November 2025 arXiv papers — page 147

Zhenxiao Fu, Chen Fan, Lei Jiang

LLMs have transformed NLP, yet deploying them on edge devices poses great carbon challenges. Prior estimators remain incomplete, neglecting peripheral energy use, distinct prefill/decode behaviors, and SoC design complexity. This paper presents CO2-Meter, a unified framework for estimating operational and embodied carbon in LLM edge inference. Contributions

Caio B. de S. Nascimento, Marilena Loverde

We explore the cosmological signatures of neutrino and Hot Dark Matter (HDM) wakes, which refers to the preferential accumulation of neutrinos (or, more broadly, HDM particles) downstream of moving cold dark matter structures. We improve on existing theoretical models, and provide forecasts for the detectability of the effect in future surveys under more rea

Shanaka Liyanaarachchi, Chathurya Wijethunga, Shihab Aaqil Ahamed, Akthas Absar

Spatial transcriptomics is an emerging field that enables the identification of functional regions based on the spatial distribution of gene expression. Integrating this functional information present in transcriptomic data with structural data from histopathology images is an active research area with applications in identifying tumor substructures associat

Mohammad Taghvaei, Ehsan Amani

For high-fidelity predictions of turbulent flows in complex practical engineering problems, the Wall-Modeled (WM) Large-Eddy Simulation (LES) has aroused great interest. In the present study, we prove that the conventional Wall-Stress Models (WSMs) developed for WMLES of Newtonian fluids fail to predict the shear-thinning-induced drag reduction in power-law

Mainak Singha, Jose Aguilera-Toste, Vinayak Lahiri

We test whether simple, interpretable state variables-trend and momentum-can generate durable out-of-sample alpha in one of the world's most liquid assets, gold. Using a rolling 10-year training and 6-month testing walk-forward from 2015 to 2025 (2,793 trading days), we convert a smoothed trend-momentum regime signal into volatility-targeted, friction-aware

Jamison Moody, James Usevitch

Kolmogorov-Arnold Networks (KANs) are a class of neural networks that have received increased attention in recent literature. In contrast to MLPs, KANs leverage parameterized, trainable activation functions and offer several benefits including improved interpretability and higher accuracy on learning symbolic equations. However, the original KAN architecture

H. Eshraghi, E. Amani, M. Saffar-Avval

In the present article, novel Coarse-Graining (CG) algorithms for the Eulerian-Lagrangian (EL) simulation of particle-laden flows are proposed. These include different variants of Reproducing Kernel Particle Methods (RKPM) and an extended Diffusion Two-Step Method (DTSM) for highly polydisperse flows. Owing to the dynamic nature of the kernel function in RKP

Jie Ren, Bin Ma, Shuangyan Yang, Benjamin Francis

Deep learning recommendation models (DLRMs) are widely used in industry, and their memory capacity requirements reach the terabyte scale. Tiered memory architectures provide a cost-effective solution but introduce challenges in embedding-vector placement due to complex embedding-access patterns. We propose RecMG, a machine learning (ML)-guided system for vec

Hanqing Zhu, Zhenyu Zhang, Hanxian Huang, DiJia Su

Reinforcement Learning with Verifiable Rewards (RLVR) reliably improves the reasoning performance of large language models, yet it appears to modify only a small fraction of parameters. We revisit this paradox and show that sparsity is a surface artifact of a model-conditioned optimization bias: for a fixed pretrained model, updates consistently localize to

Bobby Samir Acharya, Ethan Torres

In $SU(N)$ Yang-Mills theory without matter, there exist stable long electric fluxtube strings which carry a 1-form symmetry charge. Over the past decade or so, there has been increasing evidence from lattice calculations that the worldsheet theories of such QCD strings contain a massive pseudoscalar (axion), at least when $N\geq 3$. This has so far been puz

Shao Yuan Lin, Laura Pierson

We study the $H$-chromatic symmetric functions $X_G^H$ (introduced in (arXiv:2011.06063) as a generalization of the chromatic symmetric function (CSF) $X_G$), which track homomorphisms from the graph $G$ to the graph $H$. We focus first on the case of self-chromatic symmetric functions (self-CSFs) $X_G^G$, making some progress toward a conjecture from (arXiv

Joseph M. Lukens, Karthik V. Myilswamy, Alexander Miloshevsky, Hsuan-Hao Lu

We present a self-contained analytical model for biphoton generation in microring resonators. Encompassing both all-pass and add-drop geometries, identical and distinct pump and biphoton coupling coefficients, and continuous-wave and pulsed pumping, our interaction-picture-based approach reveals time-frequency biphoton correlations while also predicting abso

Shivank, Anurag Singh, Fakhteh Ghanbarnejad, Ajay K Sharma

Climate change is intensifying infectious and chronic diseases like malaria and diabetes, respectively, especially among the vulnerable populations. Global temperatures have risen by approximately $0.6^\circ$C since 1950, extending the window of transmission for mosquito-borne infections and worsening outcomes in diabetes due to metabolic stress caused by he

Cedrick Kinavuidi, Luca Peres, Oliver Rhodes

This work presents a novel spiking neural network (SNN) decoding method, combining SNNs with Hyperdimensional computing (HDC). The goal is to create a decoding method with high accuracy, high noise robustness, low latency and low energy usage. Compared to analogous architectures decoded with existing approaches, the presented SNN-HDC model attains generally

Fernanda Alves Caixeta, Keti Tenenblat

We study Laguerre isotropic hypersurfaces in the Euclidean space, which are hypersurfaces whose Laguerre form is zero and the eigenvalues of the Laguerre tensor are constant and equal to $\lambda\geq 0$. We prove a rigidity theorem for the L-isotropic hypersurfaces parametrized by lines of curvature. Moreover, we study the hypersurfaces that are L-isotropic

Shaleen Baral, Robert Kleinberg, Sylvan Martin, Henry Rogers

Reconfigurable networks are a novel communication paradigm in which the pattern of connectivity between hosts varies rapidly over time. Prior theoretical work explored the inherent tradeoffs between throughput (or, hop-count) and latency, and showed the existence of infinitely many Pareto-optimal designs as the network size tends to infinity. Existing Pareto

Yue Fang, Jin Zhi, Jie An, Hongshen Chen

Signal Temporal Logic (STL) is a powerful formal language for specifying real-time specifications of Cyber-Physical Systems (CPS). Transforming specifications written in natural language into STL formulas automatically has attracted increasing attention. Existing rule-based methods depend heavily on rigid pattern matching and domain-specific knowledge, limit

Sara Maria Brancato, Davide Salzano, Davide Fiore, Francesco De Lellis

Microbial consortia offer significant biotechnological advantages over monocultures for bioproduction. However, industrial deployment is hampered by the lack of scalable architectures to ensure stable coexistence between populations. Existing strategies rely on genetic modifications, which impose metabolic load, or environmental changes, which can reduce pro

Mohamed Siala, Abdullah Al-Nafisah, Tareq Al-Naffouri

Nyquist Signaling Modulations (NSMs) are a new signaling paradigm inspired by faster-than-Nyquist principles but based on a distinct approach that enables controlled inter-symbol interference through carefully designed finite-impulse-response filters. NSMs can operate in any number of dimensions, including mixed-dimensional configurations, offering wide flex

Ivan Butakov, Alexander Semenenko, Valeriya Kirova, Alexey Frolov

We introduce a novel Mutual Information (MI) estimator that fundamentally reframes the discriminative approach. Instead of training a classifier to discriminate between joint and marginal distributions, we learn a normalizing flow that transforms one into the other. This technique produces a computationally efficient and precise MI estimate that scales well

Bernhard Haeupler, Yonggang Jiang, Thatchaphol Saranurak

We present the first deterministic nearly-linear time algorithm for single-source shortest paths with negative edge weights on directed graphs: given a directed graph $G$ with $n$ vertices, $m$ edges whose weights are integer in $\{-W,\dots,W\}$, our algorithm either computes all distances from a source $s$ or reports a negative cycle in time $\tilde{O}(m)\c

Rachael Boyd, Guy Boyde, Oscar Randal-Williams, Robin J. Sroka

We show that the homology of a Temperley-Lieb algebra on an even number of strands has a rich algebraic structure and is highly nontrivial in general. This is achieved by proving that it is entirely governed by a differential graded algebra: the differential graded algebra of planar loops. We provide a small model for this dga, and use it to obtain consequen

Parshwa Shah, Dhaval K. Patel, Brijesh Soni, Miguel López-Benítez

Recently, Deep Learning (DL) techniques have been used for User Equipment (UE) positioning. However, the key shortcomings of such models is that: i) they weigh the same attention to the entire input; ii) they are not well suited for the non-sequential data e.g., when only instantaneous Channel State Information (CSI) is available. In this context, we propose

Simone Giombi, Anurag Pendse

We study line defects with a cusp in fermionic CFTs arising as fixed points of scalar-fermion theories with Yukawa interactions. These include the Gross-Neveu-Yukawa model and some of its generalizations with additional scalar fields, which can be thought of as UV completions of fermionic theories with quartic interactions. We compute the cusp anomalous dime

James Ingoldby, Valentin V. Khoze, Jessica Turner

Metastable cosmic strings provide a minimal and predictive origin for the stochastic gravitational-wave background reported by Pulsar Timing Array experiments. We analyse this possibility in electroweak-like dark sectors with a single-stage breaking $SU(2)\times U(1)\!\to\!U(1)$ driven by one Higgs field. In the regime with dark sector Higgs mass below the $

Sriram Srinivasan, Gautam Ramachandra

Accurate 3D reconstruction from multi-view images is essential for downstream robotic tasks such as navigation, manipulation, and environment understanding. However, obtaining precise camera poses in real-world settings remains challenging, even when calibration parameters are known. This limits the practicality of existing NeRF-based methods that rely heavi

Randall Balestriero, Yann LeCun

Learning manipulable representations of the world and its dynamics is central to AI. Joint-Embedding Predictive Architectures (JEPAs) offer a promising blueprint, but lack of practical guidance and theory has led to ad-hoc R&D. We present a comprehensive theory of JEPAs and instantiate it in {\bf LeJEPA}, a lean, scalable, and theoretically grounded training

Soumik Ghosh, Arjun Mirani, Yihui Quek, Michelle Xu

We study a counterintuitive property of 'conditioning' on the result of measuring a subsystem of a quantum state: such conditioning can boost design quality, at the cost of increased system size. We work in the setting of deep thermalization from many-body physics: starting from a bipartite state on a global system $(A,B)$ drawn from a $k$-design, we measure

Tren Baltussen, Maurice Heemels, Alexander Katriniok

This manuscript presents a dual model predictive controller (MPC) that balances the two objectives of dual control, namely, system identification and control. In particular, we propose a Gaussian process (GP)-based MPC that uses the posterior GP covariance for active learning. The dual MPC can steer the system towards states with high covariance, or to the s

David Barker, Sebastian Lehmann, Kimberly A. Dick, Peter Samuelsson

In Szilard's engine, measurement and feedback allows to extract work from an equilibrium environment, a process otherwise forbidden by the laws of thermodynamics. Recent theoretical developments have established fluctuation theorems and thermodynamic uncertainty relations that constrain the fluctuations in Szilard's engine. These relations rely on auxiliary

Xiaolei Tong, Pedram Yousefian, Ziyi Wang, Meenakshi A. Saravanan

Ferroelectric aluminum scandium nitride (Al1-xScxN, AlScN) offers CMOS-compatible integration but suffers from high coercive fields and leakage currents that hinder thickness scaling. Further reduction in thickness is essential for low-voltage embedded nonvolatile memory applications. Boron incorporation into AlScN (AlBScN) suppresses leakage current in film

Dogyoon Song

We study average treatment effect (ATE) estimation under complete randomization with many covariates in a design-based, finite-population framework. In randomized experiments, regression adjustment can improve precision of estimators using covariates, without requiring a correctly specified outcome model. However, existing design-based analyses establish asy

Yannan Bai, Kamesh Munagala, Yiheng Shen, Davidson Zhu

We study the problem of selection in the context of Bayesian persuasion. We are given multiple agents with hidden values (or quality scores), to whom resources must be allocated by a welfare-maximizing decision-maker. An intermediary with knowledge of the agents' values seeks to influence the outcome of the selection by designing informative signals and prov

Amirmohammad Omidi Galdiani, Sepehr Rezaei Melal, Mohammad Norasteh, Arash Yousefi Jordehi

This report presents a detailed methodology for constructing a high-quality Semantic Role Labeling (SRL) dataset from the Wall Street Journal (WSJ) portion of the OntoNotes 5.0 corpus and adapting it for Opinion Role Labeling (ORL) tasks. Leveraging the PropBank annotation framework, we implement a reproducible extraction pipeline that aligns predicate-argum

Yunhong He, Zhengqing Yuan, Zhengzhong Tu, Yanfang Ye

We introduce 3D4D, an interactive 4D visualization framework that integrates WebGL with Supersplat rendering. It transforms static images and text into coherent 4D scenes through four core modules and employs a foveated rendering strategy for efficient, real-time multi-modal interaction. This framework enables adaptive, user-driven exploration of complex 4D

Sen Zhang, Xiaoxiao He, Di Liu, Zhaoyang Xia

We present Large Sign Language Models (LSLM), a novel framework for translating 3D American Sign Language (ASL) by leveraging Large Language Models (LLMs) as the backbone, which can benefit hearing-impaired individuals' virtual communication. Unlike existing sign language recognition methods that rely on 2D video, our approach directly utilizes 3D sign langu

Bansari. J. Rayjada, Jekil. A. Gadhiya, Mahadityasinh. A. Sarvaiya

In graph theory, a Snark is a connected, bridgeless, Cubic graph that cannot be edge-colored with only three colors. Additionally, to avoid some trivial cases, a Snark is typically required to have a girth of minimum five and a cyclic connectivity of minimum four. In this paper, we investigate the Cordial labeling, for one of the modified structures of Snark

Panagiotis Gavriilidis, Kyriakos Stylianopoulos, George C. Alexandropoulos

This paper presents an asymptotic analysis of Multiple-Input Multiple-Output (MIMO) systems assisted by a 1-bit Reconfigurable Intelligent Surface (RIS) under Ricean fading conditions. Using random matrix theory, we show that, in the asymptotic regime, the dominant singular values and vectors of the transmitter-RIS and RIS-receiver channels converge to their

Sergey Matveenko, Nikita Rastegaev

In this paper we discuss a one-dimensional model for two-phase Enhanced Oil Recovery (EOR) floods, primarily for the polymer flood. We improve upon the method for the construction of semi-analytical solutions for the oil displacement by a water slug containing dissolved chemicals given in (Pires, Bedrikovetsky and Shapiro, 2006) and later generalized in (Apo

Yinjun Zhao, Nicholas Tatonetti, Yuanjia Wang

Electronic health records (EHRs) linked with familial relationship data offer a unique opportunity to investigate the genetic architecture of complex phenotypes at scale. However, existing heritability and coheritability estimation methods often fail to account for the intricacies of familial correlation structures, heterogeneity across phenotype types, and

Ricard Solé, Jordi Pla-Mauri

The slime mould Physarum polycephalum displays adaptive transport dynamics and network formation that have inspired its use as a model of biological computation. We develop a Lagrangian formulation of Physarum's adaptive dynamics on predefined graphs, showing that steady states arise as extrema of a least-action functional balancing metabolic dissipation and

Rong Feng, Vanisha Gupta, Vivek Patel, Viroopaksh Reddy Ernampati

Symbolic execution helps check programs by exploring different paths based on symbolic inputs. Tools like KLEE are commonly used because they can automatically detect bugs and create test cases. But one of KLEE's biggest issues is how slow it can get when programs have lots of branching paths-it often becomes too resource-heavy to run on large or complex cod

Andrew Li, Hua Wang

We consider colored compositions where only some parts are allowed different colors, depending on their locations in the composition. The counting sequences are obtained through generating functions. Connections to many other combinatorial objects are discussed, with combinatorial arguments provided and generalized for these observations.

Jack Kelly, Lenny Neyt, Sven-Ake Wegner

The notion of an LB-space was introduced by Grothendieck in his 1953 th\`{e}se, referring to a countable colimit of Banach spaces taken within the category of locally convex topological vector spaces, and refining prior work done by Dieudonn\'{e}, Schwartz and K\"othe. Recently, two different notions of `bornological LB-spaces' emerged: one, given by Stempfh

Ernesto F. Eiroa, Griselda Figueroa-Aguirre, Vasiliki Karanasou

In this article, we construct a family of spherically symmetric thin-shell wormholes within scalar-tensor theories of gravity. In the case of wormholes symmetric across the throat, we study the matter content and analyze the stability of the static configurations under radial perturbations. We apply the formalism to a particular example involving Einstein-Ma

Chenyu Fang, Phillip Gutierrez, Chung Kao

We investigate the discovery potential of flavor changing neutral Higgs (FCNH) interactions in top quark decays at the LHC. A general two Higgs doublet model (2HDM) is adopted to study the top quark decay $t \to c \phi^0$, where $\phi^0$ is either a CP-even heavy neutral scalar ($H^0$), or a CP-odd pseudoscalar ($A^0$), followed by the Higgs boson decaying i

Shu Yang, Junchao Wu, Xilin Gong, Xuansheng Wu

Large Reasoning Models (LRMs) have demonstrated remarkable performance on complex tasks by engaging in extended reasoning before producing final answers. Beyond improving abilities, these detailed reasoning traces also create a new opportunity for AI safety, CoT Monitorability: monitoring potential model misbehavior, such as the use of shortcuts or sycophanc

Mathieu Mourichoux

We give an explicit construction of the Brownian sphere biased by the distance between two distinguished points, which is based on the Miermont bijection for quadrangulations. We then describe various conditionings of this object, which are related to Vorono\"i cells in the Brownian sphere. In particular, we give a new construction of the Brownian sphere wit

Zhaojian Yu, Kaiyue Feng, Yilun Zhao, Shilin He

LLMs have made significant progress in complex but easy-to-verify problems, yet they still struggle with discovering the unknown. In this paper, we present \textbf{AlphaResearch}, an autonomous research agent designed to discover new algorithms on open-ended problems by iteratively running the following steps: (1) propose new ideas (2) program to verify (3)

Zhengyang Liang, Daoan Zhang, Huichi Zhou, Rui Huang

While specialized AI models excel at isolated video tasks like generation or understanding, real-world applications demand complex, iterative workflows that combine these capabilities. To bridge this gap, we introduce UniVA, an open-source, omni-capable multi-agent framework for next-generation video generalists that unifies video understanding, segmentation

Alexander J. H. Houston

We analyse the effect of anisotropy in elastic constants on the hydrodynamics of active nematics. Building on the multipole framework for a single elastic constant, we determine the leading effect of elastic anisotropy on the active response of generic distortions. The key findings are a new active torque, proportional to the anisotropy, in response to monop

Miranda Pikus, Paul Duffell, Soham Mandal, Abigail Polin

We systematically study how cooling creates structural changes in supernova remnants as they evolve. Inspired by the peculiar morphology of the Pa 30 nebula, we adopt a framework in which to characterize supernova remnants under different degrees of cooling. Our cooling framework characterizes remnants with a singular parameter called $\beta$ that sets how r

José Burillo, Marc Felipe

In this short note, a bound on the word metric for Thompson's group V given by Birget in 2004 is improved to a new bound, which agrees with the known bounds for Thompson's groups F and T.

Florin Teleanu, Anne M. Fabricant, Chengtong Zhang, Gary P. Centers

Nuclear-magnetic-resonance experiments can interrogate a broad spectrum of molecular-tumbling regimes and can accurately measure interatomic distances in solution with sub-nanometer resolution. In the zero- to ultralow-field (ZULF) regime, population and coherence decay reveal nontrivial behavior due to strong coupling between nuclear spins. We note, in part

Devesh Nandal, Sunmyon Chon

Supermassive stars (SMSs) are candidate progenitors of massive black hole seeds and may contribute to anomalous abundance patterns in high-redshift galaxies and globular clusters. Recent radiation-hydrodynamic simulations indicate that SMSs can form at finite metallicity, not only in metal-free direct-collapse conditions. We model SMS growth with \textsc{GEN

Manuel Bodirsky, Santiago Guzmán Pro

Extensional ESO is a fragment of existential second-order logic (ESO) that captures the following family of problems. Given a fixed ESO sentence $\Psi$ and an input structure $\mathbb A$ the task if to decide whether there is an extension $\mathbb B$ of $\mathbb A$ that satisfies the first-order part of $\Psi$, i.e., a structure $\mathbb B$ such that $R^{\ma

Caner Ünal, Doğa Veske

A universal contribution exists in the infrared (low frequency) regime of all gravitational waves, which results from nonlinear memory. Nonlinear memory is sourced by linear order gravitational waves and exists for any gravitational-wave background. We calculate the stochastic nonlinear memory signal of various stochastic backgrounds of cosmological (scalar

Ali Sonmez, Erencem Ozbey, Efe Feyzi Mantaroglu, H. Birkan Yilmaz

Transmitter localization in Molecular Communication via Diffusion is a critical topic with many applications. However, accurate localization of multiple transmitters is a challenging problem due to the stochastic nature of diffusion and overlapping molecule distributions at the receiver surface. To address these issues, we introduce clustering-based centroid

Leonie Bossemeyer, Samuel Heinrich, Grant Van Horn, Oisin Mac Aodha

Mastering fine-grained visual recognition, essential in many expert domains, can require that specialists undergo years of dedicated training. Modeling the progression of such expertize in humans remains challenging, and accurately inferring a human learner's knowledge state is a key step toward understanding visual learning. We introduce CleverBirds, a larg

Faisal Ahmed, Taymaz Akan, Fatih Gelir, Owen T. Carmichael

Now that disease-modifying therapies for Alzheimer disease have been approved by regulatory agencies, the early, objective, and accurate clinical diagnosis of AD based on the lowest-cost measurement modalities possible has become an increasingly urgent need. In this study, we propose a novel feature extraction method using persistent homology to analyze stru

Zhipeng Xu, Shengshan Qin, Kun Jiang, Jiangping Hu

Sublattice dichotomy in monolayer FeSe/SrTiO$_3$, signaling the breaking of symmetries exchanging the two Fe sublattices, has recently been reported. We propose that interband pairing serves as the origin of this dichotomy, regardless of whether the symmetry is broken in the normal state or in the pairing state. If symmetry breaking occurs in the normal stat

Anupama Sitaraman, Bharathan Balaji, Yuvraj Agarwal

Investigating the effects of climate change and global warming caused by GHG emissions have been a key concern worldwide. These emissions are largely contributed to by the production, use and disposal of consumer products. Thus, it is important to build tools to estimate the environmental impact of consumer goods, an essential part of which is conducting Lif

Xueqi Guo, Halid Ziya Yerebakan, Yoshihisa Shinagawa, Kritika Iyer

Multi-organ segmentation of 3D medical images is fundamental with meaningful applications in various clinical automation pipelines. Although deep learning has achieved superior performance, the time and memory consumption of segmenting the entire 3D volume voxel by voxel using neural networks can be huge. Classifiers have been developed as an alternative in

William J. Huggins, Tanuj Khattar, Amanda Xu, Matthew Harrigan

Holistic resource estimates are essential for guiding the development of fault-tolerant quantum algorithms and the computers they will run on. This is particularly true when we focus on highly-constrained early fault-tolerant devices. Many attempts to optimize algorithms for early fault-tolerance focus on simple metrics, such as the circuit depth or T-count.

Fedor Pakovich

Let $ R$ be a compact Riemann surface, and let $ P: R \to \mathbb P^1(\mathbb C) $ and $ Q: R \to \mathbb P^1(\mathbb C) $ be holomorphic maps. In this paper, we investigate the following problem: under what conditions do the preimages $ P^{-1}(K) $ and $ Q^{-1}(K) $ coincide for some infinite set $K$ contained in $\mathbb P^1(k)$, where $k$ is a finitely ge

Omri Koshorek, Niv Granot, Aviv Alloni, Shahar Admati

Retrieval-Augmented Generation (RAG) has become the dominant approach for answering questions over large corpora. However, current datasets and methods are highly focused on cases where only a small part of the corpus (usually a few paragraphs) is relevant per query, and fail to capture the rich world of aggregative queries. These require gathering informati

Kailong Wang, Athina Petropulu

Most works on pilot-aided orthogonal time frequency space (OTFS) channel estimation operate in the delay-Doppler (DD) domain, where embedded-pilot guard regions must cover the unknown maximum delay-Doppler spread, causing high overhead that worsens with antenna scaling: per-antenna guard regions increase overhead with the number of transmit antennas, while s

Slava G. Turyshev

Lunar dust -- the sub-millimeter fraction of the regolith -- controls the optical, thermophysical, electrical, mechanical, and environmental behavior of the Moon's surface. These properties set the performance envelopes of remote-sensing retrievals, regolith geotechnics, volatile cycles, and exploration systems, while also posing operational and biomedical r

Ruya Karagulle, Cristian-Ioan Vasile, Necmiye Ozay

Autonomous systems increasingly rely on human feedback to align their behavior, expressed as pairwise comparisons, rankings, or demonstrations. While existing methods can adapt behaviors, they often fail to guarantee safety in safety-critical domains. We propose a safety-guaranteed, optimal, and efficient approach for solving the learning problem from prefer

Oliver Janzer, Julien Portier

An $r$-cut of a $k$-uniform hypergraph is a partition of its vertex set into $r$ parts, and the size of the cut is the number of edges which have at least one vertex in each part. The study of the possible size of the largest $r$-cut in a $k$-uniform hypergraph was initiated by Erd\H{o}s and Kleitman in 1968. For graphs, a celebrated result of Edwards states

Berkcan Kapusuzoglu, Supriyo Chakraborty, Renkun Ni, Stephen Rawls

Large language models (LLMs) adapted to financial domains often suffer from catastrophic forgetting of general reasoning capabilities essential for customer interactions and complex financial analysis. We introduce Selective Parameter Evaluation and Restoration via Model Merging (SPEAR-MM), a practical framework that preserves critical capabilities while ena

Christopher M. J. Osborne

We present a novel ray acceleration structure for radiative transfer outside of local thermodynamic equilibrium (non-LTE), leveraging techniques from computer graphics to improve computational efficiency. By applying mipmapping (local recursive spatial averaging) and sparse voxel grids, we exploit spatial coherence and sparsity in astrophysical models to acc

Titir Mukherjee, Arnab Acharya, Soumitro Banerjee, Deb Shankar Ray

Dynamical chaos in a periodically driven, dissipative soft impact oscillator is investigated in the quantum regime using the complex-number quantum Langevin equation (c-number QLE). The averaged system dynamics are analyzed through a comprehensive suite of time-series diagnostics, including bifurcation diagrams, Lyapunov exponents, Fourier spectra, and the 0

Bingsong Bai, Yizhong Geng, Fengping Wang, Cong Wang

Zero-shot singing voice conversion (SVC) transforms a source singer's timbre to an unseen target speaker's voice while preserving melodic content without fine-tuning. Existing methods model speaker timbre and vocal content separately, losing essential acoustic information that degrades output quality while requiring significant computational resources. To ov

Hongjian Zhou, Xining Wang, Feng Zhu, Liming Zhang

Inspired by the $X(4140)$ structure reported in the $J/\psi \phi$ system by the CDF experiment in 2009, a series of searches have been carried out in the $J/\psi \phi$ and $J/\psi K$ channels, leading to the claim of ten structures in the $B \rightarrow J/\psi \phi K$ system. This article provides a comprehensive review of experimental developments, from the

Hongyu Guo

Tarski's first-order axiom system $\mathscr{E}_{2}$ for Euclidean geometry is notable for its completeness and decidability. However, the Pythagorean theorem -- either in its modern algebraic form $a^{2}+b^{2}=c^{2}$ or in Euclid's Elements -- cannot be directly expressed in $\mathscr{E}_{2}$, since neither distance nor area is a primitive notion in the lang

Yaozu Tang, Mazhar N. Ali, Gerrit E. W. Bauer, Yaroslav M. Blanter

Josephson junctions are essential devices in superconducting electronics and quantum computing hardware. Here we predict electrical control of the supercurrent in composite superconductor-insulator-ferroelectric-insulator-superconductor (S-I-FE-I-S) Josephson junctions. Inversion symmetry broken by unequal dielectric barrier thicknesses and/or potentials con

Li Yang, Abdallah Shami

With increasingly sophisticated cybersecurity threats and rising demand for network automation, autonomous cybersecurity mechanisms are becoming critical for securing modern networks. The rapid expansion of Internet of Things (IoT) systems amplifies these challenges, as resource-constrained IoT devices demand scalable and efficient security solutions. In thi

Mariana Smith, Tanner Watts, Susheela Sharma Stern, Brendan Burkhart

Concentric tube robots (CTRs) offer dexterous motion at millimeter scales, enabling minimally invasive procedures through natural orifices. This work presents a coordinated model-based resection planner and learning-based retraction network that work together to enable semi-autonomous tissue resection using a dual-arm transurethral concentric tube robot (the

Dmitry A. Timashev

Let $X=G/H$ be a spherical homogeneous variety for a complex reductive algebraic group $G$. We prove that the orbit space of $X$ under the action of a maximal compact subgroup $K\subset G$ is homeomorphic to the valuation cone of $X$. We also discuss the relation between the orbit type stratification of the orbit space and the face stratification of the valu

Zihan Ma, Dongsheng Zhu, Shudong Liu, Taolin Zhang

Current safety evaluations for LLM-driven agents primarily focus on atomic harms, failing to address sophisticated threats where malicious intent is concealed or diluted within complex tasks. We address this gap with a two-dimensional analysis of agent safety brittleness under the orthogonal pressures of intent concealment and task complexity. To enable this

Sayan Bhattacharya, Ruoxu Cen, Debmalya Panigrahi

In (fully) dynamic set cover, the goal is to maintain an approximately optimal solution to a dynamically evolving instance of set cover, where in each step either an element is added to or removed from the instance. The two main desiderata of a dynamic set cover algorithm are to minimize at each time-step, the recourse, which is the number of sets removed fr

Huzaifa Arif, Keerthiram Murugesan, Ching-Yun Ko, Pin-Yu Chen

We propose patching for large language models (LLMs) like software versions, a lightweight and modular approach for addressing safety vulnerabilities. While vendors release improved LLM versions, major releases are costly, infrequent, and difficult to tailor to customer needs, leaving released models with known safety gaps. Unlike full-model fine-tuning or m

Giuseppe Tomassini

The topics of Convexity and Concavity and Envelopes are central in Complex Analysis and extensively investigated. The aim of this paper is to find a possible counterpart in Algebraic Geometry. The article presents preliminary results on this topic,

David Jaklitsch, Makoto Yamashita

We consider the tube algebra of a spherical semisimple multitensor category $\mathcal{X}$, and construct a braided monoidal structure with twist for its representations. We further show that this category is braided tensor equivalent with the Drinfeld center of the ind-category of $\mathcal{X}$, extending the well-known linear equivalence.

Suyog Garg, Bidya Binay Karak, Rohan B. Mandrai

Similar to the solar cycle, the magnetic cycles of other solar-type stars are also variable. How the variability of the stellar cycle changes with the rotation rate or the dynamo number is a valuable information for understanding the stellar dynamo process. We examine the variability in the stellar magnetic cycles by studying 81 stars from the data of the Mo

Da Li, Yuxiao Luo, Keping Bi, Jiafeng Guo

Multimodal Large Language Models advance multimodal representation learning by acquiring transferable semantic embeddings, thereby substantially enhancing performance across a range of vision-language tasks, including cross-modal retrieval, clustering, and classification. An effective embedding is expected to comprehensively preserve the semantic content of

Francesco Caravenna, Rongfeng Sun, Nikos Zygouras

We review our joint work on the scaling limits of disordered systems, linking the notion of disorder relevance/irrelevance to that of sub/super-criticality of singular SPDEs. This line of research culminated in the construction of the Critical 2D Stochastic Heat Flow (SHF), a universal process which provides a non-trivial solution to the Stochastic Heat Equa

Diego Rogel Rodriguez, Francisco Alarcon, Raul Martinez, Jorge Ramirez

In this work we have characterized the phase behaviour and the dynamics of bidimensional mixtures of active and passive Brownian particles. We have evaluated state diagrams at several concentrations of the passive components finding that, while passive agents tend to hinder phase separation, active agents force crystal-like structures on passive colloids. In

Cosmin Ilie, Jillian Paulin, Andreea Petric, Katherine Freese

The James Webb Space Telescope (JWST) has begun to revolutionize our view of the Cosmos. The discovery of Blue Monsters (i.e., ultra-compact yet very bright high-z galaxies) and the Little Red Dots (i.e., very compact dustless strong Balmer break cosmic dawn sources) pose significant challenges to pre-JWST era models of the assembly of first stars and galaxi

Hadi Ghaemi, Lauren Snyder, Markus Stocker

Digital libraries for research, such as the ACM Digital Library or Semantic Scholar, do not enable the machine-supported, efficient reuse of scientific knowledge (e.g., in synthesis research). This is because these libraries are based on document-centric models with narrative text knowledge expressions that require manual or semi-automated knowledge extracti

Yangxiao Cai, Ruiyin Li, Peng Liang, Mojtaba Shahin

As the complexity of Software Engineering (SE) tasks continues to escalate, Multi-Agent Systems (MASs) have emerged as a focal point of research and practice due to their autonomy and scalability. Furthermore, through leveraging the reasoning and planning capabilities of Large Language Models (LLMs), the application of LLM-based MASs in the field of SE is ga

Jialiang Zhu, Hamza Haif, Abdelali Arous, Huseyin Arslan

According to the recent 3GPP decisions on 6G air interface, orthogonal frequency-division multiplexing (OFDM)-based waveforms are the primary candidates for future integrated sensing and communication (ISAC) systems. In this paper, we consider a monostatic sensing scenario in which OFDM is used for the downlink and its reflected echo signal is used for sensi

F. Lienhard, A. Mortier, A. Collier Cameron, M. Cretignier

We provide a detailed characterisation of the planetary system orbiting HD 85426 (TOI-1774). This bright G-type star ($M_{\ast}$: 0.99 $\text{M}_{\odot}$; $R_{\ast}$: 1.13 $\text{R}_{\odot}$; age: 7.4 Gyr; V mag: 8.25) hosts a transiting sub-Neptune, HD 85426 b, with an orbital period of 16.71 days and a blackbody equilibrium temperature of $824^{+11}_{-11}$

Paolo Bellingeri, Celeste Damiani, Oscar Ocampo, Charalampos Stylianakis

In this work, we study the relationship between congruence subgroups $B_n[m]$ and $\mathcal{N}_n(\sigma_1^m)$ the normal closure of $\sigma_1^m$, where $\sigma_1$ is the classical generator of $B_n$. We characterize the conditions under which $\mathcal{N}_n(\sigma_1^m)$ has finite index in $B_n[m]$ and provide explicit generators for these finite quotients.

Larry Riddle

Mandelbrot and Frame studied the geometry of self-contracting symmetric binary trees in which they stated that the height of such trees occurred at the branch tip of the path consisting of branches that alternate left and right. Taylor proved that this happens for both self-avoiding as well as self-contacting symmetric binary trees (if we ignore the height o

Peng Yu, Yike Chen, Chao Xu, Albert Bifet

In the context of the Classification and Regression Trees (CART) algorithm, the efficient splitting of categorical features using standard criteria like GINI and Entropy is well-established. However, using the Mean Absolute Error (MAE) criterion for categorical features has traditionally relied on various numerical encoding methods. This paper demonstrates t

Minchi Hu

Encoding static images into spike trains is a fundamental step for enabling Spiking Neural Networks (SNNs) to process visual information. However, widely used methods such as rate coding, Poisson encoding, and time-to-first-spike (TTFS) often neglect spatial correlations and produce temporally inconsistent spike patterns, limiting both efficiency and interpr

Yongchao Lü

We present exact evaluations of superconformal indices for 4d N =1 and N =2 pure Super Yang-Mills theories with arbitrary simple gauge group G. Our approach applies the Macdonald identities for untwisted affine Lie algebras to the integral formulas of the indices, yielding uniform closed formulas valid for all G, expressed both as q-series and as eta-quotien

Gabriel M. Almeida, Jacek Kibiłda, Joao F. Santos, Kleber Vieira Cardoso

The disaggregation of base stations into discrete RAN functions introduces new threats to mobile networks, as failures in one RAN function can trigger cascading failures and disrupt the entire functional chain, impacting network performance and leading to outages. In this paper, we propose the first resilience mechanism leveraging the adaptive placement of R