December 2025 arXiv papers — page 28

Vlad Temkin, Zack Weinstein, Ruihua Fan, Daniel Podolsky

We investigate the statistical physics of quantum error correction in ${\rm U}(1)$ symmetry-enriched topological quantum memories. Starting from a phenomenological error model of charge-conserving noise, we study the optimal decoder assuming the local charges of each anyon can be measured. The error threshold of the optimal decoder corresponds to a continuou

Zhi Ouyang, Dian Zheng, Xiao-Ming Wu, Jian-Jian Jiang

Inversion-based visual editing provides an effective and training-free way to edit an image or a video based on user instructions. Existing methods typically inject source image information during the sampling process to maintain editing consistency. However, this sampling strategy overly relies on source information, which negatively affects the edits in th

Esha Kundu

Supernovae (SNe), the catastrophic end of stars' lives, are among the most energetic phenomena in the universe. Mapping the aftermath of the explosions to the properties of pre-SN stars is challenging due to the lack of knowledge about the evolution of different types of stars. The immediate surroundings of pre-SN stars carry the signature of the progenitors

Yue Wu, Meng-Yuan Li, Chengshu Li, Hui Zhai

The hypergraph product (HGP) construction of quantum error-correcting codes (QECC) offers a general and explicit method for building a QECC from two classical codes, thereby paving the way for the discovery of good quantum low-density parity-check codes. In this letter, we propose a general and explicit construction recipe for QECCs from a total of D classic

Massimiliano Berti

We review recent advances regarding the long-time dynamics of space-periodic water waves, focusing on 1) bifurcation of quasi-periodic solutions, both standing and traveling; 2) long-time well-posedness results; 3) modulational instability of Stokes waves. These results rely on unconventional approaches to KAM and Birkhoff normal form theories for Hamiltonia

Matthew Riemer, Erik Miehling, Miao Liu, Djallel Bouneffouf

Although parameter-efficient fine-tuning methods, such as LoRA, only modify a small subset of parameters, they can have a significant impact on the model. Our instruction-tuning experiments show that LoRA-based supervised fine-tuning can catastrophically degrade model capabilities, even when trained on very small datasets for relatively few steps. With that

Mengkang Hu, Bowei Xia, Yuran Wu, Ailing Yu

Symbolic world models (e.g., PDDL domains or executable simulators) are central to model-based planning, but training LLMs to generate such world models is limited by the lack of large-scale verifiable supervision. Current approaches rely primarily on static validation methods that fail to catch behavior-level errors arising from interactive execution. In th

Linda M. Carpenter, Katherine Schwind

We explore models where single new exotic states interact with the Standard Model through an asymmetric Standard Model portal with couplings to at least one gluon and one lepton. We consider the complete set of effective operators up to dimension 6, and examine a few additional dimension 7 operators that contain interesting field content or potential collide

Sachin Gupta, Matthew B. Weiss

Informationally complete (IC) measurements are fundamental tools in quantum information processing, yet their physical implementation remains challenging. By the Naimark extension theorem, an IC measurement may be realized by a von Neumann measurement on an extended system after a suitable interaction. In this work, we elaborate on a simple algorithm for rea

Agniv Roy Choudhury

Pneumonia is a leading cause of mortality in children under five, with over 700,000 deaths annually. Accurate diagnosis from chest X-rays is limited by radiologist availability and variability. Objective: This study compares custom CNNs trained from scratch with transfer learning (ResNet50, DenseNet121, EfficientNet-B0) for pediatric pneumonia detection, eva

Mohamed Shalma, Engy Aly Maher, Ahmed El-Mahdy

Active Reconfigurable Intelligent Surfaces (RIS) are a promising technology for 6G wireless networks. This paper investigates a novel hybrid deep reinforcement learning (DRL) framework for resource allocation in a multi-user uplink system assisted by multiple active RISs. The objective is to maximize the minimum user rate by jointly optimizing user transmit

Zubair Shah, Noaman Khan

Neural network pruning is widely used to reduce model size and computational cost. Yet, most existing methods treat sparsity as an externally imposed constraint, enforced through heuristic importance scores or training-time regularization. In this work, we propose a fundamentally different perspective: pruning as an equilibrium outcome of strategic interacti

Haley Rosso, Talea Mayo

For complex simulation problems, inferring parameters often precludes the use of classical likelihood-based techniques due to intractable likelihoods. Simulation-based inference (SBI) methods offer a likelihood-free approach to directly learn posterior distributions $p(\bftheta \mid \xobs)$ from simulator outputs. Recently, diffusion models have emerged as p

Abhinav Parmar, Abhisek Panigrahi, Abhishek Kumar Dwivedi, Abhishek Bhattacharya

We present Mify-Coder, a 2.5B-parameter code model trained on 4.2T tokens using a compute-optimal strategy built on the Mify-2.5B foundation model. Mify-Coder achieves comparable accuracy and safety while significantly outperforming much larger baseline models on standard coding and function-calling benchmarks, demonstrating that compact models can match fro

Tomasz Denkiewicz, Hussain Gohar

We investigate the observational tests of generalized mass-to-horizon entropic cosmology by incorporating large-scale structure growth data in addition to purely geometric probes. The theoretical framework is constructed from a generalized mass-to-horizon scaling relation, $M \propto L^n$, which implies a corresponding generalized entropic functional $S_n \p

Zhaozhao Ma, Shujian Yu

Multimodal regression aims to predict a continuous target from heterogeneous input sources and typically relies on fusion strategies such as early or late fusion. However, existing methods lack principled tools to disentangle and quantify the individual contributions of each modality and their interactions, limiting the interpretability of multimodal fusion.

Shuyu Gan, Renxiang Wang, James Mooney, Dongyeop Kang

Automating end-to-end data science pipeline with AI agents still stalls on two gaps: generating insightful, diverse visual evidence and assembling it into a coherent, professional report. We present A2P-Vis, a two-part, multi-agent pipeline that turns raw datasets into a high-quality data-visualization report. The Data Analyzer orchestrates profiling, propos

Duygu Altinok

Evaluating the performance of various model architectures, such as transformers, large language models (LLMs), and other NLP systems, requires comprehensive benchmarks that measure performance across multiple dimensions. Among these, the evaluation of natural language understanding (NLU) is particularly critical as it serves as a fundamental criterion for as

Wei-Ming Chen, Yu-tin Huang, Zi-Xun Huang, Yohan Liu

In this paper we bootstrap de Sitter wavefunction coefficients (WFCs) involving fermionic operators. Starting with a fixed total-energy pole order, we systematically impose the conformal Ward identities (CWI) together with cutting-rule constraints. We derive the relevant cutting rules for fermionic exchange for the first time, enabling a complete determinati

Xiaofeng Mao, Zhen Li, Chuanhao Li, Xiaojie Xu

Recent approaches have demonstrated the promise of using diffusion models to generate interactive and explorable worlds. However, most of these methods face critical challenges such as excessively large parameter sizes, reliance on lengthy inference steps, and rapidly growing historical context, which severely limit real-time performance and lack text-contro

Alan A. Tedeev

We consider the Cauchy problem for the nonstationary discrete p-Laplacian with inhomogeneous density \r{ho}(x) on an infinite graph which supports the Sobolev inequality. For nonnegative solutions when p > 2, we prove the precise rate of stabilization in time, provided \r{ho}(x) is a non-power function. When p > 2 and \r{ho}(x) goes to zero fast enough, we p

Olaide N. Oyelade, Oliver Hoxey, Yulia Humrye

The popular use of histopathology images, such as hematoxylin and eosin (H&E), has proven to be useful in detecting tumors. However, moving such cancer cases forward for treatment requires accurate on the amount of the human epidermal growth factor receptor 2 (HER2) protein expression. Predicting both the lower and higher levels of HER2 can be challenging. M

Andrei Angelescu, Andreas Bally, Florian Goertz, Sascha Weber

We present a concise survey of the running of gauge couplings in realistic models of gauge-Higgs grand unification in a slice of AdS$_5$ space and investigate their potential unification. Besides unifying the gauge groups of the Standard Model, these models can address various unresolved puzzles, such as the lightness of the Higgs boson and the strong hierar

Bo-Ting Chen, Wei-Ming Chen, Yu-tin Huang, Zi-Xun Huang

In this work we analyze the analytic structure of tree-level flat-space wavefunction coefficients (WFCs), with particular attention to fermionic operators, and derive cutting rules for internal-fermion lines. Building on these results, we set up an iterative procedure that, starting from the flat-space S-matrix, reconstructs the 3- and 4-point WFCs with the

Yiheng Wang, Yixin Chen, Shuo Li, Yifan Zhou

We introduce SciEvalKit, a unified benchmarking toolkit designed to evaluate AI models for science across a broad range of scientific disciplines and task capabilities. Unlike general-purpose evaluation platforms, SciEvalKit focuses on the core competencies of scientific intelligence, including Scientific Multimodal Perception, Scientific Multimodal Reasonin

Aleksandr V. Pukhlikov

In this paper we study two families of three-dimensional quartics in the complex projective space ${\mathbb P}^4$: hypersurfaces with a unique quadratic singularity of rank 3, which is resolved by two blowups, and hypersurfaces with two quadratic singularities of rank 3 and 4, respectively. Both families have codimension 3 in the natural parameter space. For

Quentin Michaud, Sara Ramezanian, Dhouha Ayed, Olivier Levillain

Trusted Execution Environments (TEEs) protect sensitive code and data from the operating system, hypervisor, or other untrusted software. Different solutions exist, each proposing different features. Abstraction layers aim to unify the ecosystem, allowing application developers and system administrators to leverage confidential computing as broadly and effic

Ricardo Vasquez, Diego Riofrío-Luzcando, Joe Carrion-Jumbo, Cesar Guevara

Emotions are one of the important components of the human being, thus they are a valuable part of daily activities such as interaction with people, decision making and learning. For this reason, it is important to detect, recognize and understand emotions using computational systems to improve communication between people and machines, which would facilitate

Shukai Liu, Jian Yang, Bo Jiang, Yizhi Li

Agents based on large language models have recently shown strong potential on real-world software engineering (SWE) tasks that require long-horizon interaction with repository-scale codebases. However, most existing agents rely on append-only context maintenance or passively triggered compression heuristics, which often lead to context explosion, semantic dr

Urvi Parekh, Nadiia Didukh, Samira Dabelstein, Ronja Piehler

Copper selenide is an exceptional quasi-layered monolithic material that exhibits both semiconducting and metallic properties in adjacent visible and near-infrared (NIR) spectral ranges. Here we introduce a thiol-free colloidal synthesis for generating quasi-2D klockmannite copper selenide nanocrystals via hot injection method, achieving shape control by tun

Petar Suman, Dong-Gang Wang, Wuhyun Sohn, James R. Fergusson

Signatures of massive particles during inflation are highly informative targets for cosmological experiments. With recent progress on both theoretical and observational frontiers, we have reached the point where these novel signals of primordial non-Gaussianities (PNG) can be systematically tested with increasingly precise data. In this paper, we present the

John M. Mango, Ronald Katende

We consider the problem of restoring linear conservation laws in data-driven linear dynamical models. Given a learned operator $\widehat{A}$ and a full-rank constraint matrix $C$ encoding one or more invariants, we show that the matrix closest to $\widehat{A}$ in the Frobenius norm and satisfying $C^\top A = 0$ is the orthogonal projection $A^\star = \wideha

Alejandro Buitrago López, Alberto Ortega Pastor, David Montoro Aguilera, Mario Fernández Tárraga

Research on online social networks (OSNs) is often hindered by platform opacity, limited access to data, and ethical constraints. Simulation offer a valuable alternative, but existing frameworks frequently lack realism and explainability. This paper presents a simulation framework that models synthetic social networks with agents endowed with demographic-bas

T. Gent, S. Huber, K. Mimasu, J. M. No

We perform a detailed investigation of the viable baryogenesis parameter space of a non-minimal Higgs sector consisting of two Higgs doublets and a singlet pseudoscalar (2HDM$+a$). In such a model, an early Universe period of transient CP violation may occur, driven by a nonvanishing vacuum expectation value of the CP-odd scalar $a$. This naturally avoids th

Rafael Cavalcanti

We relate the novel concept of Topological Data Analysis in Finsler space with representability property, which is a natural obstruction to prevent spurious features in high dimensions. We use decomposition of integer matrix in order to find suitable prime integer $p$ such that persistent homology module over $\mathbb{Z}_p$ encompasses only the holes associa

Naw Sai

We investigate whether a 3-$\delta$ system with positive coupling strengths can approximate the transmission spectrum of a 2-$\delta$ resonance system with opposite-sign couplings for $k <3$. Theoretical analysis establishes exact isospectrality -- perfectly matched transmission spectrum -- is impossible for physically non-trivial configurations, while numer

Takuo Matsuoka

The purpose of these notes is to collect in one place some facts on the category of finite totally ordered sets and some related categories. More specifically, we collect some results on them which will be useful for the study of iteratedly meta theories of algebra in the style of our work arXiv:1601.00301 (arXiv:1601.00301), which is a kind of higher order

Hui Zhang, Tarik Hadzibeganovic, Xiao-Pu Han

Habitat loss and fragmentation have often been viewed as major threats to species interaction and global biodiversity conservation. However, habitat degradation can also give rise to positive ecological and behavioral responses, challenging the notion that its consequences are entirely detrimental. While controlling for the degree of total habitat loss, we s

Nir Somech, Guy Katz

Software obfuscation techniques make code more difficult to understand, without changing its functionality. Such techniques are often used by authors of malicious software to avoid detection. Reverse Engineering of obfuscated code, i.e., the process of overcoming obfuscation and answering questions about the functionality of the code, is notoriously difficul

Daiki Saito, Koki Tokeshi

The excursion-set formalism enables us to infer the mass distribution of collapsed objects, such as primordial black holes (PBHs), by the language of stochastic processes. Within the framework, this article investigates how a smooth coarse-graining procedure affects the resulting PBH mass function. As a demonstrative example, we employ a Gaussian window func

Raphaël Cerf

We revisit the proof of the de Moivre--Laplace theorem, which is the ancestor of the central limit theorem for the binomial distribution. Our goal is to provide a proof that can be reasonably presented to undergraduate students within a basic course of probability theory. We follow the strategies presented in two classical references, the books of Breiman an

Dominik Krasula

For a semiperfect ring with essential socles, the Double annihilator property encodes that the top and socle have anti-isomorphic lattices of submodules, whereas the Size condition encodes that they are isomorphic as modules. Interest in both concepts, particularly for finite rings, was revived by coding theory, where they characterise QF rings and Frobenius

Homayon Anjomshoa, Behrouz Mirza, Alireza Azizallahi

We derive exact form of accelerating Fisher-Janis-Newman-Winicour (FJNW) metric by a simple perturbative method. We also argue that by using Buchdahl transformations one can obtain the same accelerating FJNW metric. We investigate singularities of the accelerating FJNW metric and study their effects on global and local structures of this spacetime. We also s

Nikolai Peters

The paper summarizes recent results from the Belle and Belle II collaborations on semileptonic $B$ decays measurements including inclusive and exclusive determination of Cabibbo-Kobayashi-Maskawa matrix elements $|V_{cb}|$ and $|V_{ub}|$ and lepton flavor universality tests studies. The results are based on the full Belle data and 361 fb$^{-1}$ Belle II data

Youran Ye, Dejin Wang, Ajinkya Bhandare

Projected Gradient Descent (PGD) is a strong and widely used first-order adversarial attack, yet its computational cost scales poorly, as all training samples undergo identical iterative inner-loop optimization despite contributing unequally to robustness. Motivated by this inefficiency, we propose \emph{Selective Adversarial Training}, which perturbs only a

Anastasios Papazafeiropoulos, Ioannis Bartsiokas, Dimitra I. Kaklamani, Iakovos S. Venieris

We derive a novel closed-form lower bound on the ergodic capacity of holographic multiple-input multiple-output (HMIMO) systems enhanced by stacked intelligent metasurfaces (SIMs) under Rayleigh fading conditions. The proposed expression is valid for systems with a finite number of antennas and SIM elements and exhibits tightness throughout the whole signal-

Tong Li, Luping Yu

We examine whether large language models (LLMs) hold systematic beliefs about environmental, social, and governance (ESG) issues and how these beliefs compare with-and potentially influence-those of human market participants. Based on established surveys originally administered to professional and retail investors, we show that major LLMs exhibit a strong pr

Madalina I Sas, Julian H J Sutherland

Cellular automata (CA) are quintessential ALife and ubiquitous in many studies of collective behaviour and emergence, from morphogenesis to social dynamics and even brain modelling. Recently, there has been an increased interest in formalising CA, theoretically through category theory and practically in terms of a functional programming paradigm. Unfortunate

Hannah Atmer, Yuan Yao, Thiemo Voigt, Stefanos Kaxiras

Energy consumption dictates the cost and environmental impact of deploying Large Language Models. This paper investigates the impact of on-chip SRAM size and operating frequency on the energy efficiency and performance of LLM inference, focusing on the distinct behaviors of the compute-bound prefill and memory-bound decode phases. Our simulation methodology

Zhiyao Sun, Ziqiao Peng, Yifeng Ma, Yi Chen

Real-time, streaming interactive avatars represent a critical yet challenging goal in digital human research. Although diffusion-based human avatar generation methods achieve remarkable success, their non-causal architecture and high computational costs make them unsuitable for streaming. Moreover, existing interactive approaches are typically restricted to

John Cardy

The dome of the Roman Pantheon is coffered with ribs surrounding sunken lacunaria, thus forming a grid. How this is achieved given the curvature of the dome has long been a subject for study and speculation. Although detailed measurements now exist, thus far no single principle has emerged which fixes the overall geometry. Similar coffering occurs in Hawksmo

Andrey Yu. Konyaev, Vladimir S. Matveev

We construct Lax pairs for the recently (2023) introduced integrable PDE systems known as the BKM equations. As many known and previously studied integrable systems are special cases of the BKM systems, our construction provides Lax pairs for many integrable hierarchies, including previously studied ones such as Camassa-Holm, Dullin-Gottwald-Holm, cKdV, Ito,

Sravan Karthick T

Bitcoin price forecasting is characterized by extreme volatility and non-stationarity, often defying traditional univariate time-series models over long horizons. This paper addresses a critical gap by integrating Global M2 Liquidity, aggregated from 18 major economies, as a leading exogenous variable with a 12-week lag structure. Using the TimeXer architect

Runli Li, Shaojing Liu, Ximiao Wang, Hongjia Zhu

Terahertz (THz) technology shows great potential in 6G communications and imaging, but faces challenges related to detector sensitivity, noise, and cryogenic operation. Here, we integrate interferometric enhancement of absorption (IEA) from a metal reflection layer with a graphene plasmon polariton atomic cavity (PPAC)-based photodetector. The hybrid configu

Gergely Buza, George Haller

Spectral submanifolds (SSMs) are invariant manifolds of a dynamical system, defined by the property of being tangent to a spectral subspace of the linearized dynamics at a steady state. We show existence, along with certain desirable properties such as smoothness, attractivity and conditional uniqueness, of SSMs associated to a large class of spectral subspa

A. V. Kopyev, A. S. Il'yn, V. A. Sirota, K. P. Zybin

Context: During the last decades, significant progress has been made in both numerical simulations of turbulent dynamo and theoretical understanding of turbulence. However, there is still lack of quantitative comparison between the simulations and the theory of the dynamo. Results: We study the critical magnetic Reynolds number ($Rm_c$) and the growth rate n

Shuo Yang, Ziyang Yu, Yiqi Wang, Lei Wang

In this work, we study the search for charged Higgs bosons in the Two-Higgs-Doublet Model plus an additional pseudo-scalar (2HDM+a) at the Compact Linear Collider (CLIC). Focusing on the pair production of charged Higgs bosons, followed by the decays $H^\pm \to a W $ and $ H^\mp \to t b $, we analyze the signal channel of $4j+2b+E_T^{miss}$. Given the center

Dhan Raj Lawati, Prem Bahadur Karki, Jitender Kumar, Karishma Prasad

Understanding and controlling spin dynamics in two-dimensional (2D) van der Waals (vdW) ferromagnets is essential for their application in magnonics and hybrid quantum platforms. Here, we investigate the spin dynamics of the vdW ferromagnet 1T-CrTe_{2} and demonstrate their systematic tunability via niobium (Nb) substitution in Cr_{1-x}Nb_{x}Te_{2}(x=0-0.2).

Kai Ren, Pengfei Ma, Minghui Hu, Junlong Tian

The Royer law is a widely used empirical relation for calculating alpha-decay half-lives; however, it requires 12 parity-dependent parameters.It exhibits systematic deviations near the shell closure. We propose an improved Royer law by adding a shell-correction term, an odd-even pairing indicator, and an orbital-angular-momentum contribution. This unified fr

Yuefeng Lin, Kun Wang, Qinyuan Zheng, Rui Zhang

MaxCut is a canonical NP-hard combinatorial optimization problem in graph theory with broad applications ranging from physics to bioinformatics. Although variational quantum algorithms offer promising new approaches that may eventually outperform classical schemes, they suffer from resource constraints and trainability issues such as barren plateaus, making

Ronald Katende

Stability analyses of modern learning systems are frequently derived under smoothness assumptions that are violated by ReLU-type nonlinearities. In this note, we isolate a minimal obstruction by showing that no uniform smoothness-based stability proxy such as gradient Lipschitzness or Hessian control can hold globally for ReLU networks, even in simple settin

Giuseppe De Palma, Saverio Giallorenzo

This volume contains the post-proceedings of the Workshop on Adaptable Cloud Architectures (WACA 2025), held on June 20, 2025, in Lille, France, co-located with DisCoTec 2025 - 20th International Federated Conference on Distributed Computing Techniques.

V. S. Shalgin

In this paper, we consider the problem of local parameter identifiability of a parameter function in a system of ordinary differential equations. Previously, in this problem, the case where the dimensions of a parameter and a solution of a system coincide was considered, and a specific class of systems was identified, for which sufficient conditions for loca

Robynn Corveleyn, Geoffrey Janssens, Doryan Temmerman

In this series of papers, we investigate properties of a finite group which are determined by its low degree irreducible representations over a number field $F$, i.e. its representations on matrix rings $\operatorname{M}_n(D)$ with $n \leq 2$. In particular we focus on representations on $\operatorname{M}_2(D)$ where $D$ is a division algebra having an order

Ahmed Saoudi

In this paper, we introduce and study the quadratic-phase Dunkl transform, a novel integral transform on the real line parameterized by five real numbers $(a, b, c, d, e)$ and a multiplicity parameter $\mu\geq -1/2$. We define the transform and establish its fundamental properties, including continuity, a Riemann--Lebesgue lemma, linearity, scaling, and most

Jianrong Zhang, Hehe Fan, Yi Yang

Human motions are compositional: complex behaviors can be described as combinations of simpler primitives. However, existing approaches primarily focus on forward modeling, e.g., learning holistic mappings from text to motion or composing a complex motion from a set of motion concepts. In this paper, we consider the inverse perspective: decomposing a holisti

Gabrielle Lalou, Husein Natur, Uzi Pereg

This paper studies the capacity limits for quantum secret sharing (QSS). The goal of a QSS scheme is to distribute a quantum secret among multiple participants, such that only authorized parties can recover it through collaboration, while no information can be obtained without such collaboration. We introduce an information-theoretic model for the rate analy

Zhibin Qin, Zhenxiong Tan, Zeqing Wang, Songhua Liu

Diffusion Transformer models have significantly advanced image editing by encoding conditional images and integrating them into transformer layers. However, most edits involve modifying only small regions, while current methods uniformly process and denoise all tokens at every timestep, causing redundant computation and potentially degrading unchanged areas.

Oscar Meneses Rojas

A classical result by Penrose establishes that null geodesics generating a black hole event horizon can only intersect at their entrance to the horizon in ``crossover'' points. This points together with limit points of this set, namely caustics, form the so-called "crease set". Light rays enter into the horizon through the crease set, characterizing the latt

Hanzhang Zhou, Xu Zhang, Panrong Tong, Jianan Zhang

The development of GUI agents could revolutionize the next generation of human-computer interaction. Motivated by this vision, we present MAI-UI, a family of foundation GUI agents spanning the full spectrum of sizes, including 2B, 8B, 32B, and 235B-A22B variants. We identify four key challenges to realistic deployment: the lack of native agent-user interacti

Shaofei Cai, Yulei Qin, Haojia Lin, Zihan Xu

Agentic reinforcement learning (RL) holds great promise for the development of autonomous agents under complex GUI tasks, but its scalability remains severely hampered by the verification of task completion. Existing task verification is treated as a passive, post-hoc process: a verifier (i.e., rule-based scoring script, reward or critic model, and LLM-as-a-

Zongmin Zhang, Zhen Sun, Yifan Liao, Wenhan Dong

Prompt-driven Video Segmentation Foundation Models (VSFMs), such as SAM2, are increasingly used in applications including autonomous driving and digital pathology, yet their security risks remain underexplored. We study backdoor attacks against VSFMs and show that directly applying classic attacks such as BadNet is largely ineffective, yielding attack succes

Suzanne van der Veldt, Gido M. van de Ven, Sanne Moorman, Guillaume Etter

Deep artificial neural networks famously struggle to learn from non-stationary streams of data. Without dedicated mitigation strategies, continual learning is associated with continuous forgetting of previous tasks and a progressive loss of plasticity. Current approaches to continual learning have either focused on increasing the stability of representations

Alexander Rybalov

Myasnikov, Ushakov, and Won introduced power circuits in 2012 to construct a polynomial-time algorithm for the word problem in the Baumslag group, which has a non-elementary Dehn function. Power circuits are computational structures that support addition and the operation $(x,y) \mapsto x \cdot 2^y$ on integers. They also posed the question of decidability o

Vishal Baibhav

Orbital eccentricity remains one of the least accessible parameters in observations of binary black hole (BBH) systems, largely erased by gravitational radiation long before detection. We introduce a new method to recover this lost parameter by using a more accessible and routinely measurable quantity: spin-orbit misalignment. In isolated binary evolution, a

Zhangbo Long, Letian Sha, Jiaye Pan, Haiping Huang

Binary program analysis represents a fundamental pillar of modern system security. Fine-grained methodologies like dynamic taint analysis still suffer from deployment complexity and performance overhead despite significant progress. Traditional in-process analysis tools trigger severe \textbf{address-space conflicts} that inevitably disrupt the native memory

Rodrigo Nicolau Almeida, Guram Bezhanishvili, Nick Bezhanishvili

We introduce Esakia order-compactifications and study how they fit in the general theory of Priestley order-compactifications. We provide an analog of Dwinger's theorem by characterizing Esakia order-compactifications by means of special rings of upsets. These considerations naturally lead to the notion of a locally Esakia space, for which we prove that taki

L. Inácio, A. Kurumbail, S. K. Panja, I. Brevik

We commence our study with review of dispersion interactions in electrolytes. We then reflect on how background media change atom-atom excited-state systems. To highlight the impact of nonlocal media, such as salt solutions, we predict that a new contribution to the resonance interaction energy emerges in a form $\propto e^{-\kappa_{\rm D} \rho}/\rho$. Here

Lizhe Wan, Jiaqi Yang

We investigate the low regularity local well-posedness of two-dimensional irrotational deep hydroelastic waves. Building on the approach of Ifrim-Tataru [29] and Ai-Ifrim-Tataru [5], in particular by constructing a cubic modified energy that incorporates a paradifferential weight chosen carefully, we prove that the hydroelastic waves are locally well-posed i

Vanessa D'Amario, Randy Daniel, Alessandro Zanetti, Dhruv Edamadaka

Small open-source medical large language models (LLMs) offer promising opportunities for low-resource deployment and broader accessibility. However, their evaluation is often limited to accuracy on medical multiple choice question (MCQ) benchmarks, and lacks evaluation of consistency, robustness, or reasoning behavior. We use MCQ coupled to human evaluation

Mayank Ratan Bhardwaj, Vishisht Srihari Rao, Bazil Ahmed, Kartik Sagar

This paper is motivated by the need to design a robust market mechanism to benefit farmers (producers of agricultural produce) as well as buyers of agricultural produce (consumers). Our proposal is a volume discount auction with a Farmer Collective (FC) as the selling agent and high volume or retail consumers as buying agents. An FC is a cooperative of farme

Wataru Nozawa

Large-scale competitive platforms are interacting multi-agent systems in which latent skills drift over time and pairwise interactions are shaped by matchmaking. We study a controlled rating dynamics in the mean-field limit and derive a kinetic description for the joint evolution of skills and ratings. In the Gaussian regime, we prove an exact moment closure

Meng Wang, Zhichao Wang

In this paper, we establish the almost everywhere convergence of solutions to the Schr\"odinger operator with complex time $ P_{\gamma}f(x,t) $ in higher dimensions, under the assumption that the initial data $f$ belongs to the Sobolev space $ H^{s}(\mathbb{R}^d)$.

Zhuoran Zhu, Chunyang Zhu, Hao Lin, Xu Fu

Large-scale Mixture-of-Experts (MoE) models rely on \emph{expert parallelism} for efficient training and inference, which splits experts across devices and necessitates distributed data shuffling to route each token to its assigned experts. However, existing communication libraries handle this shuffling poorly; its overhead can account for over half of end-t

Yanmeng Wang, Zhiwen Dai, Shuai Wang, Jian Zhou

Federated Fine-Tuning (FFT) has attracted growing interest as it leverages both server- and client-side data to enhance global model generalization while preserving privacy, and significantly reduces the computational burden on edge devices by avoiding training from scratch. Despite these advantages, FFT performance is often degraded by unreliable server-cli

Hiroshi Nozaki, Yuta Watanabe

We study $T$-designs in the nonbinary Johnson scheme. This scheme generalizes both the Johnson and Hamming schemes and admits a bivariate $Q$-polynomial structure. Zhu (2021) provided a combinatorial characterization of $T$-designs in this scheme for certain index sets $T$, using a relationship between $T$-designs in the nonbinary Johnson scheme and relative

Jihong Liu, Hao Qi, Zhangwei Shan

Identifying codes were introduced by Karpovsky et al. as dominating sets $S\subseteq V(G)$ satisfying $N[u]\cap S \neq N[v]\cap S$ for any distinct vertices $u,v$. Later, Junnila et al. introduced the concept of \emph{self-identifying codes} (previously called $(1,\leq1)^+$-identifying codes in earlier work), a dominating set $S\subseteq V(G)$ such that $\bi

Ziyan Zhang, Nan Gao, Zhiqiang Nie, Shantanu Pal

With the rapid advancement of large language models (LLMs), intelligent conversational assistants have demonstrated remarkable capabilities across various domains. However, they still mainly rely on explicit textual input and do not know the real world behaviors of users. This paper proposes a context-sensitive conversational assistant framework grounded in

Nagham Osman, Vittorio Lembo, Giovanni Bottegoni, Laura Toni

Hit identification is a critical yet resource-intensive step in the drug discovery pipeline, traditionally relying on high-throughput screening of large compound libraries. Despite advancements in virtual screening, these methods remain time-consuming and costly. Recent progress in deep learning has enabled the development of generative models capable of lea

Somayyeh Alidoust, V. Ongun Özçelik

Perovskite solar cells (PSCs) based on methylammonium lead iodide (MAPbI3) exhibit remarkable photovoltaic performance, where interface engineering with hole transport layers (HTLs) is crucial for optimizing charge transfer and device efficiency. In this work, we present a density functional theory (DFT) study of the MAPbI3/poly(3-hexylthiophene) (P3HT) hybr

Yu-Xuan Zhang, Jing-Ling Chen

Quantum nonlocality is an essential resource in quantum information and is commonly classified into three distinct forms: quantum entanglement, Einstein-Podolsky-Rosen (EPR) steering, and Bell&#39;s nonlocality. Gisin&#39;s theorem shows that pure-state entanglement implies Bell nonlocality, and it motivates the question of how entanglement and EPR steering

Wenbin Li, Shangge Liu, Borui Kang, Yiyang Chen

A fundamental challenge in Continual Learning (CL) is catastrophic forgetting, where adapting to new tasks degrades the performance on previous ones. While the field has evolved with diverse methods, this rapid surge in diverse methodologies has culminated in a fragmented research landscape. The lack of a unified framework, including inconsistent implementat

Hanmo You, Zan Wang, Zishuo Dong, Luanqi Mo

Deep Learning (DL) has been widely adopted in diverse industrial domains, including autonomous driving, intelligent healthcare, and aided programming. Like traditional software, DL systems are also prone to faults, whose malfunctioning may expose users to significant risks. Consequently, numerous approaches have been proposed to address these issues. In this

O. Vince, C. M. Raiteri, M. Villata, A. C. Gupta

(Shortened)Context: We analyze the optical variability of the FSRQ Ton 599 using BVRI photometry from the WEBT collaboration (2011-2023), complemented by photometric and spectroscopic data from the Steward Observatory.\\ Aims: To characterize short- and long-term optical variability -- including flux distributions, intranight changes, color evolution, and sp

Huanhuan Yuan, Yang Ping, Zhengqin Xu, Junyi Cao

The rapid advancement of generative artificial intelligence has enabled the creation of highly realistic fake facial images, posing serious threats to personal privacy and the integrity of online information. Existing deepfake detection methods often rely on handcrafted forensic cues and complex architectures, achieving strong performance in intra-domain set

S. Vrbaški, G. Stanić, S. Molinelli, M. Bhattarai

In this work, we proposed virtual imaging simulators as an alternative approach to experimental validation of beam range uncertainty in complex patient geometry using a computational model of a human head and a photon-counting CT scanner. We validate the accuracy of stopping power ratio (SPR) calculations using a conventional stoichiometric calibration appro

M. A. Korolev

We give an elementary proof of some identities that express the squares of Riemann zeta function at integer points in terms of the series involving hyperbolic functions, digamma function, Bernoulli numbers etc. In this version, inaccuracies in the text have been corrected and one of the bibliographic references has been updated.

Wesley S. Leite, Rodrigo C. de Lamare, Yuriy Zakharov, Wei Liu

In this work, we introduce a variable window size (VWS) spatial smoothing framework that enhances coarray-based direction of arrival (DOA) estimation for sparse linear arrays. By compressing the smoothing aperture, the proposed VWS Coarray MUSIC (VWS-CA-MUSIC) and VWS Coarray root-MUSIC (VWS-CA-rMUSIC) algorithms replace part of the perturbed rank-one outer

Michail Kalntis, George Iosifidis, José Suárez-Varela, Andra Lutu

While traditional handovers (THOs) have served as a backbone for mobile connectivity, they increasingly suffer from failures and delays, especially in dense deployments and high-frequency bands. To address these limitations, 3GPP introduced Conditional Handovers (CHOs) that enable proactive cell reservations and user-driven execution. However, both handover

Ryo Watanabe, Toshiya Hikihara, Hiroshi Ueda

In the approaches based on matrix-product states (MPSs), such as the density-matrix renormalization group (DMRG) method, the ordering of the sites crucially affects the computational accuracy. We investigate the performance of an algorithm that searches for the optimal site order by iterative local site rearrangement. We improve the algorithm by expanding th