control, we characterize the resulting latency tradeoff through a tractable surrogate of average inference latency. Minimizing the surrogate yields a novel two-sided water-filling strategy that determines whether each active expert donates, retains, or receives power according to the water level and its WPT link efficiency. Simulations with Mixtral-8x7B and Phi-3.5-MoE show up to 29.25% lower inference latency than without WPT, with larger gains under more compact formations and higher transfer efficiencies.

cs.DC↗
arXiv2026

REGATTA: A Cross-Domain Measurement Framework and Fault-Injection Harness for Enterprise AI Agents

Enterprise AI-agent evaluation needs evidence about whether test suites expose specified faults and whether their reference rules are trustworthy. We introduce REGATTA (Regulatory Evaluation of Agents via Two-stage Testing and Assurance), a measurement framework with three instruments. A 125-item source-cited checklist spans five industry-by-jurisdiction cells across the United States and Vietnam. A 25-fault corpus covers five categories, and a harness scores suite-level fault detection at the design and execution stages. We examine four measurement diagnostics on 180 generated suites (5,400 scenarios) from three LLM generators. A source audit finds 24 items misstating the rule in force, 46 additional citation defects, 31 items with only non-binding, unidentifiable, or unverifiable sources, and 24 verified clean. Only initially substance-flagged items received author review; all 24 clean rows received AI review only. Results therefore describe the instruments as written, not legal compliance. Item coverage varies, execution-stage fault profiles are stable across generators, and the historical stages agree on $66$--$69\%$ of fault--suite pairs. A current-model US-insurance retest finds stage disagreement above repeated-execution variation, but stage agreement $.536$ falls below pooled chance $.549$, failing V3's lower inequality. Blind adjudication by one expert supports the design-stage judging protocol ($κ=.72$, 95% interval $[.52,.88]$); 40 current-model execution probes give $κ=.75$. The instruments, audit, and reusable code are released.

cs.SE↗