Peer Revenue Play credits. Real papers.

Papers

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

All of arXiv

Tuesday 8 September

Cryptography and Security

Safety in Self-Evolving Agents: A Survey

Large language models (LLMs) exhibit strong general capabilities, yet their parameters typically remain fixed after deployment, limiting learning from new interactions. In open-ended environments, this motivates self-evolving agents that continually update reusable state-including model parameters, memories, tool definitions, skills, and workflows-from data, feedback, and accumulated experience. This shift changes the safety problem: once experience becomes reusable state, past events become future causes, and information harmless in one context may later influence decisions with greater persistence, authority, or scope. Self-evolving agent safety therefore asks not only whether a response is aligned or an action authorized, but whether safety properties survive the accumulation, generalization, and cross-context reuse of locally useful experience. We introduce SAVER, a transition-centered framework in which Substrate locates reusable influence, Adaptation captures how it changes, Violation identifies compromised safety attributes, Exposure marks where failures become observable, and Response assesses containment, repair, or revocation. Our survey reveals that failures need not originate from harmful information: legitimate state can become unsafe when adaptation expands its persistence, authority, or scope beyond the conditions under which it was valid. Existing work provides comparatively strong evidence for admission, retrieval, activation, exposure, and local containment, but much less for descendant repair and evaluation after adaptation resumes. We therefore argue for longitudinal evaluation that traces unsafe influence to its originating transition, verifies repair across descendants, and tests whether it can re-emerge under continued evolution.
Read the full abstract

All authors Jiahao Chen, Zhou Feng, Oubo Ma, Yichen Yan, Ruixiao Lin, Hangtao Zhang, Linkang Du, Hengyu An, Yong Yang, Jun Liu, Junhao Li, Naen Xu, Chunyi Zhou, Yuan Su, Zehao Jin, Qianli Ma, Leyi Qi, Yiming Wang, Zhe Ma, Yuwen Pu, Mengyao Du, Yuanyi Song, Enhao Huang, Zhihui Fu, Jun Wang, Jinfeng Li, Yuefeng Chen, Hui Xue, Yiming Li, Tianyu Du, Shouling Ji

Computation and Language

BudgetSchemaBench: A Budget-Swept Diagnostic for Schema Context in Text-to-SQL

Data agents over structured sources must fit database schema into the model's context window. Large catalogs can span many databases and thousands of columns, so cost constraints may require choosing between table coverage and serialization detail well before the context window is full. We introduce BudgetSchemaBench, an execution-grounded diagnostic for this setting. Its construction derives relevance labels mechanically from gold SQL, without human- or LLM-authored ground truth. Using a pooled 80-database catalog, we sweep four schema-context budgets and compare three representations while keeping each retriever's table ranking fixed. A source-namespace check rejects queries that obtain the correct result from the wrong database. The evaluation covers three conditions: end-to-end retrieval; frozen-gold, in which the required tables are guaranteed; and a probe that removes those tables. For the primary solver with raw serialization, raising the budget from 2.5% to 50% of the catalog improves execution accuracy on 1,279 held-out questions by 18 percentage points under lexical retrieval but only 3 under dense retrieval; the dense retriever already finds most required tables at the smallest budget. When the required tables are removed, 94.6% of correct predictions name one of them exactly, consistent with reconstruction of absent schema from parametric knowledge. For the two main solvers in the frozen-gold condition, the three representations differ by at most 2 percentage points, and the widest paired 95% confidence interval bounds the difference within +/-4 points. We observe the same qualitative patterns with one reasoning model from a different family. When retrieval is coverage-limited, execution accuracy is more sensitive to the schema budget than to the tested serializations. The diagnostic and the code used to construct and evaluate it are publicly available.
Read the full abstract

Data Analysis and Statistics

A State-Transition Information Space for Time-Series Dynamics: Theory and Application

Characterizing dynamical organization in time series requires distinguishing the diversity of accessible states from uncertainty in their temporal transitions. Here we introduce a state-transition information space based on two normalized entropy measures derived from ordinal patterns: K_q, quantifying ordinal-state diversity, and K_t, quantifying transition uncertainty. Their joint K_t-K_q representation provides a two-dimensional framework in which dynamical regimes and their temporal evolution can be examined. Theoretical properties establish the bounded relation 0 <= K_t <= K_q <= 1, while canonical time series identify distinct empirical domains ranging from ordered dynamics to near-maximal randomness. Sliding-window analysis further shows that systems can exhibit temporal trajectories through the K_t-K_q plane rather than remaining at a fixed dynamical state. Application to 1,879 monthly naturalized streamflow records from the U.S. rivers shows that river dynamics occupy a distinct intermediate region between ordered and highly disordered regimes. Moreover, river positions shift systematically toward higher K_t and K_q with increasing drainage area, revealing scale-dependent organization of streamflow dynamics. The framework therefore provides a compact means of comparing state diversity, transition uncertainty, and their temporal and spatial organization across time-series systems.
Read the full abstract

Computation and Language

A Mechanistic Study of AI-Text Detection Neurons in Frozen BERT: Sparse Probing and Activation Patching on RAID

AI-generated text detectors achieve high accuracy on standard benchmarks, yet the internal representations that drive these predictions remain poorly understood. We study which neurons in a frozen BERT-base-uncased encoder support AI-text detection, using the RAID benchmark across six generators spanning pure-base and instruction-tuned models. We apply the L1-to-L2 sparse-probing protocol of Gurnee et al. (2023) to all 9,216 CLS hidden-state dimensions (12 layers x 768), which we call neurons. The procedure recovers a stable set of under 1% of neurons per generator, consistent across folds and seeds; a probe restricted to that set retains most of the full-feature detection accuracy. Bidirectional activation patching confirms this set's causal relevance: in both directions it flips predictions an order of magnitude more often than size-matched random sets. Mean-ablating the same neurons leaves accuracy largely intact; the signal is therefore redundantly distributed. Cross-generator analysis reveals a bipartite structure: instruction-tuned generators concentrate 30-36% of stable neurons in BERT's final layer while both base generators fall below 14%, consistent with a layer-12 footprint of post-training alignment. Leave-one-family-out evaluation shows the selected neurons retain 86-94% of the full-feature ceiling on unseen generator families, so a detector can operate on a small fixed subspace without re-identifying neurons per generator.
Read the full abstract

Machine Learning

When Does Advection-Aware Graph Nowcasting Help? A Controlled Study of Distributed Solar Ramp Forecasting with a Self-Supervised Cloud-Motion Estimator

Short-term forecasting of cloud-induced power ramps across a network of distributed photovoltaic (PV) or irradiance sensors is a recognised pain point for grid operators. A natural idea is to make the graph neural network (GNN) advection-aware: connect each site to the sites upwind of it, with edge time-lags set by the cloud-motion vector (CMV), so that a ramp is propagated forward before it physically arrives. Using a controlled synthetic testbed with a known wind field, we show that (i) with a realistic cross-correlation CMV estimate, an explicit advection graph does not beat a plain static or learned-adjacency spatiotemporal GNN; (ii) roughly half of the benefit available from a perfect CMV comes simply from providing an accurate motion vector as an input feature, not from graph structure; and (iii) advection helps only when the advective displacement over the forecast horizon, v*H, fits inside the sensor network. Motivated by (ii), we introduce a small self-supervised cloud-motion estimator – a position-aware encoder trained only on a multi-lag optical-flow reconstruction objective with an annealed kernel – that recovers the true wind vector to 2-4 degrees median angular error, 2-4x better than the classical cross-correlation method across every wind regime. Freezing this estimator and feeding its vector to the forecaster closes about 60% of the oracle-CMV RMSE gap at moderate wind (8-15% RMSE reduction over no advection), with no external wind data. We also report a negative result for a spatially-coherent probabilistic head. All claims are established on a single synthetic simulator; we discuss why real-network validation is the necessary next step and outline it.
Read the full abstract

Algebraic Topology

Complex and Stable Complex Structures on Real Vector Bundles over Connected Sums of Quaternionic Projective Planes

Let $M_{\ell,m}=\ell\,\mathbb{HP}^{2}\# m\,\overline{\mathbb{HP}^{2}}$ be the connected sum of $\ell$ copies of $\mathbb{HP}^{2}$ with $m$ copies of $\mathbb{HP}^{2}$ endowed with the opposite orientation. We characterise, in terms of Pontryagin classes, the real vector bundles over $M_{\ell,m}$ admitting a stable complex structure, and deduce that $M_{\ell,m}$ is stably almost complex if and only if $\ell-m$ is even. We then determine when an oriented real vector bundle of even rank over $M_{\ell,m}$ admits a complex structure. Consequently, $M_{\ell,m}$ admits an almost complex structure if and only if $\ell=2m+1$ and $m$ is odd. This corrects an assertion made by Sato and Suzuki in 1974, according to which $M_{\ell,m}$ is never almost complex. We also identify the unjustified step in their argument.
Read the full abstract

Computation and Language

ChainDoRA: Tensor-Train Factorized Weight-Decomposed Low-Rank Adaptation for Parameter-Efficient LLM Fine-Tuning

Parameter-efficient fine-tuning (PEFT) adapts large language models (LLMs) to downstream tasks while updating only a small fraction of their pretrained parameters. Low-Rank Adaptation (LoRA) uses two trainable low-rank matrices, while Weight-Decomposed Low-Rank Adaptation (DoRA) further separates weight magnitude and direction but retains the dense LoRA-style factorization in its directional branch. We propose ChainDoRA, a weight-decomposed adaptation framework that constructs the directional low-rank factors from a connected Tensor-Train (TT) chain, where the adapter rank forms the boundary rank between input- and output-side TT contractions and an independent TT rank controls representation capacity and parameter cost. Under a controlled 15,119-example response-only adaptation setting with LLaMA-7B, ChainDoRA is evaluated against matched LoRA and DoRA baselines on seven commonsense reasoning benchmarks. ChainDoRA with TT rank 16 achieves a seven-task average accuracy of 72.30%, compared with 69.88% for LoRA and 69.39% for DoRA, while requiring only 5.35M trainable parameters versus 56.10M for LoRA and 56.98M for DoRA, corresponding to a 90.62% reduction relative to DoRA. Ablations over TT rank and adapter placement show controllable parameter-accuracy trade-offs, indicating that connected TT parameterization can substantially reduce the parameter cost of magnitude-direction adaptation while preserving, and in this setting improving, downstream reasoning performance.
Read the full abstract

Signal Processing

Physics-guided deep metric learning with continuous time embeddings for open-world radar pulse de-interleaving

Radar pulse de-interleaving is a foundational Electronic Support Measures (ESM) task that aims to separate chronologically interleaved pulse streams from multiple non-cooperative transmitters under unknown emitter cardinality in dense, contested electromagnetic environments. Classical histogram transforms and closed-world deep classifiers degrade under severe pulse loss, agile Pulse Repeti tion Interval (PRI) modulation, and spurious clutter. In this paper, we systematically characterise continuous temporal representations and physics-guided model selection in deep metric learning for open-world radar de-interleaving. Building on the transformer-based metric-learning framework for open-world deinterleaving introduced by Gunn et al. [1], we introduce a continuous Time-of-Arrival (ToA) sinusoidal positional encoding that directly models physical inter-pulse durations rather than ordinal token indices, a design choice that contrasts with [1], who found ordinal positional encodings provided no benefit and omitted them entirely. Neural network parameters are optimised solely via Supervised Contrastive (SupCon) learning, while scale-aware physical domain priors based on PRI Consistency and Angle-of-Arrival (AoA) continuity serve as physics-guided validation and checkpoint-selection criteria operating on unsupervised HDBSCAN cluster assignments.
Read the full abstract

Computation and Language

Graph-Based Inference for Feedback-Driven Word Deduction: A Scalable Framework for the Jotto Problem

A feedback-based word deduction framework based on the Jotto problem is proposed, and the problem space is represented as a weighted graph where all valid words correspond to nodes, and the edge weight is defined by the number of common letters between the two words. Finally, the gameplay is defined as an iterative constraint propagation mechanism where feedback is used to iteratively narrow the incompatible space of the graph, facilitating the reduction of the hypothesis space in a structured and interpretable manner. In contrast to existing approaches, where the problem space is typically defined for fixed-length isograms, the proposed framework generalizes to variable-length words (between 3 and 8 letters) and naturally extends to repeated letter cases, facilitating the treatment of realistic Jotto problem instances within a unified framework for the first time. The proposed framework's applicability and solver dynamics are also discussed through an interactive implementation and a qualitative case study, respectively. Significant automated tests on approximately 3,000 simulated gameplay scenarios identify a novel convergence behavior: the expected number of iterations diminishes with increasing word length. A strong relationship is confirmed using statistical tests to verify a logarithmic relationship, which is also verified using regression modeling and goodness-of-fit tests. In addition to the initial problem statement, this formulation introduces graph pruning as a viable paradigm for feedback-driven inference with interpretability and its association with symbolic reasoning and interactive intelligent systems.
Read the full abstract

Computation and Language

ICDAR2026 Competition on Multimodal Reasoning over Documents in Multiple Domains

In this report we present results of the ICDAR2026 Competition on Multimodal Reasoning over Documents in Multiple Domains. This competition aimed to advance research in document understanding through the task of Visual Question Answering (VQA). Building upon previous DocVQA benchmarks, this competition introduces challenging reasoning questions over a diverse collection of documents spanning eight domains, including business reports, scientific papers, slides, posters, maps, comics, infographics, and engineering drawings. The competition concluded with 20 valid submissions from 8 teams spanning zero-shot VLMs, OCR and parser-augmented pipelines, agentic retrieval systems, multi-agent ensembles, and fine-tuned multimodal models. The results show that the strongest systems move beyond single-pass prompting and instead rely on structured evidence extraction, retrieval, verification, and orchestration across multiple components.
Read the full abstract

All authors Artemis Llabrés, Marc Serra Ortega, Tomàs Ockier, Samuel Ortega Cuadra, Amritpal Singh, Christos Georgakilas, Andrey Barsky, Ernest Valveny, Dimosthenis Karatzas

Computer Vision

Did You Steal My Shot? Pioneering Camera Motion Plagiarism Detection in Generative Videos

Camera motion often reflects directorial intent and requires professional equipment, making it a high value form of intellectual property. However, generative video models can imitate such high value camera motions with simple prompts, while existing similarity detection methods mainly operate on visual content and fail to capture deeper motion similarity. This is mainly because their training data entangles camera motion with visual content. Moreover, traditional optical flow is insufficient to represent complex camera motions. We therefore build the first benchmark for camera motion analysis, including a motion dataset with 11 motion styles and evaluation protocols. Furthermore, we propose a motion representation that augments optical flow with vorticity cues from fluid dynamics, thereby better capturing motions. Experiments show that our detector achieves a 3.02* improvement in plagiarism detection over the strongest baseline and remains effective on generative videos. We believe our work extends copyright protection beyond static content to dynamic camera motion.
Read the full abstract

Signal Processing

Dynamic Mode Decomposition by Tensor Evolution for Source Depth

Depth estimation of shallow acoustic sources using a deep-ocean near-bottom vertical line array relies on depth-sensitive interference, which appears as an oscillatory structure in the frequency-angle domain in broadband matched beam-intensity processing (MBIP). For low-SNR, the nonlinear feature extraction in MBIP transforms array-domain noise into feature-domain distortions, making depth estimation fragile. Addressing this issue, we propose a tensor evolution-based depth estimation (TEDS) method inspired by dynamic mode decomposition (DMD), which interprets a sequence of broadband MBIP surfaces as the output of a dynamic system from a time-varying autoregressive operator. The operator is represented by a low-rank Tucker decomposition, which separates coherent feature modes and their temporal modes from incoherent noise. The target depth is inferred via Fourier summation applied to a dominant feature mode. Deep-ocean numerical experiments demonstrate that TEDS improves robustness at low-SNR, consistently outperforming matched field processing and broadband MBIP.
Read the full abstract

All authors Wenqian Wu, Xiaohong Yang, Guangying Zheng, Lei Cheng, Peter Gerstoft

Algebraic Topology

A Tale of Two Paths

We provide a new presentation for a cylinder object inside of the category of co-globular $\infty$-groupoids. We prove that the cylinder construction that Lanari provides is isomorphic to the one we build in this paper.

Algebraic Topology

Binary Minor Certificates for the $\mathcal A(1)$-hit problem of $MO(2)$

Let $M=\widetilde H^*(MO(2);\mathbb F_2)$ and let $\mathcal A(1)$ be the subalgebra of the mod $2$ Steenrod algebra generated by $\operatorname{Sq}^1$ and $\operatorname{Sq}^2$. In this work, we determine the quotient $\mathbb F_2\otimes_{\mathcal A(1)}M$ in every internal degree $h\geq0$, where internal degree $h$ corresponds to cohomological degree $h+2$. To overcome the limitations of finite experimental extrapolation, our proof is strictly certifying: in each degree, we exhibit a determinant-one minor of the binary hit matrix and a complementary family of linear dual functionals annihilating every hit column, mathematically guaranteeing the exact rank for all degrees. This rigorously establishes the monomial generating families and module decomposition proposed in Shih's 2012 report. Furthermore, we derive a closed binomial formula for every Steenrod square from the length-two symmetric-polynomial model, correct a boundary omission in the $\mathcal A(0)$ calculation, compute the exact Hilbert series of the $\mathcal A(1)$-cohit, and prove an intrinsic direct-sum decomposition of $M$ into two stable $\mathcal A(1)$-submodules.
Read the full abstract

Astrophysics of Galaxies

SPHEREx 3.3 Micron Aromatic Emission in the Magellanic Clouds: Separating Dust Column, Excitation, and Environmental Suppression

Bright emission in the 3.3um unidentified infrared band, often described as an aromatic infrared band, can trace dust column, radiative excitation, or the survival and emissivity of its carriers. We combine SPHEREx maps of the Magellanic Clouds with HERITAGE dust columns, a TIR heating proxy, Br\alpha, H\alpha, H II-region catalogs, and SAGE 8um imaging. We measure continuum-subtracted 3.3um feature excess, X_{3.3}, and Br\alpha excess, X_{Br\alpha}, in matched cells. In each Cloud, a quiescent diffuse baseline depending on dust surface density, \Sigma_d, and TIR/\Sigma_d predicts X_{3.3} with R^2=0.628 in the LMC and 0.288 in the SMC. Relative to this expectation, LMC H II interiors are under-luminous by -0.223^{+0.035}_{-0.040} dex; local-background measurements give -0.153^{+0.014}_{-0.043} dex despite a total-X_{3.3} enhancement of +0.749 dex. SMC H II-region residuals are consistent with zero over its lower-X_{Br\alpha} regime. A matched comparison with stellar-subtracted 8um dust intensity, I_{8,dust}, shows that column and heating predict I_{8,dust} more tightly than X_{3.3}, and that the high-X_{Br\alpha} LMC deficit is stronger for X_{3.3} than for I_{8,dust}. The deficit is therefore selective to the isolated 3.3um feature, not uniform in all aromatic-band-dominated emission. This result suggests that the small aromatic/aliphatic carbonaceous carriers traced by the isolated 3.3um feature are more susceptible to processing in active regions than the larger or differently emitting carbonaceous-grain population dominating the 8um aromatic complexes.
Read the full abstract

General Relativity and Quantum Cosmology

Non-equilibrium thermodynamics of an expanding Swiss-Cheese braneworld cosmology with a matter bounce

We investigate the non-equilibrium thermodynamics of a Swiss-cheese (SC) braneworld universe. By employing Hayward's unified first law and the Clausius relation at the apparent horizon, we first show that the standard equilibrium treatment of the effective SC braneworld fluid, together with the usual Hawking temperature, reproduces the Bekenstein-Hawking area law without any braneworld correction to the entropy. This demonstrates that the brane quadratic energy density corrections cannot, by themselves, generate a modified entropy area relation within the equilibrium framework. We therefore formulate a non-equilibrium thermodynamic description in which deviations from the Bekenstein-Hawking entropy are accompanied by an internal entropy production term. For a general horizon entropy functional, we derive the corresponding entropy production rate and then specialize to a logarithmic and inverse area quantum-corrected entropy. The resulting irreversible contribution is expressed in terms of the Hubble parameter and its derivative. We apply the formalism to a nonsingular matter-bounce solution in the SC braneworld. During the expanding phase, the total entropy production remains positive satisfying the GSLT. Moreover, the late-time behavior of the second derivative of the total entropy becomes negative, indicating a decelerating entropy growth and an asymptotic approach toward thermodynamic equilibrium. Far from the bounce, the equation of state parameter $w \approx -1$, indicating a dark energy-dominated regime. Significant departures from $w = -1$ are confined to the bounce vicinity, where high energy braneworld corrections drive the nonsingular transition. Thus, the model naturally transitions from a high energy bounce into an asymptotic dark energy phase during late-time expansion.
Read the full abstract

Atmospheric and Oceanic Physics

Scale analysis for the Madden$-$Julian oscillation

This study performs a scale analysis for the dynamics of equatorial flow systems like the Madden$-$Julian oscillation (MJO). This analysis makes three assumptions about the geometry of the flow and focuses on the leading order terms of the mass continuity and momentum equation system. First, the flow is meridionally narrow. Second, the zonal structure is planetary wavenumber 1. Last, the flow is vertically shallow. When constrained by these geometric assumptions, the system shows a time scale much longer than a day depending on the meridional narrowness. Specifically, for equatorial Kelvin waves, the oscillation period ranges from 36 to 64 days given a ratio of the planetary radius to the meridional width scale from 6 to 8 (i.e., an e-folding latitude from 9.59$^\circ$ to 7.18$^\circ$ either side of the equator). Given the long, narrow, and shallow geometry, the dynamics can sustain Kelvin waves only at low frequencies. This result suggests that the intraseasonal time scale of the MJO arises from the Kelvin wave frequency limit due to the planetary wave geometry. This planetary intraseasonal scale relation is independent from thermodynamics, and an open question remains on what physical processes coordinate the thermodynamics in the same length or time scale.
Read the full abstract

Hardware Architecture

Grouped Value Attention: Efficient KV Caching via On-Demand Key Reconstruction

The KV cache is a primary bottleneck for Transformer decoding: its memory footprint and cache-read traffic grow with sequence length. Grouped-query attention (GQA) reduces this cost by sharing key-value heads, but still stores both a key and a value at every step. We introduce Grouped Value Attention (GVA), which stores grouped values and reconstructs content keys with a learned linear map. At inference, the map can be absorbed into the query, eliminating the need to materialize content keys in the intended decode path. A small shared decoupled RoPE channel retains positional information through a separately cached positional key. For the configurations studied, this representation reduces persistent cache scalars by approximately 45-47% relative to matched GQA. At the 350M-parameter scale with 30B FineWeb-Edu tokens, the 16-dimensional positional variant reaches 44.18 average accuracy across five tasks, compared with 44.36 for GQA and 43.88 for MLA. These results demonstrate near-GQA benchmark accuracy with a more compact cache representation. To translate this compact representation into faster autoregressive inference, we have developed custom decoding kernels and are currently evaluating their end-to-end inference performance with an open-source release planned soon.
Read the full abstract

Logic in Computer Science

Decoupling Evidence Sources in Paraconsistent Logic: A Generalized PAL2v Framework for Educational Assessment

Paraconsistent Annotated Logic with two values (PAL2v) annotates a proposition with a pair $(\mu,\lambda)$ of degrees of favorable and unfavorable evidence, drawn from a bilattice whose four extreme points are the values of Belnap and Dunn. In applications, however, the second coordinate is almost always taken to be the complement of the first, and several such annotations are then fused. We show that this practice determines the sign of the degree of contradiction in advance, independently of the data: for aggregation functions $F$ and $G$ the fused degree of contradiction equals $F-G^{d}$, where $G^{d}$ is the De Morgan dual of $G$, so that it vanishes identically when the two are dual, and has a constant sign otherwise. In particular the componentwise minimum used throughout the literature yields only consistent annotations, and the conjunction of annotated logic only exact ones; five of the twelve logical states, contradiction among them, are essentially unreachable by construction. We also give a decision procedure for the twelve states that is provably total, in place of the inequality lists in current use, which assign two states to about half of the non-extreme region. We then propose a construction in which the two coordinates are computed from disjoint families of evidence, show that every logical state is attainable. Educational assessment provides a natural instance, since summative and formative instruments are institutionally distinct: mapping examinations to $\mu$ and disengagement to $\lambda$ makes the degree of contradiction equal to the gap between what a student demonstrates and what that student invests. Simulations over three cohort profiles illustrate the classification.
Read the full abstract

All authors Arthur Nakamura, David Drummond, Welbert Pereira, Gustavo Lahr, Kaique Matias de Andrade Roberto

Computer Vision

Multimodal-Multiresolution Foundation Model for Lunar Remote Sensing

We present a multimodal foundation model for lunar remote sensing, pretrained from scratch on SomBench, a geographically partitioned corpus of nearly two million co-registered tile bundles spanning 11 modalities at two spatial scales (1 m/pixel and 100 m/pixel). The model adapts the TerraMind masked-token architecture with two lunar-specific extensions: acquisition geometry is provided as explicit context, and meter- and hundred-meter-scale tiles are trained jointly so that a single set of weights covers both resolutions. FlexiViT patch embeddings allow adaptation to different patch sizes without retraining, while modality-wise inputs enable flexible multimodal fine-tuning. Qualitative generation experiments suggest the model learns meaningful cross-modal correspondences, including terrain derivatives from elevation and illumination-consistent reflectance from geometry. We evaluate on four benchmarks: crater detection at WAC and NAC scales, irregular mare patch (IMP) segmentation, and polar ice prospectivity regression. Across tasks, the pretrained model matches or outperforms ImageNet-pretrained baselines and an architecturally identical random-init control. On multimodal ice prospectivity regression, pretrained variants achieve the best results, while the random-init model outperforms most baselines, suggesting gains arise from both the architecture and pretraining. Label efficiency is notable for WAC crater detection, where the pretrained model trained on 50% of the data exceeds the strongest ImageNet baseline trained on the full dataset. Among adaptation strategies, LoRA matches or surpasses full fine-tuning on crater detection and IMP segmentation while using far fewer trainable parameters, whereas full fine-tuning performs best for ice prospectivity regression. We release the pretrained checkpoint, benchmark datasets, and fine-tuning code to support reproducible lunar AI research.
Read the full abstract

All authors Paolo Fraccaro, Gabby Nyirjesy, Daniela Szwarcman, Himanshu Patil, Vishal Gaur, Rohit Lal, Rachel A. Slank, Geoffrey Dawson, Hiyam Debary, Michael K. Barker, Andrew Annex, Vishnu Viswanathan, Zachary Morse, Ethan I. Schaefer, Nikolaos Dionelis, Ankur Kumar, Campbell D. Watson, Manil Maskey, Rebekah I. Dawson-Rigas, Juan Bernabé-Moreno, Rahul Ramachandran, Sujit Roy

Computer Vision

GaugeDefect: Detecting Surface Anomalies by Curvature of Feature Transport

Industrial anomaly localization has advanced rapidly with feature-based, reconstruction-based, and distillation-based methods. Most of these methods score a region by asking how unusual its local appearance or feature representation is with respect to normal training images. This is a strong and practical formulation. In this work, we study a complementary geometric cue for cases where an abnormal region may still contain locally plausible visual features. Thin scratches, small dents, and disrupted repeated patterns often do not make every local patch individually abnormal; instead, they disturb how nearby features vary and connect across the surface. We propose GaugeDefect, a geometric method for surface anomaly localization based on the curvature of feature transport. Given a feature lattice, we estimate a local feature frame at each node and compute orthogonal transports between neighboring frames. The accumulated transport around a small closed loop gives a holonomy matrix, whose deviation from identity measures feature-transport curvature. After calibration on normal training images, unusually large curvature indicates a local inconsistency in the feature field. The curvature here is not the physical curvature of the inspected object, but a representation-space measure of neighborhood inconsistency. This makes the method applicable to curved surfaces, textured materials, and non-planar industrial objects. Its main role is to improve localization of subtle surface disruptions, while often producing sharper responses near defect boundaries as a natural consequence of the curvature signal.
Read the full abstract

Geophysics

Large-scale bioacoustic detection using semantic segmentation: a deep learning framework applied to fin whale calls in ocean-bottom seismometer recordings

Ocean-bottom seismometers (OBS), originally deployed for geophysical research, continuously record low-frequency sound for months to years across broad areas of ocean, offering a largely untapped resource for passive acoustic monitoring (PAM) of baleen whales. Realising this potential requires automated detection methods that operate reliably across the varied conditions in large sensor networks. We present a deep learning semantic segmentation framework that detects the 20-Hz notes of fin whales (Balaenoptera physalus) in OBS spectrograms, assigning each pixel a probability of belonging to a call and converting the resulting probability maps into time-frequency bounding boxes describing individual detections. We trained the model on hydrophone data from one OBS deployment in the Azores-Madeira-Canaries region and applied it without retraining to vertical-component seismometer data from a second, geographically distinct deployment, showing that a single trained model generalises across sensor types and recording environments. Applied to 378,912 h of recordings from 46 OBS sites, the detector identified 6.3 million calls, forming the largest fin whale call catalogue assembled to date, with high precision (~97%) across both deployments. The resulting catalogue resolves call timing and spectral structure accurately enough to support ecological analyses, revealing coherent seasonal shifts in three persistent inter-note interval (INI) groups across the singing season and basin-scale patterns in calling activity. By transforming existing geophysical infrastructure into a scalable sensing network, our approach substantially expands the spatial and temporal reach of PAM without new hardware investment, offering a transferable framework for tracking other low-frequency vocalising species and informing conservation planning, marine spatial management, and abundance estimation across large scales.
Read the full abstract

All authors Jocelyn Japnanto, Alex A. Saoulis, Miriam Romagosa, Rita Leitão, Gabrielle Arrieta, Mónica A. Silva, Matthew Graham, Ana M. G. Ferreira

Computer Vision

RxScribe Bench: A Multi-Axis Benchmark for Evaluating Vision-Language Models on Indian Outpatient Prescriptions

Prescription transcription errors are not interchangeable. A model that fabricates a drug and a model that misreads a legible dose pose very different clinical risks, yet prescription-transcription accuracy is typically reported as a single blended figure that treats the two as equivalent. We introduce RxScribe Bench, a benchmark for evaluating vision-language models on handwritten prescription digitization that decomposes performance into four axes tied to clinical severity, rather than folding everything into a single aggregated score. Given only a prescription image and an output schema, a model produces a structured record, which is then compared field by field against a human-authored ground truth of identical shape, with each field also labeled for visibility and legibility. The four axes isolate distinct failure modes, namely Correctness, Hallucination, Engagement, and Robustness. The Robustness axis withholds its hard-field results rather than reporting one when the supporting sample falls below a minimum-evidence threshold. We evaluate frontier vision-language models on real prescriptions across independent cold runs per image, and find that no single model wins across all four axes.
Read the full abstract

All authors Somil, Vijay Saini, Vidit Verma, Riya, Aastha Batta, Vibhuti Malhotra, Chayan Khetan, Piyush Mittal, Puneet Poonia

Computer Vision

The MODA General Attribute Suite: A Four-Track Evaluation Benchmark for Fashion Attribute Extraction

Fashion attribute extraction is evaluated inconsistently: results are reported as single aggregate numbers across image types that pose different problems, fields that are not visible in an image are scored as ordinary negatives, and the effect of vocabulary mismatch between datasets is acknowledged but not measured. We release the MODA General Attribute Suite, a four-track benchmark that keeps these problems separate by construction. Each track (localized garment crops, catalogue product images, applicability-aware full-body photographs, and product text) carries its own frozen test set, input contract, metric, and leakage unit, and the tracks are never averaged. The protocol requires label-blind prediction, SHA-256 commitment of prediction files before any label is opened, a fail-closed scorer, and a 10,000-sample paired cluster bootstrap at each track's natural leakage unit; promotion requires a positive interval on every track rather than a favourable mean. We release the scorers, the split builders, our own prediction files with their hashes for every image-track system in the main results and for runs we lose, and the MODA_NER(V) model checkpoints for three of the four tracks. The text model is not distributed; its benchmark is. We report baseline results, six interventions that did not improve them, and the limitations of the evidence. On the garment-crop track we also measure how far the headline depends on the benchmark's own design: the order of the two strongest systems depends on whether fields are micro- or macro-averaged, two thirds of the cells the protocol judges carry no annotation, and disabling the applicability decision cuts attribute micro-F1 from 0.63 to 0.27, so most of that score rests on absence decisions the source labels cannot adjudicate. This version also corrects one figure in an external comparison that an inference preprocessing defect had affected.
Read the full abstract

Chaotic Dynamics

Tensor-Based Chaotic Perception

Symbols can provide compact representations of complex sensory information. Mathematically, chaotic attractors can provide a basis for such representations, with object identity encoded in the topological organization of the attractor. In three-dimensional phase space, this organization can be characterized by the linking matrix of unstable periodic orbits, which serves as a topological fingerprint. Because the linking matrix has no closed-form dependence on the learnable parameters, we learn the chaotic series generating the target attractors rather than optimizing the invariant directly; the reconstructed attractor then determines its unstable periodic orbits and linking matrix through its topology. In this paper, we propose a third-order tensor that maps inputs to connection matrices that generate the corresponding chaotic series. Exploiting the duality between a vector and its sequential representation as a series, we lift inputs into a higher-dimensional space and reorder their components, allowing the input to be presented sequentially rather than simultaneously. We first demonstrate that this representation can discriminate between classes in the \href{https://archive.ics.uci.edu/dataset/151/connectionist+bench+sonar+mines+vs+rocks}{Sonar} dataset. We then examine perceptual constancy using the \href{https://www.csc.kth.se/cvap/databases/kth-tips/index.html}{KTH-TIPS2-a} dataset, where variations in color and texture within a class are mapped to distinct chaotic series associated with the same topological attractor class. The results demonstrate that the proposed tensor-based mapping can preserve attractor class while accommodating variations in the sensory realization of an object.
Read the full abstract

Computer Vision

SomBench: Benchmark Dataset for Advancing Machine Learning in Lunar Science

Lunar orbital missions, such as Lunar Reconnaissance Orbiter, Kaguya/SELENE, Gravity Recovery and Interior Laboratory, and Lunar Prospector, among others, provide rich multi-instrument observations, but their heterogeneity in sampling, projection, and conventions limits reproducible machine learning (ML). We introduce SomBench, a unified, spatially-aligned, ML-ready lunar dataset aggregating 30+ co-registered layers from ten instruments across four missions, spanning 1 meter to 20 kilometer/pixel and covering 82 degree latitude in 90 Lunar Transverse Mercator zones with two polar stereographic caps. An image-anchored tiling pipeline yields pretraining-ready multimodal tile views with leakage-safe splits, distributed as netCDF with Parquet catalogs. An application benchmark suite spans impact processes, volcanic history, and polar volatiles. Baseline experiments with ResNet-50 and SwinV2-B models confirm that each benchmark task is learnable from the released inputs, establishing reference points for future model development.
Read the full abstract

All authors Himanshu Patil, Gabby Nyirjesy, Rachel A. Slank, Vishal Gaur, Daniela Szwarcman, Paolo Fraccaro, Nikolaos Dionelis, Michael K. Barker, Andrew Annex, Vishnu Viswanathan, Zachary Morse, Ethan I. Schaefer, Hiyam Debary, Ankur Kumar, Rohit Lal, Geoffrey Dawson, Campbell Watson, Rebekah I. Dawson-Rigas, Manil Maskey, Juan Bernabé-Moreno, Rahul Ramachandran, Sujit Roy

Emerging Technologies

E-ALS: Approximation-Potential-Aware E-Graph Rewriting for Approximate Logic Synthesis

Approximate logic synthesis (ALS) improves circuit power, performance, and area by trading exact correctness for bounded functional error. However, existing structural ALS methods largely overlook structural bias: even functionally equivalent netlists can expose markedly different approximation opportunities and lead to substantially different outcomes under the same downstream ALS flow. Our experiments show that this effect can induce final area gaps of up to 42.77%. To unlock this opportunity, we propose E-ALS, an e-graph-based framework for approximation-aware structural search. E-ALS combines Function-Reduced Saturation, an ALS-coupled surrogate, search-based extraction, and budget-guided refinement to identify approximation-friendly equivalent structures. Experiments on well-established arithmetic and logic benchmarks show that E-ALS achieves additional area reductions of 3.2 and 7.0 percentage points under maximum Hamming-Distance and Error-Distance constraints, respectively. Code is available in https://github.com/ZenuSunB/Ecompile.git.
Read the full abstract

All authors Bin Sun, Jianan Mu, Jiaxi Zhang, Rengang Zhang, Zhiteng Chao, Jing Ye, Huawei Li

Physics and Society

Assessment of Non-Institutional AI Tool Usage Among Clinicians

Generative artificial intelligence (AI) tools are increasingly accessible and have the potential to improve efficiency across clinical workflows. However, clinicians may also use non-institutional AI tools that are not provided, managed, or governed by their healthcare institutions, creating potential concerns related to privacy, security, accuracy, and clinician-AI interaction. Little is known about how clinicians currently use these tools for work-related tasks. We conducted a descriptive survey of clinicians recruited from the University of Arizona College of Medicine-Tucson and Banner University Medical Center-Tucson between May 20 and June 26, 2026. Participants reported their use of AI tool categories and the frequency with which they used AI for specific tasks across five workload categories: administrative work, clinical work, research, studying/continued education, and teaching. Forty-four respondents completed the survey. Forty-three respondents reported using AI for at least one work-related task during the preceding 6 months. Conversational AI and clinical decision support/diagnostic AI were the most used tool categories, each reported by 28 respondents. Administrative and clinical tasks demonstrated the most frequent use. AI was also used for higher-risk activities, including diagnostic assistance and clinical decision support. Non-institutional AI use was common among surveyed clinicians and extended across a broad range of work-related activities, including tasks with potential implications for clinical reasoning and patient care. Further research is needed to characterize how clinicians use these tools, how they evaluate AI-generated outputs, and how AI can be safely and effectively integrated into clinical workflows.
Read the full abstract

Neurons and Cognition

The Computational Primitives of Adaptation

Research on adaptive systems has traditionally focused on behavior (what organisms do) and mechanism (how their machinery works). This paper focuses on a third level, computation, which considers what adaptive systems must compute to survive and reproduce. It is proposed that adaptation has its own computational structure, comprising a small set of primitive operations common to all adaptive systems, regardless of their physical form. Six primitives, Arouse, Orient, Valence, Position, Boundary, and Attune, were selected using four criteria: necessity for existence, universality across independently evolved lineages, evolutionary conservation, and irreducibility. From this, two main implications follow. First, in biological systems, the primitives provide a substrate-neutral method for cross-species comparisons, reframing elaborative behaviors like attention, memory, and decision-making as combinations of these computations. Second, in artificial systems, the primitives offer a new way to view current challenges in machine intelligence, such as confabulation, prompt injection, distractibility, reward hacking, and catastrophic forgetting, suggesting that these issues may arise from a lack of these computations. Thus, for biological and artificial systems that persist, no specific physical substrate is necessary; rather, these primitives must be implemented if the systems are to be adaptive. What biology offers to inform machine intelligence, then, is not the brain's neural design but the computational functions it evolved to perform. The Computational Primitives Theory (CPT) presented here is a working hypothesis, intended for further refinement through discussion, empirical testing, and application.
Read the full abstract

Statistical Mechanics

Disordered hyperuniform modulated phases and the cosmic web from one free energy

Disordered hyperuniformity spans twenty-five orders of magnitude, from the microscopic world of soft and condensed matter to the distribution of matter in the universe. We introduce a free-energy functional with no adjustable coupling, a phase-field-crystal ordering term and a Newtonian tail, and show that its two kinetic limits generate two of these structures. Under conserved overdamped dynamics it arrests in a bicontinuous labyrinth of the kind found in confined fluids, block copolymers, active matter and vegetation patterns, which is a class III disordered hyperuniform state; under inertial dynamics in an expanding background it generates a cosmic web whose tidal skeleton matches that of gravity alone to within $0.03$ in every morphological class. In both limits the suppressed long-wavelength fluctuations are inherited from the initial condition rather than made by the dynamics: the large-scale exponent of the arrested state equals the primordial one, and gravity rescales the infrared spectrum by the linear growth factor to better than one per cent over a factor $1363$ of growth. The two sectors are continuously connected, and their distance is measurable: once the ordering modulation saturates, its mean-field back-reaction leaves the medium with a residual pressure proportional to $\lambda_0^2$ that suppresses the growth of the largest scales by a smooth factor and vanishes in the cold-matter limit $\lambda_0\to0$.
Read the full abstract