Peer Revenue Play credits. Real papers.

Papers

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

All of arXiv

Thursday 10 September

Machine Learning

REVA: Reusable Evidence View Aggregation for Context-Efficient RAG Serving

Retrieval-augmented generation (RAG) improves knowledge-intensive large language model (LLM) applications by conditioning generation on retrieved documents, but longer contexts increase latency, key-value (KV) cache memory, and token cost. Post-retrieval compression can reduce this cost, yet existing compressors often operate independently for each query, rely on auxiliary models or rewriting, and introduce online overhead that can offset the benefit of shorter prompts. We revisit RAG compression from a data-mining perspective by aggregating historical query–document–model interactions into reusable evidence views. We first show that modern compressors have unstable gains over simple truncation and can add substantial inference-time latency. We then propose Reusable Evidence View Aggregation (REVA), a framework that mines the target generator's historical attention traces into a document-keyed, budget-agnostic score store. REVA maps token-level attention to readable word units, aggregates importance across repeated document accesses, and renders budget-specific plain-text views that preserve document order and the standard RAG interface. Across four representative benchmarks and modern LLMs, REVA improves generation quality by 1.0–5.8 points over existing advances, while reducing compression overhead by a factor of 5.3 to 15.6, adding less than 40 ms of latency.
Read the full abstract

All authors Tuan Nguyen, Qiran Hu, Banruo Liu, Khoa D. Doan, Kok-Seng Wong, Fan Lai

Solar and Stellar Astrophysics

Non-axisymmetric Transport of Magnetic Flux Tubes: A Mechanism for Joy's Law and Poloidal Field Generation via Meridional Flows

We present a new scenario for the solar $\alpha\omega$-dynamo based on the interaction of asymmetric magnetic flux tubes with the near-surface meridional flow. Standard thin flux tube simulations indicate that the radial gradient of differential rotation ($\partial\Omega/\partial r > 0$) breaks the spatial symmetry of emerging $\Omega$-loops. The leading (western) leg becomes nearly vertical, while the trailing (eastern) leg is heavily stretched and flattens along the longitude. Upon emergence at the photosphere, the horizontal meridional flow toward the poles acts more efficiently on the trailing leg because the efficiency of the hydrodynamic entrainment by the plasma (the “sail effect”) is directly proportional to the total spatial length of the magnetic flux tube arc elements. This creates a differential torque that drives the trailing sunspot poleward while the leading sunspot remains anchored due to its compact spatial geometry. Our model yields a tilt angle evolution rate of $\approx 0.3^\circ\,\mathrm{day}^{-1}$ at a latitude of $20^\circ$ for a bipolar separation of $15^\circ$, reproducing Joy's law without invoking the classical Coriolis-driven twist during the rise phase.
Read the full abstract

Machine Learning

Convex Optimization with Nested Evolving Feasible Sets (CONES) under Time-Varying Loss Functions

Convex Optimization with Nested Evolving Feasible Sets (CONES)} was introduced in \cite{CONESVaze} where the objective function \(f\) remains fixed but the feasible region evolves over time as a nested sequence \(S_1 \supseteq S_2 \supseteq \cdots \supseteq S_T\). The goal of an online algorithm is to simultaneously minimize the regret with respect to hindsight static optimal benchmark and the total movement cost $M_\cA(T)$ while ensuring feasibility at all times. CONES is an optimization-oriented generalization of the well-known nested convex body chasing (NCBC). In this paper, we extend CONES to allow for loss functions $f_t'$s to also change over time. When all loss functions are convex, we show that the projected proximal algorithm achieves $O(T^{1-\beta}), O(T^\beta)$ simultaneous regret and movement cost, respectively, for any $\beta \in [0,1)$, over a time horizon of $T$. We also show that any weakly adaptive online algorithm with $O(T^\beta)$ regret has a movement cost of $\Omega\left(T^{\frac{1-\beta}{2}}\right)$ for any $\beta \in [0,1)$. When all loss functions are strongly convex, we show that the projected proximal algorithm simultaneously achieves $O(1)$ regret and a movement cost of $O(\log T)$. To complement this, we show that any online algorithm with sublinear anytime regret has a movement cost of $\Omega\left(\log T\right)$.
Read the full abstract

Artificial Intelligence

CryptoL: Towards Scale Dominance and Physics Constraints Mitigation in Financial Multivariate Time Series Forecasting

Cryptocurrency forecasting presents a distinctive combination of extreme cross-asset scale heterogeneity, non-stationary dynamics, and structural dependencies among Open, High, Low, and Close (OHLC) variables. We present CryptoL, a unified framework designed to address these challenges within multivariate time-series forecasting. CryptoL evaluates forecasting error in context-normalized coordinates within the RevIN pipeline, preventing inverse normalization from introducing an additional squared-scale weighting into the MSE objective. We formally characterize this effect through the empirical risk and parameter-gradient geometry, establishing the conditions under which large-scale assets can disproportionately influence shared-model optimization. Beyond loss-space normalization, CryptoL examines channel-independent and channel-dependent normalization for OHLC data, showing that a shared channel-dependent affine transformation preserves candle-order relations that independent channel transformations need not preserve. The framework further incorporates scale-adaptive numerical stabilization to reduce distortions caused by a fixed normalization constant across assets spanning many orders of magnitude, together with a soft feasibility loss that penalizes violations of the defining OHLC inequalities. Experiments across heterogeneous cryptocurrency assets evaluate these components through controlled ablations and demonstrate improvements in forecasting accuracy, training stability, and the frequency of financially valid OHLC predictions relative to the considered baselines. CryptoL therefore provides an integrated approach to scale-balanced optimization, structure-preserving normalization, numerical stabilization, and constraint-aware cryptocurrency forecasting.
Read the full abstract

All authors Yalda Taheri, Mohammad Hassan Heydari, Armon Rasooli, Maryam Amirshahkarami, Mohammad Ebrahim Mahdavi, Hossein Karshenas

Analysis of PDEs

Viscosity and variational approaches in free boundary problems: A Volume in Honor of Sandro Salsa

Free boundary problems (FBPs) are differential equations where the domain depends on the solution, with applications in diverse fields such as flame propagation, financial mathematics, and tumor growth. Classical examples include the Bernoulli problem, the obstacle problem, and the Stefan problem. The field has advanced significantly through variational and geometric approaches, notably by Alt, Caffarelli, and collaborators, who developed monotonicity formulas and viscosity solutions to address free boundary regularity. Extensions have included nonlinear operators, inhomogeneous problems, and lower-dimensional cases like the thin Bernoulli problem. Evolution problems, such as the Hele-Shaw flow, also connect to FBPs. To celebrate these developments and honor Sandro Salsa's pioneering contributions, a special session on FBPs was held at the AMS-UMI joint meeting in Palermo (2024). This is the editorial of a volume dedicated to Sandro Salsa on the occasion of his 75th birthday.
Read the full abstract

Optics

Two-step high-accuracy microwave frequency measurement and time-frequency analysis based on optical frequency combs

Broadband microwave frequency measurement and time-frequency analysis are crucial for applications such as electronic warfare. However, when it comes to ultra wideband signal analysis, traditional electronic methods have high analysis accuracy, but intrinsic electronic bottlenecks limit their real-time analysis. Here, we propose and experimentally demonstrate a two-step microwave frequency measurement and time-frequency analysis method based on optical frequency combs. The system first performs coarse frequency localization over the 0-40 GHz range using stimulated-Brillouin-scattering-assisted frequency-to-time mapping (FTTM) and dual-comb channelized reception. The dual-comb is then reapplied for downconverting the signal under test, followed by digital signal processing to achieve high-accuracy unambiguous frequency extraction. Experimental results show that the system achieves mean single-tone frequency measurement errors of less than 10 kHz over 0-40 GHz. We further experimentally measure multi-tone, linearly frequency-modulated, and V-shaped frequency-modulated signals, demonstrating the proposed method's capability for analyzing complex signals.
Read the full abstract

All authors Wentao Ma, Taixia Shi, Chi Jiang, Yang Chen

Materials Science

Interplay of spin-lattice and electronic coupling far above Neel ordering in 2D antiferromagnetic CrPS4 and its interface manifestation

A short-range spin correlation driven, strongly intercoupled spin-phonon-electronic state far above TN (~38K) is identified in the low dimensional van der Waals antiferromagnet CrPS4. Temperature-dependent Raman spectroscopy reveals spin-phonon coupling persisting up to T*~120K, concomitant with local lattice distortion. The setting of vibronic progression in photoluminescence spectra suggests the strengthening of electron-phonon coupling around T*. Furthermore, both the electrical transport and optoelectronic response also change significantly at T*. The results indicate that spin-phonon coupling above TN in CrPS4 originates from local lattice distortion induced by short range magnetic correlations which in turn enhances the electron-phonon interaction. Furthermore, using a CrPS4/In2Se3 heterostructure, we demonstrate that the anomaly associated with the coupled degrees of freedom in CrPS4 also influences the adjacent In2Se3 layer. The lattice dynamics of In2Se3 is significantly modified across the magnetic anomaly of CrPS4, and the coupled dynamics is observed at T*. These interfacial manifestation opens up new possibilities for achieving correlated multifunctionalities in artificially designed heterostructure.
Read the full abstract

All authors Divya Jangra, Binoy Krishna De, Mukesh Kumar Dasoundhi, Sourav Chowdhury, Pragati Sharma, Kartick Biswas, Arup Basak, Dheeraj Kumar Gupta, Koushik chakraborty, Praveen Kumar Velpula, Pavan Nukala, U. Chandni, Arvind Kumar yogi, Mukul Gupta, Markus Hucker, Vasant G. Sathe

Computation and Language

Automated Identification of Competing Narratives in Political Discourse on Social Media

Social media platforms have become central to shaping political discourse, serving as arenas where narratives form and evolve, influencing public opinion. Identifying and analyzing these narratives, particularly when they compete across different political ideologies, is crucial for understanding the dynamics of modern political communication. This paper presents an unsupervised framework for identifying and characterizing competing narratives in political discourse on social media, focusing on German politicians' tweets. The framework employs a multi-stage pipeline that integrates natural language processing techniques such as topic modeling, event detection, and event linking. By forming data into coherent stories and uncovering the distinct perspectives of user communities, the system is able to detect the key competing narratives, highlighting the divergent framings and conflicts surrounding trending political topics. Two case studies on polarizing political issues demonstrate the efficacy of the methodology, showcasing its ability to uncover and analyze divergent viewpoints. The findings contribute to the broader understanding of how narratives propagate within the digital public sphere and offer insights for policymakers, social media platforms, and researchers interested in monitoring political discourse.
Read the full abstract

Computer Vision

CEM-TUDASR: Computationally efficient multi-modality transformer based unsupervised domain adaptive super-resolution approach

Wireless Capsule Endoscopy (WCE) enables non-invasive visualization of the gastrointestinal tract, but its miniaturized optics, sensor limitations, and wireless transmission constraints result in low-resolution images with reduced visibility of diagnostically important structures. This paper proposes CEM-TUDASR, a computationally efficient unsupervised Transformer-based super-resolution framework for WCE image enhancement without paired low-resolution (LR) and high-resolution (HR) training data. A domain-adaptive degradation network synthesizes realistic WCE-like LR images from HR conventional endoscopy images, reducing the domain gap and enabling effective unpaired learning. The SR generator integrates Deep Attention Blocks (DABs) and a Fusion Attention Block (FAB) to capture long-range contextual dependencies and fine local structures while preserving perceptual and structural fidelity. The model is trained on a curated dataset derived from Kvasir Capsule and evaluated on KID and GIANA for cross-dataset generalization. No-reference quality metrics, including BRISQUE, PIQE, NIQE, and the domain-specific EndoQM, show that CEM-TUDASR consistently outperforms existing unsupervised SR methods. Qualitative results further demonstrate improved restoration of mucosal textures, vascular patterns, and clinically relevant anatomical details. Cross-domain experiments on retinal images additionally demonstrate the adaptability of the framework. With only 2.67 million parameters and 169.94 GFLOPs, CEM-TUDASR achieves high-quality reconstruction while maintaining computational efficiency, making it suitable for resource-constrained clinical and embedded endoscopic applications.
Read the full abstract

All authors Anjali Sarvaiya, Jay Kadel, Kishor Upla, Kiran Raja

Representation Theory

Obstructions to Jacobi-Finiteness of Quivers with Potentials

We show that Jacobi-finite potentials need not exist on finite $2$-acyclic quivers. Our main tool is a matrix-valued Golod–Shafarevich–Vinberg inequality for quotients of completed path algebras by finitely many, possibly nonhomogeneous, topological relations. Applied to cyclic derivatives, it yields a potential-dependent obstruction to the finite-dimensionality of completed Jacobian algebras. We then construct a purely quiver-level criterion excluding every Jacobi-finite potential on a given quiver, and exhibit a family of quivers for which every potential has an infinite-dimensional Jacobian algebra.
Read the full abstract

Artificial Intelligence

An AI-Powered Culturally Aware Chatbot for Stress Detection and Wellness Support among Pakistani University Students Using NLP and Machine Learning

With the existing digital mental health tools specifically developed for Western settings, Pakistani students are exposed to a uniquely compounded stress situation in their university that includes academic, financial, familial, and relational stressors, which have become a serious concern for academic and psychological development of students in Pakistani universities. This paper introduces a new, AI-driven and culturally sensitive stress detection and wellness support system that is tailored to the context of Pakistani university students. The system is based on a machine learning model called Random Forest which is trained using a validated student stress data set of 1100 responses on 20 features from psychological, physiological, academic, environmental and social aspects, with an accuracy of 89.09% and a macro F1-score of 0.89, in three stress severity levels. The classification outputs are passed on to an open-source large language model through OpenRouter API, where an appropriately crafted system prompt, culturally aware, gives the model a conversation about wellness, in English, Urdu and Roman Urdu. The second most predictive stress factor in this population identified by feature importance analysis was teacher-student relationship, which is a culturally important stress factor highlighting the need for region-aware mental health systems. Future research will involve primary data collection from students at various academic levels of Pakistani Universities with the validated DASS-21 instrument focusing on the students who are moving from FSc to undergraduate studies, which is a time of being psychologically vulnerable which is under-researched.
Read the full abstract

Computers and Society

(Whose defaults?) Is artificial intelligence reorienting archaeological methods?

Generative AI and the practice of "vibe coding" are changing how archaeologists carry out computational research, but their effects on the discipline's range of methods is still understudied. In this paper, we evaluate whether large language models (LLMs) are narrowing the variety of methods archaeologists use. We first analysed approximately 119,000 archaeology abstracts from Scopus, covering publications from 2010 to 2025. Using a locally run LLM, we identified the computational methods reported in each abstract and organised them into 25 broad categories (L2) and 241 finer clusters (L3). A Bayesian Dirichlet-multinomial model of method composition within sub-disciplines found a small but credible shift in method use after 2023. However, this shift was smaller than the variation already present across the full study period. No individual technique showed a significant change, and overall methodological diversity increased rather than declined. We then ran a controlled experiment to see whether LLMs recommend a narrower set of methods than archaeologists have used in practice. Two different open-weight models were asked to suggest methods for 28 archaeological research problems, with prompts providing three levels of methodological guidance: novice, intermediate, and expert. Recommendation diversity was much lower than in the published literature, particularly without methodological guidance. The models also tended to favour methods that were widely used before 2023, and their recommendations more closely resembled the post-2023 literature. Taken together, these results are consistent with LLMs pushing methodological choice towards convergence, although our study cannot establish a causal effect. They raise a broader question: how can archaeology retain methodological diversity as LLMs become more involved in research?
Read the full abstract

Optimization and Control

Optimal Control for Cancer Chemotherapy Using Hybrid Quantum Particle Swarm Optimization

Optimal control in cancer chemotherapy is challenged by tumor heterogeneity and mutations, which complicate treatment effectiveness. Traditional methods, such as Pontryagin's maximum principle (PMP), are often hindered by their reliance on an initial guess for the costate equation, affecting accuracy and convergence. To address these limitations, this work introduces a hybrid quantum particle swarm optimization (QPSO) method based on regularization. QPSO is employed for global exploration to approximate the optimal control trajectory, followed by a regularization-based refinement to ensure smoothness and consistency with optimality conditions. The Hamiltonian function is used for first- and second-order optimality checks, verifying solution quality. Numerical case studies explore various drug effectiveness functions, demonstrating the role of periodic and localized drug delivery in achieving robust tumor suppression and providing insights into the impact of different drug combinations on optimal chemotherapy strategies.
Read the full abstract

High Energy Physics (Phenomenology)

Doubling down on proton structure: what double parton scattering can and cannot tell us about partonic separation

Double parton scattering provides a unique probe of proton structure and is commonly characterized by the effective cross section $\sigma_\mathrm{eff}$. We examine, under the factorized ansatz for the double-parton-scattering cross section and without assuming any specific parametric form for the transverse profile, what the effective cross section can reveal about the transverse separation, $r$, between partons in the proton. We derive a sharp lower bound on the transverse partonic separation, $\langle r^2\rangle \ge 4\sigma_\mathrm{eff}/(9\pi)$, and demonstrate through explicit counterexamples that previously posited upper bounds do not hold. Our results establish, within the effective-cross-section parametrization of double parton scattering framework, that $\sigma_\mathrm{eff}$ can provide a sharp lower bound on partonic separation but cannot provide a general upper bound.
Read the full abstract

Software Engineering

FST Pay: Deterministic Safety-Gated Architecture for Youth Digital Payments

Digital payment infrastructures increasingly provide adolescent users with direct access to real-time financial services. While early access promotes financial literacy and digital inclusion, it exposes young users to severe risks of impulsive spending, social engineering frauds, unauthorized transactions, and merchant exploitation. Conventional countermeasures rely on probabilistic machine learning or rigid static controls. However, allowing probabilistic or generative artificial intelligence (AI) models to directly influence real-time payment authorization introduces non-determinism, unpredictable edge-case behavior, and critical audit vulnerabilities. This paper introduces Financial Safety for Teens Pay (FST Pay) as an architectural and formal specification. FST Pay is founded on an immutable operational boundary: strict deterministic safety gating on the real-time authorization path coupled with decoupled downstream AI explanation. Transactions initiated via rails like UPI are subjected to six deterministic invariant checks covering spending limits, guardian co-sign policies, transaction amount thresholds, merchant category codes, temporal access intervals, and hardware integrity constraints. Transactions are classified strictly into ALLOW, REVIEW, or BLOCK outcomes through an ordered, mutually exclusive decision function. High-risk transactions trigger an asynchronous guardian co-sign workflow. Generative AI is relegated entirely downstream of settlement, consuming published post-decision events solely to generate natural-language financial insights without holding mutation privileges over the ledger.
Read the full abstract

Audio and Speech Processing

Domain-Incremental Learning for Multi-Channel Replay Speech Detection

Replay attacks are the most accessible threat to voice-controlled systems, and the acoustic cues that expose them are strongly modulated by the environment in which the attack is mounted. A detector deployed in the field therefore has to absorb new acoustic conditions over time, ideally without revisiting past recordings, since retaining speech indefinitely is both expensive and legally constrained. We frame this as Domain-Incremental Learning (DIL) over acoustic environments and present the first continual learning benchmark for multi-channel replay speech detection, evaluating a state-of-the-art beamformer-based detector over all 24 environment orderings of the ReMASC corpus with five seeds. Sequential fine-tuning forgets severely, raising the error rate on previously learned environments by 18.8 points. Elastic weight consolidation (EWC) halves forgetting but loses plasticity, gradient projection memory (GPM) is statistically indistinguishable from naive fine-tuning, and the proposed task-specific beamformer (TSB) that keeps one spatial front-end per environment significantly improves final and incremental accuracy. We further show that the last environment of the sequence dominates final performance. Code, results, and analysis are available at https://github.com/michaelneri/replay-speech-continual.
Read the full abstract

Human-Computer Interaction

Conceptualising an Initial Design Space for Guidance in Digital Physical Activity Support

Providing guidance is frequently referenced as a key capability of digital health interventions targeting physical activity, yet the term remains poorly defined and inconsistently applied. Existing work often conflates guidance with related constructs such as personalisation, feedback, or persuasion, limiting both theoretical clarity and design progress. This paper conceptualises an initial design space of guidance in the context of digital physical activity support. We define guidance for physical activity as situated, action-oriented support that scaffolds users' embodied engagement in physical activity. Drawing on literature from behaviour change, human-computer interaction, embodied cognition, and digital health, we outline a design space that characterises guidance along multiple dimensions: scope, purpose, timing, context, modality, embodiment, adaptivity, autonomy, and affective quality. By offering a structured vocabulary and conceptual foundation, this work aims to support more coherent research, comparisons, and responsible design of digital health interventions featuring guidance for physical activity support.
Read the full abstract

All authors Faith Young, Markus Tatzgern, Alexander Meschtscherjakov, Jan Smeddinck

Computation and Language

FlexComp: One Model for Every Ratio in Context Compression

Soft context compression condenses a context into a few memory tokens that a frozen LLM consumes in place of the raw text, but existing compressors fix the compression ratio at training and inference: each deployed ratio requires a separately trained model, and the chosen ratio is applied uniformly to all inputs, whose actual needs vary drastically. We propose FlexComp, a method-agnostic framework that decouples the ratio from both training and deployment: Matryoshka-style training samples the memory budget $K$ per instance, turning one model into an any-ratio compressor, and the budget is then chosen per input by: (1) confidence-based cascade routing or (2) a lightweight learned $K$ predictor. Across ICAE, 500xCompressor, and SAC on MRQA, a single FlexComp model matches separately trained fixed-ratio specialists with minimal degradation. Cascade routing preserves over 98% of the mildest ratio's accuracy at up to 266x average compression; the $K$ predictor, in a single compression-decoding pass, reaches 158-236x within 0.7 F1 of the mildest ratio. At serving-scale batch sizes, the $K$ predictor cuts context KV cache by 50% and improves decoding throughput by 47%.
Read the full abstract

All authors Kaiyan Zhao, Zhongtao Miao, Akiko Aizawa, Yoshimasa Tsuruoka

Astrophysics of Galaxies

Black Hole-Galaxy Correlations in Cluster Zoomed-in Simulations: GIZMO-SIMBA and TNG-Cluster

We investigate the co-evolution of supermassive black holes (SMBHs) and central galaxies in massive clusters using the GIZMO-SIMBA and TNG-Cluster zoom-in simulations at $z=0-5$. We find that the distinct subgrid physics of these two models suggest fundamentally different evolutionary pathways. On the one hand, GIZMO-SIMBA, employs torque-limited accretion and predicts a supply-driven scenario where the SMBHs rapidly assemble synchronized with dark matter halo ($M_{200c}$) growth (i.e. the halo mass-BH mass relation is set by $z=3.0$ and similar to the present-day relationship). On the other hand, TNG-Cluster, exhibits a feedback-regulated growth phase delayed by an early thermal suppression. While both models successfully reproduce some local black hole-galaxy scaling relations, they imply significantly different evolution for these relations. Analysis of the BH mass-gas mass ratio relations suggests that TNG-Cluster's isotropic kinetic winds efficiently deplete cold gas, resulting in a "hard quench" of star formation. In the black hole accretion rate (BHAR)-star formation rate (SFR) relation we find that both simulations successfully reproduce the decoupling of BHAR and star formation observed in recent massive cluster ellipticals. The divergent evolutionary trends emphasize the importance of the multiphase intracluster medium; while these subgrid models do not have the necessary resolution and employ distinct formalisms, the sustained BHAR in quenched systems resemble outcomes broadly consistent with modern multiphase feeding paradigms such as chaotic cold accretion in turbulent cluster cores. Furthermore, we demonstrate that for both models, black hole mass is a primary regulator of atomic and molecular gas depletion in galaxy clusters.
Read the full abstract

All authors Hangxin Pu, Antonios Katsianis, Weiguang Cui, Romeel Davé, Massimo Gaspari, Weishan Zhu, Xiaohu Yang, Jinyi Shangguan, Qingshan Wang, Ying Tang, Yuchang Li

Artificial Intelligence

Agentic Share-of-Search: A Multi-Agent AI System for Competitive Decision-Making in LLM-Mediated E-Commerce

AI shopping assistants increasingly redirect consumer discovery, creating an urgent need for tools that support seller-side competitive decision-making. We present a multi-agent AI system that automates competitive visibility measurement and root cause diagnosis in LLM-mediated ecommerce. The system introduces Agentic Share-of-Search (ASoS) as the decision target, deploys query agents across leading AI platforms, and uses a ReAct-based diagnostic agent to recommend prioritized merchandising interventions. A 100-trial ablation study, presented as a feasibility evaluation of this prototype, shows the agent recovers the ablated signal in 39% of trials (95% CI: 30.0% - 48.8%, 5.5x over chance), rising to 63.9% among high-correlation ablations.
Read the full abstract

Combinatorics

Vector Balancing via Directional Total Variation

Our main result is a $3\sqrt{2\pi}$ bound for the Komlós signing problem: every finite family of real vectors of Euclidean norm at most one admits a signed sum of $\ell_\infty$-norm less than this constant, independently of the dimension and the family size. For any $\kappa\ge0$, if a bounded open convex set supports a probability density with directional total variation at most $\kappa$ in every unit direction, then its open-set Banaszczyk transform supports another such density with the same $\kappa$, provided the translation vector $v$ satisfies $\kappa\|v\|_2\le1/3$. As a consequence, every finite set system in which each element belongs to at most $t$ sets, where $t\ge1$ is an integer, admits a two-coloring whose imbalance in each set is less than $3\sqrt{2\pi t}$. This gives the square-root dependence predicted by the Beck-Fiala conjecture. The proof was discovered by the Odin Automatic AI Research Agent.
Read the full abstract

Computer Vision

A Multi-View and Confusion-Guided Ensemble Framework for Robust Synthetic Image Attribution

Synthetic image attribution (SIA) has become increasingly important with the rapid advancement of text-to-image generation models. However, accurately identifying the source model of a generated image remains challenging due to the growing similarity among modern diffusion-based generators and the presence of diverse post-processing operations. In this report, we present a multi-view and confusion-guided ensemble framework for the Synthetic Image Attribution Challenge of the DLMMDD Workshop at ICANN 2026. Our approach integrates multiple complementary architectures, including FFT-ConvNeXt, DINOv2, CLIP, and Xception, to capture diverse attribution cues from frequency, semantic, and forensic perspectives. To improve robustness against unknown degradations and image manipulations, extensive data augmentation strategies are employed during training, simulating realistic post-processing operations such as compression, resizing, grayscale conversion, and blur. Furthermore, we analyze the confusion patterns of the ensemble model and observe severe ambiguity between Stable Diffusion 3 and Stable Diffusion 3.5. To address this issue, we introduce a dedicated binary expert classifier that is selectively activated under low-confidence conditions. We additionally apply class-adaptive confidence calibration to improve the discrimination of challenging classes such as Tencent Hunyuan. The proposed framework achieved 99.53% on the public leaderboard and 99.20% on the private leaderboard. The source code and implementation details are publicly available at https://github.com/ZOMIN28/SIA.
Read the full abstract

Information Theory

Physics of Information Geometry - Part II: Small-Step Active Inference on the Probability Simplex

This paper is the second in a two-part investigation of the physics of information geometry. While Part I develops a physical foundation for distributional motion on the probability simplex, the present paper studies how that framework manifests in active inference. The treatment is fully self-contained and does not require familiarity with Part I. We focus in particular on active inference through small distributional steps and the geometric structure induced by such local motion. Starting from an initial distribution, an agent evolves its belief state toward a final target distribution through a sequence of constrained updates. We define a relative free energy functional with respect to the preferred distribution and extend it to a relative potential energy analogous to the Helmholtz/Gibbs free-energy decomposition. The evolution is subject to a per-step kinetic constraint expressed through the Kullback-Leibler (KL) divergence between consecutive distributions, which serves as a discrete kinetic energy on the probability simplex. Using the information-geometric Pythagorean theorem on KL balls, we show that sufficiently small local moves dominate large direct jumps, and that greedy maximization of free-energy reduction is globally optimal under the kinetic constraint. This leads to a sequential variational principle in which the optimal trajectory minimizes the associated Lagrangian of the optimization problem. Similar to classical mechanics, the Lagrangian takes on the form as the difference between the kinetic and potential terms, establishing a least-action principle for distributional motion on the simplex. The resulting optimal update admits a closed form as an exponentially tilted version of the current distribution toward the preferred distribution, parametrized by an inverse-temperature-like multiplier. We further extend the framework to incorporate state-dependent geodesic...
Read the full abstract

Classical Analysis and ODEs

Riesz capacity ratios with negative exponents

We investigate sharp inequalities for ratios of Riesz capacities with negative exponents by combining computational experiments with rigorous analysis. For finite subsets of the line, we prove positivity of equilibrium masses when $-1<p<0$, enabling numerical tests of conjectured extremal ratios. In the plane, comparisons of the disk with regular polygon vertex sets reveal a cascade of transitions among the tested competitors and suggest a precise conjecture for the equilibrium measure of odd polygons, for which we give a partial proof. Numerical intersections of equality curves show that the regions where these sets outperform the disk are not simply nested. Similar numerical intersections occur in three dimensions between the regular-simplex equality curve and those of explicit five-point and six-point configurations. Motivated by the dimensional dependence of these comparisons, we prove that for each fixed $p<-2<q<0$, the regular simplex has a larger capacity ratio than the ball in all sufficiently large dimensions. Accompanying Python and Mathematica code supports reproduction and further testing of the conjectures.
Read the full abstract

Artificial Intelligence

Can LLMs Follow Medical Expert Logic? A Benchmark for Hierarchical Logical Consistency in Risk-of-Bias Assessment

Evidence-based medicine demands strict logical consistency, yet current evaluations of large language models (LLMs) prioritize superficial label matching over genuine reasoning. We introduce LogiMed-RoB, a benchmark grounded in Cochrane Risk of Bias (RoB) 2.0 expert logic, comprising 860 randomized controlled trials (RCTs) and 14,820 queries. It evaluates models under the Hierarchical Logical Consistency (HLC) framework across four dimensions: Atomic Consistency, Domain Consistency, Aggregation Consistency, and Evidential Faithfulness. Experiments on 10 state-of-the-art LLMs reveal a catastrophic Error Compounding Effect: despite the top model reaching 98.88% Atomic Consistency, its end-to-end consistency collapses to 45.13%, with several open-weight architectures plummeting to nearly 0%. We further uncover a systematic evidence-reasoning gap: even when models retrieve high-quality evidence, they fail to deduce correct outcomes in 18.63-40.05% of cases, while Blind Guess Rates reach 48.28%. LogiMed-RoB demonstrates that high outcome accuracy can conceal critical reasoning flaws, underscoring the necessity of white-box logical verification for clinical deployment.
Read the full abstract

All authors Jiayu Huang, Zichen Tang, Qianhui Ling, Zemin Kuang, Haihong E

Signal Processing

Trellis-Based Noise Modulation with Soft-Decision Viterbi Detection

Noise modulation conveys information through the statistical properties of noise rather than conventional deterministic signal parameters. This paper investigates a trellis-based noise modulation framework that exploits temporal dependencies between successive noise-power symbols. A binary filtering-based configuration is first developed as an illustrative finite-state model for comparing hard- and soft-decision sequence detection. A joint soft Viterbi receiver is then proposed, which directly incorporates the received noise energy into a likelihood-based branch metric, avoiding the information loss associated with intermediate hard decisions. The framework is further extended to an N-ary state-dependent trellis-based noise modulation scheme, where the transmitted noise-power level depends jointly on the current input symbol and trellis state. A Bhattacharyya-distance-based method is employed for systematic power-level design. Simulation results demonstrate the advantage of soft-decision sequence detection over hard-decision processing and investigate the effects of power-level design, traceback depth, energy-per-bit-to-noise ratio, symbol duration, and modulation order. The results also reveal the trade-off between spectral efficiency, detection reliability, and trellis complexity in higher-order noise modulation.
Read the full abstract

All authors Razieh Torkamani, Hadi Zayyani, Mohammad Salman, Felipe A. P. de Figueiredo, Rausley A. A. de Souza

Optimization and Control

Single-Loop Gradient Algorithms for Pessimistic Bilevel Optimization Problems

Bilevel optimization has recently attracted growing attention, particularly in the development of efficient numerical methods. Despite substantial progress on optimistic bilevel optimization, pessimistic bilevel optimization (PBO) remains much less explored, especially the design of fully first-order, single-loop gradient-based methods. To address this gap, we propose a smooth approximation of PBO through reformulation, penalization and regularization, and establish convergence guarantees in terms of both minimizers and stationarity. Building on this framework, we then develop two single-loop algorithms for deterministic and stochastic PBOs, respectively. Both use only first-order gradient information and avoid second-order derivatives and inner-loop subproblem solves. Non-asymptotic convergence rates for the proposed algorithms are established to provide theoretical guarantees. Through a systematic empirical study of both synthetic and practical problem instances, we demonstrate that our algorithms are highly effective and efficient. In particular, our results on spam classification and Smart Predict-then-Optimize further illustrate PBO's advantages over its classical optimistic bilevel counterpart, highlighting its strong potential for practical modeling and the delivery of robust solutions.
Read the full abstract

All authors Qichao Cao, Bo Zeng, Shangzhi Zeng, Jin Zhang

Audio and Speech Processing

Exploring Second-Order Pattern Recognition in Speaker Recognition

In traditional pattern recognition tasks, neural networks are trained to recognise human-defined patterns (e.g. audio categories) in model inputs (e.g. audio). Meanwhile, some Explainable AI (XAI) methods explain latent patterns characterising the network's recognition of inputs as human-defined patterns; this work calls these latent patterns second-order patterns and proposes to discover them. Accordingly, we apply a hierarchical clustering algorithm to analyse whether our speaker recognition network's representations learned from known utterances naturally form hierarchical clusters. Each resulting cluster is a second-order pattern that characterises a context in our network's recognition of the known utterances as speaker identities. All discovered second-order patterns are then interpreted using the Hierarchical Cluster-Class Matching (HCCM) method. Moreover, we propose a new task, second-order pattern recognition, to identify which of the discovered second-order patterns characterising the recognition of known utterances also apply to unseen utterances, thereby characterising the recognition of unseen utterances. Accordingly, we design the Hierarchical Cluster Navigation and Assignment (HCNA) method. HCNA recognises a second-order pattern as applying to an unseen utterance when the utterance's network representation lies within the extrapolation space of the cluster regarded as that second-order pattern. Experimental results demonstrate that the extrapolation space introduced in HCNA substantially improves task performance.
Read the full abstract

Cosmology and Nongalactic Astrophysics

SMART : Spherical Mass ApeRture Toolkit

The aperture mass filter is a widely used tool in weak lensing studies. While previous surveys have typically relied on the flat-sky approximation, forthcoming Stage IV surveys such as Euclid will cover sufficiently large areas to require the construction of curved-sky aperture mass maps. In this paper, we extend the shear-based aperture mass formalism to the sphere and introduce a software package for a fast and precise computation of spherical aperture mass maps from the shear field, defined directly on the celestial sphere. Unlike existing approaches, our method does not rely on planar projections and does not require the reconstruction of convergence maps. The software provides four different implementations, all of which are independent of the choice of aperture mass filter function. Starting from the brute-force approach, whose computational cost renders it impractical for future surveys covering more than 10,000 deg2, we investigate alternative galaxy-grouping strategies to improve computational efficiency. We demonstrate speed-up factors of up to 40 at a resolution of NSIDE = 4096 while achieving significantly higher precision than existing state-of-the-art methods. In particular, we show that the spherical shear-based implementations using a pixel-based approach outperforms the existing methods in terms of both precision and computational time. Finally, we show that, for weak-lensing cluster detection, the cluster sample produced using the spherical shear-based aperture mass method consistently outperforms those obtained using projected shear-based approaches in the purity-completeness plane, while simplifying the resulting selection function. Our spherical mass aperture software, called SMART, is made available on GitHub.
Read the full abstract

All authors Igor Bezmaternykh, Sandrine Pires, Nicolas Dagoneau, Gabriel W. Pratt, Loris Chappuis, Gavin Leroy

Artificial Intelligence

SemVerBench: Benchmarking LLM Comprehension of Version-Constraint Resolution Semantics

Large language model (LLM) coding agents constantly decide whether a version satisfies a constraint such as ^1.2.3 or >=2.0,<3, yet their grasp of version-constraint semantics has never been measured directly. We introduce SemVerBench, the first benchmark of LLM version-constraint resolution semantics across three ecosystems (npm, PEP 440, Cargo): 240 machine-checkable items with unique answers, built author-neutrally from four balanced sources (each ecosystem's official test suite plus three frontier LLM proposers) and labeled by a non-circular two-implementation oracle. Evaluating six frontier models, we find systematic, predictable per-mechanism blind spots: a partial-comparator carry rule (>1.2 means >=1.3.0) traps every model on Cargo (near 60%), and although standard PEP 440 prefix matching is universal, on zero-pad/post-release corner cases GPT-5.1 collapses (0/26) while Claude stays at 97-100% (verified on a 67-item oracle-validated set). Opus significantly outperforms all other models, and Sonnet outperforms the OpenAI models (McNemar). The failures look more like an activation/application gap than a knowledge gap: injecting the rule or a light correct hint recovers most errors, whereas interval decomposition does not, and models are at ceiling on the basic forms of the same rules. An author-stratified analysis finds no statistically significant self-favoritism. Because the task is verifiable and a free, 100%-correct resolver exists, tool delegation reaches ~100%: coding agents should delegate version resolution to a resolver rather than reason about versions in-head.
Read the full abstract