April 2026 arXiv papers — page 120

David Fajman, Liam Urban

We analyse the effect of expansion in Newtonian cosmology on the asymptotic behaviour of charged self-interacting plasmas close to Poisson equilibria. To this end, we study the Vlasov-Poisson system on the phase space of a $3$-torus which is expanding with respect to the scale factor $a(t)$. We show that, for $a(t)=t^q$ with $q\in(0,\frac12)$, solutions to t

Rui Li, Bingyu Li, Yuanzhi Liang, Haibin Huang

Achieving high-fidelity generation in extremely few sampling steps has long been a central goal of generative modeling. Existing approaches largely rely on distillation-based frameworks to compress the original multi-step denoising process into a few-step generator. However, such methods inherently constrain the student to imitate a stronger multi-step teach

Xinpeng Yang, Meng Hao, Chenkai Weng, Robert H. Deng

Private set intersection (PSI) enables a sender holding a set $Q$ of size $m$ and a receiver holding a set $W$ of size $n$ to securely compute the intersection $Q \cap W$. Fuzzy PSI (FPSI) is a PSI variant where the receiver learns the items $q \in Q$ for which there exists some $w \in W$ satisfying $\mathsf{dist}(q, w) \le \delta$ under a given distance met

Emre Özyıldırım, Barış Yaycı, Umut Eren Akturk, Cem Tekin

We study downlink beam and rate adaptation in a multi-user mmWave MISO system where multiple base stations (BSs), each using analog beamforming from finite codebooks, serve multiple single-antenna user equipments (UEs) with a unique beam per UE and discrete data transmission rates. BSs learn about transmission success based on ACK/NACK feedback. To encode se

Evaldas Vaiciukynas, Paulius Danenas, Linas Ablonskis, Algirdas Sukys

Online hate speech and abusive language pose a growing challenge for content moderation, especially in multilingual settings and for low-resource languages such as Lithuanian. This paper investigates to what extent modern multilingual sentence embedding models can support accurate hate speech detection in Lithuanian, Russian, and English, and how their perfo

Armin Gießler, Pol Jané-Soneira, Sören Hohmann

This paper studies the data-driven synthesis of linear quadratic integral (LQI) controllers for continuous-time systems. The objective is to achieve optimal state-feedback control with integral action for reference tracking using only measured data. To this end, we derive a data-driven closed-loop parameterization of the augmented dynamics that incorporates

Rahul Chakraborty, Paweł Tecmer

The pair-coupled-cluster doubles (pCCD) method has emerged as a viable approach for quantum-chemical studies of strongly correlated systems. Despite its lower formal scaling (O(N$^4$)) compared to other versions of coupled cluster (CC) theory, applications to large chemical structures are still expensive. Fragmentation and embedding strategies offer a viable

Henry Bradford

We propose a quantitative notion of permutation stability for finitely generated groups. Our notion is related to, but distinct from, the ``stability rate'' introduced by Becker and Mosheiff (which is valid within the class of finitely presented groups). We construct a family of finitely generated stable groups which exhibit, quantitatively, arbitrarily ``ba

Pei-An Chen, Yong-Ching Liang, Jia-Fong Yeh, Hung-Ting Su

Intelligent embodied agents should not simply follow instructions, as real-world environments often involve unexpected conditions and exceptions. However, existing methods usually focus on directly executing instructions, without considering whether the target objects can actually be manipulated, meaning they fail to assess available affordances. To address

Rikard Rosenbacke, Carl Rosenbacke, Victor Rosenbacke, Martin McKee

Large language models have advanced rapidly, from pattern recognition to emerging forms of reasoning, yet they remain confined to linguistic simulation rather than grounded understanding. They can produce fluent outputs that resemble reflection, but lack temporal continuity, causal feedback, and anchoring in real-world interaction. This paper proposes a comp

Boyu Han, Xu Du, Karl H. Johansson, Apostolos I. Rikos

This paper addresses distributed consensus optimization problems with mixed-integer variables, with a specific focus on Boolean variables. We introduce a novel distributed algorithm that extends the Consensus Augmented Lagrangian Alternating Direction Inexact Newton (CALADIN) framework by incorporating specialized techniques for handling Boolean variables wi

Marta Sumyk, Oleksandr Kosovan

We present an initial investigation into Agentic Retrieval-Augmented Generation (RAG) for Ukrainian, conducted within the UNLP 2026 Shared Task on Multi-Domain Document Understanding. Our system combines two-stage retrieval (BGE-M3 with BGE reranking) with a lightweight agentic layer performing query rephrasing and answer-retry loops on top of Qwen2.5-3B-Ins

Ziwu Sun, Zhen Gao, Jiyong Zhang, Jiaheng Li

We propose a new perspective on policy optimization: rather than reweighting all samples by their importance ratios, an optimizer should select which samples are trustworthy enough to drive a policy update. Building on this view, we introduce Rejection-Gated Policy Optimization (RGPO), which replaces the importance sampling ratio r_theta = pi_theta / pi_old

Chuanlin Liu, Dan Liu, Jie Guan, Chao Lian

We investigate optically induced phase transitions in the two-dimensional (2D) ferroelectric (FE) material Nb2O2I4 using real-time time-dependent density functional theory (rt-TDDFT). Our results demonstrate that tailored laser pulses can activate specific coherent phonon modes. Specifically, the anharmonic atomic distortions of the A1-1 and A1-2 modes at th

Ruisen Qian

This paper investigates a class of generalized mean-reflected McKean-Vlasov type backward stochastic differential equations (BSDEs). Our new framework combines a mean reflection constraint on the solution's expectation with a generalized integral with respect to a continuous non-decreasing process. We establish the existence and uniqueness of the solution. T

Anton Rebhan, Luigi Cappiello, Josef Leutgeb, Jonas Mager

We review the recent progress made with regard to the hadronic light-by-light (HLbL) contribution to the Standard Model prediction of the muon anomalous magnetic moment and how well this compares with predictions from holographic QCD models, which had predicted larger contributions from axial vector mesons and short-distance constraints than the White Paper

Lars H. Halle, Klaus Hulek, Ziyu Zhang

We present a method to construct explicit degenerations of higher-dimensional generalized Kummer varieties. We start with a simple degeneration $f: \mathcal Y \to C$ of abelian surfaces. Then $ \mathcal{Y} \setminus \mathcal{Y}_0$ is an abelian scheme over $C \setminus 0$ and we can form the relative generalized Kummer variety $K^{n-1}_{\circ} = \mathrm{Kum}

Hardy Halbauer, Le Hang Nguyen, Thorsten Wild

Sub-THz bands are promising high bandwidth and data rates, and in the recent years the device technologies made large progress and provided a multitude of transceiver, power amplifier (PA) and phased array devices supporting the frequency bands above 100 GHz. The more painful aspect of sub-THz transmission is the increased power consumption, caused by the la

Thanh Linh Nguyen, Nguyen Van Huynh, Quoc-Viet Pham

In data-sensitive domains such as healthcare, cross-silo federated learning (CFL) allows organizations to collaboratively train AI models without sharing raw data. However, practical CFL deployments are inherently coopetitive, in which organizations cooperate during model training while competing in downstream markets. In such settings, training contribution

Zihong Zhang, Zuchao Li, Lefei Zhang, Ping Wang

Autoregressive decoding in Large Language Models (LLMs) generates one token per step, causing high inference latency. Speculative decoding (SD) mitigates this through a guess-and-verify strategy, but existing training-free variants face trade-offs: retrieval-based drafts break when no exact match exists, while logits-based drafts lack structural guidance. We

Jianchao Huang, Fengming Zhang, Haibo Zhu, Tao Yan

Small object detection remains a significant challenge due to feature degradation from downsampling, mutual occlusion in dense clusters, and complex background interference. To address these issues, this paper proposes FSDETR, a frequency-spatial feature enhancement framework built upon the RT-DETR baseline. By establishing a collaborative modeling mechanism

Radhika Khatri, Adit Tewari, Nikhil Sharma, M. B. Srinivas

Rapid urbanization and continuous population growth have made municipal solid waste management increasingly challenging. These challenges highlight the need for smarter and automated waste management solutions. This paper presents the design and evaluation of an integrated waste management framework that combines two connected systems, a robotic waste segreg

Rikard Rosenbacke, Carl Rosenbacke, Victor Rosenbacke, Martin McKee

Large language models are increasingly integrated into decision-making in areas such as healthcare, law, finance, engineering, and government. Yet they share a critical limitation: they produce fluent outputs even when their internal reasoning has drifted. A confident answer can conceal uncertainty, speculation, or inconsistency, and small changes in phrasin

Ertugrul Kececi, Tufan Kumbasar

Recent advances in Deep Learning (DL) have boosted data-driven System Identification (SysID), but reliable use requires Uncertainty Quantification (UQ) alongside accurate predictions. Although UQ-capable models such as Fuzzy ODE (FODE) can produce Prediction Intervals (PIs), they offer limited interpretability. We introduce Explainable Type-2 Fuzzy Additive

Murat Furkan Mansur, Tufan Kumbasar

Nonlinear system identification must balance physical interpretability with model flexibility. Classical methods yield structured, control-relevant models but rely on rigid parametric forms that often miss complex nonlinearities, whereas Neural ODEs are expressive yet largely black-box. Physics-Informed Neural Networks (PINNs) sit between these extremes, but

Yanyan Zou, Junbo Qi, Lunsong Huang, Yu Li

Generative Retrieval (GR) offers a promising paradigm for recommendation through next-token prediction (NTP). However, scaling it to large-scale industrial systems introduces three challenges: (i) within a single request, the identical model inputs may produce inconsistent outputs due to the pagination request mechanism; (ii) the prohibitive cost of encoding

Zhiyuan Zhai, Wenjing Yan, Xiaodan Shao, Xin Wang

Does reinforcement learning genuinely expand what LLM agents can do, or merely make them more reliable? For static reasoning, recent work answers the second: base and RL pass@k curves converge at large k. We ask whether this holds for agentic tool use, where T rounds of interaction enable compositional strategies that re-sampling cannot recover. We introduce

Subhodip Panda, Shubhada Agrawal

We study the tail behavior of regret in stochastic multi-armed bandits for algorithms that are asymptotically optimal in expectation. While minimizing expected regret is the classical objective, recent work shows that even such algorithms can exhibit heavy regret tails, incurring large regret with non-negligible probability. Existing sharp characterizations

Michael A. Seaton, Benjamin T. Speake, Ilian T. Todorov

Modelling micro- and meso-scopic scale thermodynamic and transport properties of soft condensed matter hinges upon its representation. This is especially relevant for polar solvents such as water, since these require effective representation of their dielectric nature as driven by molecular charge distributions and molecular network structuring. The dielectr

Qijia Chen, Andrea Bellucci, Zhida Sun, Giulio Jacucci

LLM-based mobile GUI agents treat every task invocation as an independent reasoning episode, requiring a full LLM inference call at each action step. This per-step dependence makes them stateless: a task completed successfully yesterday is re-derived from scratch today, with no improvement in reliability or speed. We present SkillDroid, a three-layer skill a

Nikita Kiselev, Andrey Grabovoy

Local loss-landscape stabilization under sample growth is typically measured either pointwise or through isotropic averaging in the full parameter space. Despite practical value, both choices probe directions that contribute little to the dominant local deformation of strongly anisotropic neural landscapes. We recast stabilization as an observational problem

Tijl Schepens, Thomas Feys, Thomas Eriksson, Gilles Callebaut

Sub-Terahertz radio-stripe and distributed MIMO architectures promise extreme spatial reuse and multi-GHz bandwidths, but the cascaded fiber front-haul and RF hardware impairments strongly shape end-to-end performance. This paper presents an open-source, configuration-driven simulator that models the full waveform-level signal chain from CP-OFDM baseband gen

Fernando Amodeo, Luis Merino, Fernando Caballero

4D millimeter-wave (mmWave) radars are increasingly used in robotics, as they offer robustness against adverse environmental conditions. Besides the usual XYZ position, they provide Doppler velocity measurements as well as Radar Cross Section (RCS) information for every point. While Doppler is widely used to filter out dynamic points, RCS is often overlooked

Michal Töpfer, František Plášil, Tomáš Bureš, Petr Hnětynka

Vibe coding inherently assumes iterative refinement of LLM-generated code through feedback loops. While effective for conventional software tasks, its reliability in runtime-adaptive systems is unclear -- especially when generated code is not manually inspected. This paper studies feedback-based automated verification of LLM-generated adaptation managers in

Meng-Xun Li, Wen-Hui Deng, Zhi-Xing Wu, Chun-Xiao Jin

Vision-Language Models (VLMs) have demonstrated significant potential in medical image analysis, yet their application in intraoral photography remains largely underexplored due to the lack of fine-grained, annotated datasets and comprehensive benchmarks. To address this, we present MetaDent, a comprehensive resource that includes (1) a novel and large-scale

Xuanli He, Bilgehan Sel, Faizan Ali, Jenny Bao

Large Language Models (LLMs) are increasingly exposed to adaptive jailbreaking, particularly in high-stakes Chemical, Biological, Radiological, and Nuclear (CBRN) domains. Although streaming probes enable real-time monitoring, they still make systematic errors. We identify a core issue: existing methods often rely on a few high-scoring tokens, leading to fal

Hongsik Kim, Sunghwan Jung

Interfacial deformation under electric fields is a common phenomenon in many industrial processes. Particularly, we are interested in the dynamics of sessile soap bubbles in a parallel-plate electric field which exhibits a stable deformation regime followed by conical instability. Using side-view imaging, we track the equilibrium shapes, the transition to th

Yifan Le

Constrained decoding is widely used to make large language models produce structured outputs that satisfy schemas such as JSON. Existing work mainly treats schemas as structural constraints, overlooking that schema-key tokens also enter the autoregressive context and may guide generation. To the best of our knowledge, we present the first systematic study of

Xu Du, Boyu Han, Ivano Notarnicola, Karl H. Johansson

This paper investigates distributed resource allocation optimization over directed graphs with limited communication bandwidth. We develop a novel distributed algorithm that integrates the centralized Proximal Jacobian Alternating Direction Method of Multipliers (PJ-ADMM) with a finite-level quantized consensus scheme, enabling nodes to cooperatively solve t

Ravi Kiran Kadaboina

As autonomous AI agents increasingly call other agents to complete tasks on behalf of a human principal, a structural accountability gap has emerged: the calling agent accepts the terms of service of the callee without any protocol-level mechanism to prove that it understood those terms or that it subsequently honoured them. Authentication protocols such as

Yasin Abbasi-Yadkori, Peter L. Bartlett, Victor Gabillon, Alan Malek

We study bandit best-arm identification with arbitrary and potentially adversarial rewards. A simple random uniform learner obtains the optimal rate of error in the adversarial scenario. However, this type of strategy is suboptimal when the rewards are sampled stochastically. Therefore, we ask: Can we design a learner that performs optimally in both the stoc

Mare Hirose, Shoko Imaizumi, Hitoshi Kiya

This paper proposes a novel privacy-preserving semantic segmentation method that can use independent keys for each client and image. In the proposed method, the model creator and each client encrypt images using locally generated keys, and model training and inference are conducted on the encrypted images. To mitigate performance degradation, an image encryp

Juraj Krsnik, Dino Novko, Fabian B. Kugler, Osor S. Barišić

Plasmon modes, their dispersion, and the onset of damping when approaching the electron-hole continuum are well understood when electron correlations are weak. However, we know little about how this picture is modified and what additional features emerge in strongly correlated materials. Here, we present a fully ab initio approach to plasmon excitations that

Zhonghao Yang, Yu Li, Yanxu Zhu, Tianyi Zhou

As agent systems move into increasingly diverse execution settings, trajectory-level safety evaluation and diagnosis require benchmarks that evolve with them. ATBench is a diverse and realistic agent trajectory benchmark for safety evaluation and diagnosis. This report presents ATBench-Claw and ATBench-Codex, two domain-customized extensions that carry ATBen

Georg Dorndorf, Daniel Adolfsson, Masrur Doostdar

Automotive 4D imaging radar is well suited for operation in dusty and low-visibility environments, but scan registration remains challenging due to scan sparsity and spurious detections caused by noise and multipath reflections. This difficulty is compounded in feature-poor open-pit mines, where the lack of distinctive landmarks reduces correspondence reliab

Mayeul Arminjon

The theory considered interprets gravity as a pressure force. Thus, the scalar gravitational field defines the gravity acceleration field. However, it also determines the relation between the flat ``background metric'' and a curved ``physical metric''. Here we derive the equations of motion of the mass centers of a system of weakly gravitating bodies in the

Liesbeth Allein, Nataly Pineda-Castañeda, Andrea Rocci, Marie-Francine Moens

Understanding climate change requires reasoning over complex causal networks. Yet, existing causal discovery datasets predominantly capture explicit, direct causal relations. We introduce ClimateCause, a manually expert-annotated dataset of higher-order causal structures from science-for-policy climate reports, including implicit and nested causality. Cause-

Yildiz Berk Ates

We investigate the non-monotonic temperature sensitivity of a coherently driven two-level quantum system coupled to an Ohmic phonon environment. By employing a unitary polaron transformation, we account for phonon-induced renormalization effects that go beyond the standard weak-coupling approximations. Our analysis reveals that the Quantum Fisher Information

Armin Gießler, Pol Jané-Soneira, Sören Hohmann

We study state-feedback design for continuous-time LTI systems with a control input and an external input-output pair. Our objective is to determine feedback gains that render the closed-loop system (strictly) passive with respect to the external port while minimizing the standard LQR cost in the disturbance-free case. The resulting constrained optimization

Zhiyuan Zhai, Bingcong Li, Bingnan Xiao, Ming Li

Test-time compute scaling, the practice of spending extra computation during inference via repeated sampling, search, or extended reasoning, has become a powerful lever for improving large language model performance. Yet deploying these techniques under finite inference budgets requires a decision that current systems largely ignore: which inputs deserve mor

Annie Millet, Svetlana Roudenko

We investigate the focusing and defocusing energy-critical stochastic nonlinear Schr\"odinger equation, subject to random perturbations in the form of either additive or multiplicative (Stratonovich) noise. We establish local well-posedness for random or deterministic initial data $u_0$ in $\dot{H}^1(\mathbb{R}^n)$ or $H^1(\mathbb{R}^n)$, depending on the no

Zhenhao Cai, Eviatar B. Procaccia, Yuan Zhang

We present and study the Pool model in $\mathbb{R}^2$, a rotationally symmetric analogue of Multi-Particle Diffusion-Limited Aggregation (MDLA), in which particles ("droplets") perform continuous-time random walks and are absorbed upon entering a circular pool initially centered at the origin. Each absorbed particle increases the pool's mass, and the pool ex

Emil Benedykciuk, Marcin Denkowski, Grzegorz M. Wójcik

Purpose: Adaptive skip modules can improve medical image segmentation, but searching for them is computationally costly. Implantable Adaptive Cells (IACs) are compact NAS modules inserted into U-Net skip connections, reducing the search space compared with full-network NAS. However, the original IAC framework still requires a 200-epoch differentiable search

Benedikt Remlein, Massimiliano Esposito, Francesco Avanzini

We address a fundamental question: under which conditions do the dynamics and thermodynamics of open chemical reaction networks (CRNs), grounded on the notion of idealized chemostats that exchange selected species, emerge from underlying closed CRNs? While open CRNs provide the standard framework to describe out-of-equilibrium chemical systems, real systems

Yi Zhao, Yajuan Peng, Cam-Tu Nguyen, Zuchao Li

Large Reasoning Models (LRMs) achieve strong performance on complex tasks through extended chains of thought but suffer from high inference latency due to autoregressive reasoning. Recent work explores using Small Reasoning Models (SRMs) to accelerate LRM inference. In this paper, we systematically characterize the capability boundaries of SRMs and identify

Haileab Yagersew

Retail theft costs the global economy over \$100 billion annually, yet existing AI-based detection systems require expensive custom model training on proprietary datasets and charge \$200-500/month per store. We present Paza, a zero-shot retail theft detection framework that achieves practical concealment detection without training any model. Our approach or

Bin Han, Hao Chen, Muxia Sun, H. V. Poor

We study matched and Euclidean-mismatched decoding on finite Fourier-curve constellations with tangent-space artificial noise. Each hypothesis induces a Gaussian law with symbol-dependent rank-one covariance. We derive exact Euclidean pairwise errors for arbitrary pairs and an exact Gaussian-expectation representation for matched decoding on bilaterally tang

Yufeng Wu

Behavioral Profile (BP) annotation is difficult to automate because it requires simultaneous coding across multiple linguistic dimensions. We treat BP annotation as a bundle of annotation skills rather than a single task and evaluate LLM-assisted BP annotation from this perspective. Using 3,134 concordance lines of 30 Chinese metaphorical color-term derivati

Mahsa Farjadnia, Katayoun Eshkofti, Albin Apell, Tilde Hjalmarsson

This paper investigates non-intrusive occupancy detection methods for residential buildings using environmental sensor data from the KTH Live-In Lab in Stockholm, Sweden. Three machine learning approaches, namely, logistic regression (LR), support vector machines (SVM), and long short-term memory (LSTM) network enhanced with an attention mechanism, are evalu

Pavel Martynyuk

We study the minimization problem for eigenvalues of the Dirac operator within a fixed conformal class on a closed spin Riemannian manifold. We establish a criterion for the existence of a minimizer for this variational problem, focusing specifically on the case of closed surfaces. Furthermore, we apply our results to derive isoperimetric inequalities for th

Jinfeng Xu, Zheyu Chen, Shuo Yang, Jinze Li

Recent advancements in multimodal recommendations, which leverage diverse modality information to mitigate data sparsity and improve recommendation accuracy, have gained significant attention. However, existing multimodal recommendations overlook the critical role of user representation initialization. Unlike items, which are naturally associated with rich m

Vincenzo Yuto Civale, Roberto Semeraro, Andrew David Bagdanov, Alberto Magi

Current single-cell foundation model benchmarks universally extract final layer embeddings, assuming these represent optimal feature spaces. We systematically evaluate layer-wise representations from scFoundation (100M parameters) and Tahoe-X1 (1.3B parameters) across trajectory inference and perturbation response prediction. Our analysis reveals that optima

Zihan Li, Alexander Kotz, Adrian Schwarzenberger, Christian Koos

Modern communication networks demand ever-increasing transmission bandwidth, placing stringent requirements on low-cost, high-performance electro-optic modulators. Substantial advances have been made in integrated photonics employing lithium niobate on insulator. In contrast, photonic integrated circuits based on lithium tantalate -- a material already comme

Konstantinos Efstathiou, Gabriela Jocelyn Gutierrez-Guillen, Pavao Mardešić, Dominique Sugny

Singular Lagrangian fibrations arising from three-degree-of-freedom integrable Hamiltonian systems remain largely unexplored. While several results describe the global structure of large classes of systems with two degrees of freedom, only a few examples are understood in higher dimensions. We present a three-degree-of-freedom system derived from the two-spi

Yuen-Fui Lau, Qihan Zhao, Yinhuai Wang, Runyi Yu

Recent advancements in whole-body control through deep reinforcement learning have enabled humanoid robots to achieve remarkable progress in real-world chal lenging locomotion skills. However, existing approaches often struggle with flexible transitions between distinct skills, cre ating safety concerns and practical limitations. To address this challenge, w

Lei Guo, Hongyun Yang, Pengjie Ren, Tong Chen

Large Language Models have shown great success in recommender systems. However, the limited and sparse nature of user data often restricts the LLM's ability to effectively model behavior patterns. To address this, existing studies have explored cross-domain solutions by conducting Cross-Domain Recommendation tasks. But previous methods typically assume domai

Pooriya Jamie, Amir Ghasemian, Homa Hosseinmardi

Recommender systems on social media increasingly mediate how users encounter mental health content, yet it remains unclear whether they distinguish help-seeking from distress expression. We conduct a controlled 7-day audit of TikTok's "For You" page using 30 fresh accounts and LLM-guided agents that vary initial search framing (distress- vs. help-initiated)

Da Chen, Pietro Andreozzi, Giulia Frigerio, Daniele Perilli

In this systematic density functional theory study, we compare a standard gradient corrected functional (PBE) with a long-range hybrid functional (HSE06), with and without correction for the dispersion forces, relative to their ability to correctly reproduce structural and electronic properties of different bulk 3D C3N4 phases, encompassing diamond- and grap

Anjan Daimari, Diana Thongjaomayum

The response of a complex system to a slow varying external force often displays a jump discontinuity in the order parameter near the critical point. However, this discontinuity is not usually a single jump but rather breaks into smaller jumps which makes it difficult to locate the critical point on approaching its vicinity based only on simulations, in the

Bhavik Vachhani, Kush Shrisvastava, Pranshu Nema, Sai Chiranthan

Evaluating large language models (LLMs) for clinical documentation tasks such as SOAP note generation remains challenging. Unlike standard summarization, these tasks require clinical abstraction, normalization of colloquial language, and medically grounded inference. However, prevailing evaluation methods including automated metrics and LLM as judge framewor

Dinghao Li, Wenlong Zhou, Zhimin Chen, Yuehan Peng

Educational assistants should spend more computation only when the task needs it. This paper rewrites our earlier draft around the system that was actually implemented and archived in the repository: a sample-level 1B to 7B cascade for the shared-8 EduBench benchmark. The final system, Pangu-ACE, uses a 1B tutor-router to produce a draft answer plus routing

DL Ferrario

A. Albouy and R. Moeckel in 2000 found some interesting inequalities related to the inverse problem for collinear (Moulton) central configurations: the Pfaffian of a certain matrix is positive since all coefficients of some polynomials are positive, for the Newtonian (interaction potential $1/r$ and $n\leq 6$). They conjectured that for all $n$ such Pfaffian

Cristina Blaga, Paul A. Blaga

Weyl conformal gravity was originally proposed in the early twentieth century as an attempt to unify gravitation and electromagnetism. Since 1989, renewed interest in this fourth-order theory of gravity has emerged following the discovery of several exact black hole solutions. In this work, we investigate the timelike circular geodesics of a spherically symm

Yifan Zhao, Yuchen Yang, Matei Budiu, Sasa Misailovic

We present Nautilus, a novel tensor compiler that moves toward fully automated math-to-kernel optimization. Nautilus compiles a high-level algebraic specification of tensor operators into efficient tiled GPU kernels. Nautilus's successive lowering design allows high-level optimizations, expression rewrites, and tile optimizations to be jointly applied in a s

Tsunehide Kuroki

We analytically study the effect of backreaction from analog Hawking radiation on its entanglement entropy in the Bose-Einstein condensate (BEC). The backreaction is expected to play an essential role in the decrease of the entanglement entropy and in realizing the Page curve. Since the BEC theory has microscopic Hamiltonian and thus exhibits unitarity, it i

Ramgopal Madugani, Amal Jose, Christophe Pin, Metin Ozer

Similar to microspheres, thin-walled microbubble resonators support whispering gallery modes (WGMs) that combine ultrahigh Q-factors and small effective mode volumes. In contrast, their hollow nature enables enhanced interactions with encapsulated materials and lower spectral mode density due to the tight radial confinement of the optical modes. However, the

Hao Han, Jin Xie, Xuehao Ma, Weiquan Zhu

Resolving real-world software engineering (SWE) issues with autonomous agents requires complex, long-horizon reasoning. Current pipelines are bottlenecked by unoptimized demonstration data, sparse execution rewards, and computationally prohibitive inference scaling, which collectively exacerbate token bloat, reward hacking, and policy degradation. We present

Jinjin Li, Chaoying Zhao

We propose an orbital angular momentum (OAM) quantum holography scheme based on multi-mode Bessel-Gaussian (MBG) beams. Entangled photon pairs are generated through spontaneous parametric down-conversion (SPDC) process, and the axis prism parameters and topological charges of the idler photons are used for encoding to construct Bessel-Gaussian quantum select

Benito J. González, Emilio R. Negrín

This note presents a new equivalence to the Riemann Hypothesis by means of the Salem integral equation.

Jiwon Lee, Hugo Matias, Daniel Silvestre, Thinh T. Doan

Safe navigation for an ego vehicle in uncertain environments characterized by dynamic obstacles with unknown nonlinear dynamics is a challenging problem of significant practical interest. Existing approaches in the literature either lack formal safety guarantees, require full model knowledge, or fail to account for the risk associated with the vehicle's exac

Yue Zhao, Bin Xiao, Jiawei Liu, Hui Zeng

Magnetic tunnel junction (MTJ) is the key component to enable information access and increasing number of MTJs is integrated to develop high-density spintronic devices. However, continuous miniaturization of the conventional MTJs is hindered by stray magnetic fields. Altermagnets, combining the advantages of both ferromagnets and antiferromagnets, provide a

Andrey Moskalenko, Alexey Bryncev, Ivan Kosmynin, Kira Shilovskaya

This paper presents an overview of the NTIRE 2026 Challenge on Video Saliency Prediction. The goal of the challenge participants was to develop automatic saliency map prediction methods for the provided video sequences. The novel dataset of 2,000 diverse videos with an open license was prepared for this challenge. The fixations and corresponding saliency map

Rami Luisto, Liisa Petäinen, Tommi Grönholm, Jan Böhm

In NLP classification tasks where little labeled data exists, domain fine-tuning of transformer models on unlabeled data is an established approach. In this paper we have two aims. (1) We describe our observations from fine-tuning the Finnish BERT model on Finnish medical text data. (2) We report on our attempts to predict the benefit of domain-specific pre-

S. Sakinah, Sang-Ho Kim, H. M. Choi

We investigate near-threshold $J/\psi$ photoproduction off the nucleon, focusing on hadronic rescattering effects induced by open-charm meson-baryon intermediate states. Beyond the conventional Pomeron-exchange mechanism, the $\bar D^0\Lambda_c^+$ and $\bar D^{*0} \Lambda_c^+$ channels are incorporated within an effective Lagrangian framework. The relevant p

Ovaisa Jan, Idrees Qasim

This paper establishes new upper bounds for the right eigenvalues of monic matrix polynomials over the quaternion division algebra. The noncommutative nature of quaternion multiplication presents fundamental challenges in eigenvalue analysis, distinguishing this problem from the classical complex case. We use spectral norm inequalities for partitioned quater

Juan Carlos del Valle, Paul Bergold, Karolina Kropielnicka

We present a time-dependent extension of logarithmic perturbation theory for nonrelativistic quantum dynamics governed by the Schrödinger equation, in which the logarithm of the wave function is expanded in powers of a coupling constant. The resulting hierarchy of equations defining the perturbative corrections is governed by a gauge-rotated Hamiltonian of t

Can Karacelebi, Yusuf Talha Sahin, Elif Surer, Ertan Onur

Ad hoc wireless networks exhibit complex, innate and coupled dynamics: node mobility, energy depletion and topology change that are difficult to model analytically. Model-free deep reinforcement learning requires sustained online interaction whereas existing model based approaches use flat state representations that lose per node structure. Therefore we prop

Connie Trojan, Pavel Myshkov, Paul Fearnhead, James Hensman

In online clustering problems, there is often a large amount of uncertainty over possible cluster assignments that cannot be resolved until more data are observed. This difficulty is compounded when clusters follow complex distributions, as is the case with text data. Sequential Monte Carlo (SMC) methods give a natural way of representing and updating this u

Shahar Cohen, David M. Steinberg, Yael Radzyner, Yochai Ben Horin

We study a classification problem with three key challenges: pervasive informative missingness, the integration of partial prior expert knowledge into the learning process, and the need for interpretable decision rules. We propose a framework that encodes prior knowledge through an expert-guided class-conditional model for one or more classes, and use this m

Zeguan Xiao, Siqing Li, Yong Wang, Xuetao Wei

Machine unlearning for large language models (LLMs) aims to remove targeted knowledge while preserving general capability. In this paper, we recast LLM unlearning as an asymmetric two-task problem: retention is the primary objective and forgetting is an auxiliary. From this perspective, we propose a retention-prioritized gradient synthesis framework that dec

Hyunwoo Kim, Harin Yu, Hanau Yi

The rapid integration of large language models (LLMs) into everyday workflows has transformed how individuals perform cognitive tasks such as writing, programming, analysis, and multilingual communication. While prior research has focused on model reliability, hallucination, and user trust calibration, less attention has been given to how LLM usage reshapes

Jieyi Wang, Yazhe Niu, Dexuan Xu, Zhongyu Wei

Recent Large Audio Language Models have demonstrated impressive capabilities in audio understanding. However, they often suffer from perceptual errors, while reliable audio reasoning is unattainable without first grounding the model's perception in structured auditory scenes. Inspired by Auditory Scene Analysis, we first introduce a Perception-Aware Question

Yili Ren, Shiqi Wen, Li Hou, Dingwen Xiao

Grain-edge segmentation (GES) and lithology semantic segmentation (LSS) are two pivotal tasks for quantifying rock fabric and composition. However, these two tasks are often treated separately, and the segmentation quality is implausible albeit expensive, time-consuming, and expert-annotated datasets have been used. Recently, foundation models, especially th

Xinjie Jiang, Shengliang Pan, Yanlong Zhang

In this paper, inspired by the work of Guan and Li (2015), we introduce a fourth-order centro-equiaffine invariant curve flow via the affine Minkowski formula. Without any smallness assumptions on the initial curve, we establish the long-time existence of the flow and prove that, as $t \to +\infty$, the evolving curve preserves its enclosed area and converge

Jonathan Frey, David Kiessling, Katrin Baumgärtner, Moritz Diehl

Although Anderson acceleration (AA) is known to speed up fixed-point iterations, it is rarely applied in constrained optimization, in particular sequential quadratic programming (SQP). We show that the local convergence behavior of a general family of (inexact) SQP-type methods can benefit from AA and introduce a simple heuristic to alleviate slower converge

Linlin Ye, Zhaoqi Wu, Shao-Ming Fei

Quantum coherence is a fundamental issue in quantum mechanics and quantum information processing. We explore the coherence dynamics of the evolved states in HHL quantum algorithm for solving the linear system of equation $A\overrightarrow{x}=\overrightarrow{b}$. By using the Tsallis relative $\alpha$ entropy of coherence and the $l_{1,p}$ norm of coherence,

Marco Schlimbach, Moritz Rempe, Jessica Mnischek, Lukas T. Rotkopf

Objective. Standard Magnetic Resonance Imaging (MRI) reconstruction pipelines discard phase information captured during acquisition, despite evidence that it encodes tissue properties relevant to tumor diagnosis. Current machine learning approaches inherit this limitation by operating exclusively on reconstructed magnitude images. The aim of this study is to

Nishanth Madhusudhan, Vikas Yadav, Alexandre Lacoste

Effective abstention (EA), recognizing evidence insufficiency and refraining from answering, is critical for reliable multimodal systems. Yet existing evaluation paradigms for vision-language models (VLMs) and multi-agent systems (MAS) assume answerability, pushing models to always respond. Abstention has been studied in text-only settings but remains undere

David Amaro-Alcalá, Carlos Pineda

We demonstrate that the Ising all-to-all (ATA) model exhibits a range of dynamics, from integrable to chaotic, including mixed behaviour across symmetry blocks within a single system. While other works have explored the dynamics of all-to-all systems by varying parameters, we analyse a fixed set of parameters and examine the dynamics within different blocks.

Luiz M. Faria, Carlos Perez-Arancibia, Svetlana Tlupova

This paper extends and analyzes the high-order kernel regularization framework of Beale & Tlupova (arXiv:2510.13639) to all four on-surface boundary integral operators of the Helmholtz Calderon calculus in three dimensions: the single-layer, double-layer, adjoint double-layer, and hypersingular operators. To the best of our knowledge, this work provides the