id
large_stringlengths 9
16
| title
large_stringlengths 1
382
| abstract
large_stringlengths 3
6.09k
| publish_date
date32 | update_date
date32 | categories
large listlengths 1
13
| authors
large_stringlengths 3
62.8k
|
|---|---|---|---|---|---|---|
2501.07774
|
Transforming Indoor Localization: Advanced Transformer Architecture for NLOS Dominated Wireless Environments with Distributed Sensors
|
Indoor localization in challenging non-line-of-sight (NLOS) environments often leads to poor accuracy with traditional approaches. Deep learning (DL) has been applied to tackle these challenges; however, many DL approaches overlook computational complexity, especially for floating-point operations (FLOPs), making them unsuitable for resource-limited devices. Transformer-based models have achieved remarkable success in natural language processing (NLP) and computer vision (CV) tasks, motivating their use in wireless applications. However, their use in indoor localization remains nascent, and directly applying Transformers for indoor localization can be both computationally intensive and exhibit limitations in accuracy. To address these challenges, in this work, we introduce a novel tokenization approach, referred to as Sensor Snapshot Tokenization (SST), which preserves variable-specific representations of power delay profile (PDP) and enhances attention mechanisms by effectively capturing multi-variate correlation. Complementing this, we propose a lightweight Swish-Gated Linear Unit-based Transformer (L-SwiGLU-T) model, designed to reduce computational complexity without compromising localization accuracy. Together, these contributions mitigate the computational burden and dependency on large datasets, making Transformer models more efficient and suitable for resource-constrained scenarios. Experimental results on simulated and real-world datasets demonstrate that SST and L-SwiGLU-T achieve substantial accuracy and efficiency gains, outperforming larger Transformer and CNN baselines by over 40% while using significantly fewer FLOPs and training samples.
| 2025-12-26
| 2025-12-29
|
[
"cs.LG",
"cs.AI",
"eess.SP"
] |
Saad Masrur, Jung-Fu, Cheng, Atieh R. Khamesi, Ismail Guvenc
|
2512.21955
|
Superradiant and dynamical spin-down of neutron stars with gravitational wave implications
|
Neutron stars such as pulsars and magnetars lose angular momentum primarily through electromagnetic dipole radiation, gravitational waves, $r$-mode oscillation, and also affected by fallback accretion processes. However, anomalous spin variations, particularly sudden enhanced spin-down rates, indicate additional spin-down mechanisms. We propose superradiant spin-down as a potential explanation for these events. By modelling the interplay between conventional and superradiant spin-down channels, we evaluate their impact on neutron star rotational evolution. We also discuss gravitational-wave emission produced by quadrupole deformation, $r$-mode oscillations, and axion-induced bosonic clouds around an isolated neutron star, highlighting their potential as distinct multimessenger probes in upcoming detectors.
| 2025-12-26
| 2025-12-29
|
[
"astro-ph.HE",
"astro-ph.SR",
"gr-qc"
] |
Indra Kumar Banerjee, Sandeep Chatterjee, Biswarup Das, Ujjal Kumar Dey
|
2512.22120
|
See Less, See Right: Bi-directional Perceptual Shaping For Multimodal Reasoning
|
Large vision-language models (VLMs) often benefit from intermediate visual cues, either injected via external tools or generated as latent visual tokens during reasoning, but these mechanisms still overlook fine-grained visual evidence (e.g., polylines in charts), generalize poorly across domains, and incur high inference-time cost. In this paper, we propose Bi-directional Perceptual Shaping (BiPS), which transforms question-conditioned masked views into bidirectional where-to-look signals that shape perception during training. BiPS first applies a KL-consistency constraint between the original image and an evidence-preserving view that keeps only question-relevant regions, encouraging coarse but complete coverage of supporting pixels. It then applies a KL-separation constraint between the original and an evidence-ablated view where critical pixels are masked so the image no longer supports the original answer, discouraging text-only shortcuts (i.e., answering from text alone) and enforcing fine-grained visual reliance. Across eight benchmarks, BiPS boosts Qwen2.5-VL-7B by 8.2% on average and shows strong out-of-domain generalization to unseen datasets and image types.
| 2025-12-26
| 2025-12-29
|
[
"cs.CV"
] |
Shuoshuo Zhang, Yizhen Zhang, Jingjing Fu, Lei Song, Jiang Bian, Yujiu Yang, Rui Wang
|
2512.22085
|
Searching for Cosmological Collider in the Planck CMB Data II: collider templates and Modal analysis
|
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 results of improved CMB data analysis for cosmological collider signals using Planck CMB data. To set the stage, we first construct a set of simplified but characteristic collider templates which are accurate over a broad range of particle masses, spins and sound speeds. In order to break degeneracies with single-field PNG, we propose an orthogonalization scheme such that the collider templates are uncorrelated with the highly constrained equilateral and orthogonal shapes. On this basis, we deploy the Modal bispectrum estimator for the Planck analysis and perform a systematic scan of parameters to search for the most significant collider signal. The maximum signal-to-noise ratio is found to be $2.35Ï$ for massive spin-0 exchange after taking into account the look-elsewhere effect. In addition, we cross-validate the Modal analysis with the CMB-BEST pipeline, which demonstrates the consistency of results across the benchmark examples of collider templates. Given the low signal-to-noise ratio regime we find at the current stage of PNG observations, we believe the orthogonalization procedure provides an optimized strategy for future tests of the cosmological collider with the ability to rule out single field inflation.
| 2025-12-26
| 2025-12-29
|
[
"astro-ph.CO",
"gr-qc",
"hep-ph",
"hep-th"
] |
Petar Suman, Dong-Gang Wang, Wuhyun Sohn, James R. Fergusson, E. P. S. Shellard
|
2512.21966
|
Topological constraints on the electronic band structure of hexagonal lattice in a magnetic field
|
The impact of projective lattice symmetry on electronic band structures has attracted significant attention in recent years, particularly in light of growing experimental studies of two-dimensional hexagonal materials in magnetic fields. Yet, most theoretical work to date has focused on the square lattice due to its relative simplicity. In this work, we investigate the role of projective lattice symmetry in a hexagonal lattice with rational magnetic flux, emphasizing the resulting topological constraints on the electronic band structure. We show that, at pi flux, the symmetry in the hexagonal lattice enforces novel Dirac band touchings at E not equal to zero, and for general rational flux it constrains the number of Dirac points at E = 0. We further analyze the symmetry-imposed constraints on the Chern numbers of both isolated gapped bands and band multiplets connected by Dirac-point touchings. Our results demonstrate that these constraints in the hexagonal lattice differ substantially from those in the square lattice.
| 2025-12-26
| 2025-12-29
|
[
"cond-mat.mes-hall",
"cond-mat.mtrl-sci"
] |
Qi Gao, Wei Chen
|
2503.00363
|
Characterizing dynamical behaviors in topological open systems with boundary dissipations
|
We investigate the dynamics of the Su-Schrieffer-Heeger model with boundary dissipations described by Lindblad master equations and unravel distinct dynamical features in the topologically different phases of the underlying Hamiltonian. By examining the long-time damping dynamics, we uncover a dynamical duality phenomenon between the weak and strong dissipation region, which exists only in the topologically non-trivial phase, linked to the structure of the Liouvillian spectra,particularly the stripe closest to the steady state. When dissipation is confined to a single boundary, the dynamical duality phenomenon still exists. Under this condition, the Liouvillian gap fulfills an exponential size scaling relation in the topologically non-trivial phase and a power-law size scaling relation in the topologically trivial phase. Within the topologically non-trivial region, we identify the existence of boundary-localized dark states in the thermodynamical limit, which is responsible for the exponential size decay of Liouvillian gap.
| 2025-12-26
| 2025-12-29
|
[
"quant-ph",
"cond-mat.other"
] |
Zhen-Yu Zheng, Xueliang Wang, Shu Chen
|
2512.22029
|
LibContinual: A Comprehensive Library towards Realistic Continual Learning
|
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 implementations, conflicting dependencies, and varying evaluation protocols, makes fair comparison and reproducible research increasingly difficult. To address this challenge, we propose LibContinual, a comprehensive and reproducible library designed to serve as a foundational platform for realistic CL. Built upon a high-cohesion, low-coupling modular architecture, LibContinual integrates 19 representative algorithms across five major methodological categories, providing a standardized execution environment. Meanwhile, leveraging this unified framework, we systematically identify and investigate three implicit assumptions prevalent in mainstream evaluation: (1) offline data accessibility, (2) unregulated memory resources, and (3) intra-task semantic homogeneity. We argue that these assumptions often overestimate the real-world applicability of CL methods. Through our comprehensive analysis using strict online CL settings, a novel unified memory budget protocol, and a proposed category-randomized setting, we reveal significant performance drops in many representative CL methods when subjected to these real-world constraints. Our study underscores the necessity of resource-aware and semantically robust CL strategies, and offers LibContinual as a foundational toolkit for future research in realistic continual learning. The source code is available from \href{https://github.com/RL-VIG/LibContinual}{https://github.com/RL-VIG/LibContinual}.
| 2025-12-26
| 2025-12-29
|
[
"cs.LG",
"cs.AI"
] |
Wenbin Li, Shangge Liu, Borui Kang, Yiyang Chen, KaXuan Lew, Yang Chen, Yinghuan Shi, Lei Wang, Yang Gao, Jiebo Luo
|
2409.17008
|
Andreev qubit readout from dynamic interference supercurrent
|
Nondemolition protocols use ancilla qubits to identify the fragile quantum state of a qubit without destroying its encoded information, thus playing a crucial role in nondestructive quantum measurements particularly relevant for quantum error correction. However, the multitude of ancilla preparations, information transfers, and ancilla measurements in these protocols create an intrinsic overhead for information processing. Here we consider an Andreev qubit defined in a quantum-dot Josephson junction and show that the macroscopic time-dependent oscillatory supercurrent arising from the quantum interference of the many-body eigenstates, can be used to probe the qubit itself-arbitrarily close to the nondestructive limit-under currently available experimental capabilities. This readout of arbitrary superposition states of Andreev qubits avoids ancillae altogether and significantly reduces experimental overhead as no repetitive qubit reinitialization is needed. Our prediction of an AC-like Josephson effect without an applied external voltage, which enables the nondestructive qubit readout, is a unique macroscopic manifestation of the microscopic dynamics of the Andreev quantum state. Our findings should have an unprecedented impact on advancing research and applications involving Andreev dots, thus positioning them as promising qubit contenders for quantum processing and technologies.
| 2025-12-26
| 2025-12-29
|
[
"cond-mat.supr-con",
"quant-ph"
] |
Xian-Peng Zhang, Chuanchang Zeng, Zhen-Biao Yang, Jose Carlos Egues, Yugui Yao
|
2512.21972
|
On Convergence of Regularized Barzilai-Borwein Method
|
The regularized Barzilai-Borwein (RBB) method represents a promising gradient-based optimization algorithm. In this paper, by splitting the gradient into two parts and analyzing the dynamical system of difference equations governing the ratio of their magnitudes, we establish that the RBB method achieves R-linear convergence for strongly convex quadratic functions of arbitrary dimensions. Specifically, for the two-dimensional case, we provide a concise proof demonstrating that the method exhibits at least R-linear convergence. We propose a simple yet effective adaptive regularization parameter scheme to further improve its performance. A typical numerical example verifies the effectiveness of this scheme.
| 2025-12-26
| 2025-12-29
|
[
"math.OC"
] |
Xin Xu
|
2512.22381
|
PHANTOM: Physics-Aware Adversarial Attacks against Federated Learning-Coordinated EV Charging Management System
|
The rapid deployment of electric vehicle charging stations (EVCS) within distribution networks necessitates intelligent and adaptive control to maintain the grid's resilience and reliability. In this work, we propose PHANTOM, a physics-aware adversarial network that is trained and optimized through a multi-agent reinforcement learning model. PHANTOM integrates a physics-informed neural network (PINN) enabled by federated learning (FL) that functions as a digital twin of EVCS-integrated systems, ensuring physically consistent modeling of operational dynamics and constraints. Building on this digital twin, we construct a multi-agent RL environment that utilizes deep Q-networks (DQN) and soft actor-critic (SAC) methods to derive adversarial false data injection (FDI) strategies capable of bypassing conventional detection mechanisms. To examine the broader grid-level consequences, a transmission and distribution (T and D) dual simulation platform is developed, allowing us to capture cascading interactions between EVCS disturbances at the distribution level and the operations of the bulk transmission system. Results demonstrate how learned attack policies disrupt load balancing and induce voltage instabilities that propagate across T and D boundaries. These findings highlight the critical need for physics-aware cybersecurity to ensure the resilience of large-scale vehicle-grid integration.
| 2025-12-26
| 2025-12-30
|
[
"cs.ET",
"cs.LG"
] |
Mohammad Zakaria Haider, Amit Kumar Podder, Prabin Mali, Aranya Chakrabortty, Sumit Paudyal, Mohammad Ashiqur Rahman
|
2512.22072
|
Rotationally invariant dynamical lattice regulators for Euclidean quantum field theories
|
We introduce a dynamical-lattice regulator (DLR) for Euclidean quantum field theories on a fixed hypercubic graph $Î\simeq \mathbb{Z}^d$, in which the embedding $x:Î\to \mathbb{R}^d$ is promoted to a dynamical field and integrated over subject to shape-regularity constraints. The total action is local on $Î$, gauge invariant, and depends on $x$ only through Euclidean invariants built from edge vectors (local metrics, volumes, etc.), hence the partition function is exactly covariant under the global Euclidean group SE(d) at any lattice spacing. The intended symmetry-restoring mechanism is not rigid global zero modes but short-range *local twisting* of the embedding that mixes local orientations; accordingly, our universality discussion is conditioned on a short-range geometry hypothesis (SR): after quotienting the global SE(d) modes, connected correlators of local geometric observables have correlation length O(1) in lattice units.
We prove Osterwalder-Schrader reflection positivity for the coupled system with embedding $x$ and generic gauge/matter fields $(U,Φ)$ in finite volume by treating $x$ as an additional multiplet of scalar fields on $Î$. Assuming (SR), integrating out $x$ at fixed cutoff yields a local Symanzik effective action in which geometry fluctuations generate only SO(d)-invariant irrelevant operators and finite renormalizations; in particular, in $d=4$ we recover the standard one-loop $β$-function in a scalar $Ï^4$ test theory. Finally, we describe a practical local Monte Carlo update and report $d=2$ proof-of-concept simulations showing a well-behaved geometry sector and a substantial reduction of axis-vs-diagonal cutoff artifacts relative to a fixed lattice at matched bare parameters.
| 2025-12-26
| 2025-12-29
|
[
"hep-lat",
"hep-th"
] |
Tsogtgerel Gantumur
|
2512.19443
|
D2Pruner: Debiased Importance and Structural Diversity for MLLM Token Pruning
|
Processing long visual token sequences poses a significant computational burden on Multimodal Large Language Models (MLLMs). While token pruning offers a path to acceleration, we find that current methods, while adequate for general understanding, catastrophically fail on fine-grained localization tasks. We attribute this failure to the inherent flaws of the two prevailing strategies: importance-based methods suffer from a strong positional bias, an inherent model artifact that distracts from semantic content, while diversity-based methods exhibit structural blindness, disregarding the user's prompt and spatial redundancy. To address this, we introduce D2Pruner, a framework that rectifies these issues by uniquely combining debiased importance with a structural pruning mechanism. Our method first secures a core set of the most critical tokens as pivots based on a debiased attention score. It then performs a Maximal Independent Set (MIS) selection on the remaining tokens, which are modeled on a hybrid graph where edges signify spatial proximity and semantic similarity. This process iteratively preserves the most important and available token while removing its neighbors, ensuring that the supplementary tokens are chosen to maximize importance and diversity. Extensive experiments demonstrate that D2Pruner has exceptional efficiency and fidelity. Applied to LLaVA-1.5-7B for general understanding tasks, it reduces FLOPs by 74.2\% while retaining 99.2\% of its original performance. Furthermore, in challenging localization benchmarks with InternVL-2.5-8B, it maintains 85.7\% performance at a 90\% token reduction rate, marking a significant advancement with up to 63. 53\% improvement over existing methods.
| 2025-12-26
| 2025-12-29
|
[
"cs.CV"
] |
Evelyn Zhang, Fufu Yu, Aoqi Wu, Zichen Wen, Ke Yan, Shouhong Ding, Biqing Qi, Linfeng Zhang
|
2512.23744
|
Acoustic Black Holes in a Shock-Wave Exciton-Polariton Condensate
|
We demonstrate the spontaneous formation of acoustic black holes in exciton-polariton condensates triggered by discontinuous Riemann-type initial conditions. Starting from a quasi-conservative Gross-Pitaevskii model, we show that nonlinear dispersive shock waves naturally generate spatial regions where the local flow velocity exceeds the speed of sound, creating a self-induced transonic interface that functions as an acoustic horizon. Unlike previous schemes relying on externally engineered potentials or pump-loss landscapes, our approach reveals that the intrinsic nonlinear hydrodynamics of polariton fluids alone can lead to horizon formation. Using Whitham modulation theory and numerical simulations, we characterize the transition between subsonic and supersonic regimes and estimate the corresponding surface gravity and Hawking temperature. This mechanism opens a new route toward realizing polariton black holes and studying analogue gravitational effects, including Hawking-like emission, in Bose-Einstein quantum liquids.
| 2025-12-26
| 2026-01-01
|
[
"cond-mat.mes-hall",
"cond-mat.quant-gas",
"gr-qc"
] |
Junhui Cao, Jinling Wang, Kirill Bazarov, Chenqi Jin, Huijun Li, Anton Nalitov, Alexey Kavokin
|
2512.21969
|
From emission to absorption: the FAST observation of the OH 18-cm lines from the Comet C2025/A6
|
We observed comet C/2025 A6 with FAST telescope equipped with the ultra-wideband receiver from 23$^{\rm rd}$ October to 8$^{\rm th}$ November 2025 and detected the OH 18-cm lines for the first time. The OH lines underwent a reversal from emission to absorption from 23$^{\rm rd}$ October to 5$^{\rm th}$ November, which is mainly caused by variations in the heliocentric velocity. Through trapezoidal fitting of the OH line profiles, we derive expansion velocities of the water that rise as the heliocentric distance decreases. Based on these results, we estimated the OH production rates of C/2025 A6 for 23$^{\rm rd}$ October, 26$^{\rm th}$ October, 4$^{\rm th}$ November, and 5$^{\rm th}$ November and it presents a significant upward trend.
| 2025-12-26
| 2025-12-29
|
[
"astro-ph.EP",
"astro-ph.GA"
] |
Dongyue Jiang, Lei Qian, Minglei Guo, Qiaoli Hao, Menglin Huang, Peng Jiang, Hongfei Liu, Chun Sun, Xingyi Wang, Qingliang Yang, Naiping Yu, Lei Zhao, Yutao Zhao, Liyun Zhang, Yichi Zhang, Tongjie Zhang, Zhichen Pan
|
2512.22358
|
Can high-$p_\perp$ theory and data constrain $η/s$?
|
Understanding the temperature dependence of the specific shear viscosity $(η/s)$ is crucial for characterizing the properties of the QCD matter produced in ultrarelativistic heavy-ion collisions. Since, low-$p_\perp$ theory and data are only weakly sensitive to the typical forms of $η/s(T)$, especially at high temperatures, we use high-$p_\perp$ data and theory to impose additional constraints on it. Our approach, based on dynamical radiative and collisional energy loss of high-$p_\perp$ particles, provides promising results in constraining the temperature dependence of $η/s$.
| 2025-12-26
| 2025-12-30
|
[
"hep-ph"
] |
Bithika Karmakar, Dusan Zigic, Igor Salom, Jussi Auvinen, Pasi Huovinen, Marko Djordjevic, Magdalena Djordjevic
|
2405.05512
|
Characteristic Learning for Provable One Step Generation
|
We propose the characteristic generator, a novel one-step generative model that combines the efficiency of sampling in Generative Adversarial Networks (GANs) with the stable performance of flow-based models. Our model is driven by characteristics, along which the probability density transport can be described by ordinary differential equations (ODEs). Specifically, we first estimate the underlying velocity field and use the Euler method to solve the probability flow ODE, generating discrete approximations of the characteristics. A deep neural network is then trained to fit these characteristics, creating a one-step map that pushes a simple Gaussian distribution to the target distribution. In the theoretical aspect, we provide a comprehensive analysis of the errors arising from velocity matching, Euler discretization, and characteristic fitting to establish a non-asymptotic convergence rate in the 2-Wasserstein distance under mild data assumptions. Crucially, we demonstrate that under a standard manifold assumption, this convergence rate depends only on the intrinsic dimension of data rather than the much larger ambient dimension, proving our model's ability to mitigate the curse of dimensionality. To our knowledge, this is the first rigorous convergence analysis for a flow-based one-step generative model. Experiments on both synthetic and real-world datasets demonstrate that the characteristic generator achieves high-quality and high-resolution sample generation with the efficiency of just a single neural network evaluation.
| 2025-12-26
| 2025-12-29
|
[
"cs.LG",
"cs.AI",
"cs.NA",
"math.NA",
"math.ST",
"stat.TH"
] |
Zhao Ding, Chenguang Duan, Yuling Jiao, Ruoxuan Li, Jerry Zhijian Yang, Pingwen Zhang
|
2512.21929
|
Impact of the sodium and calcium chlorides uptake on the interfacial behavior of ice: premelting, structure, and dynamics
|
Hypothesis: Seawater ice and frozen aqueous solutions in contact with air can exhibit a thin quasi-brine surface layer intruding between ice and vapor, but a detailed characterization of surface properties and its relation to three phase coexistence has been lacking. Using thermodynamic arguments we show how it is possible to characterize the surface layers by comparison to the three phase ice-brine-air bulk phase diagram, despite the difficulty to control or monitor all of the relevant thermodynamic fields of the two component system.
Simulations: We performed computer simulations of surface briny layers of sodium and calcium chloride adsorbed on ice. Using suitable order parameters and a rigorous geometrical dividing surface, we are able to characterize the layer's thermodynamic state, measure its properties and relate them to the corresponding properties of the bulk solution.
Results: Our results confirm that undersaturated briny surface layers can form down to the eutectic point, with a maximum concentration that is bound by the liquidus line of the ice-brine phase diagram. Such layers are distinct from finite size realizations of three phase coexistence, and can be regarded as genuine surface states, but their salt content can increase the premelting layer thickness by a factor of two or more. Owing to this significant thickness, these layers can be related to bulk electrolyte solutions of similar concentration, both as regards the structural organization of ions and the dynamical properties of the quasi-liquid film.
| 2025-12-26
| 2025-12-29
|
[
"cond-mat.soft",
"physics.ao-ph",
"physics.chem-ph",
"physics.geo-ph"
] |
Åukasz Baran, Luis G. MacDowell
|
2512.21815
|
Few Tokens Matter: Entropy Guided Attacks on Vision-Language Models
|
Vision-language models (VLMs) achieve remarkable performance but remain vulnerable to adversarial attacks. Entropy, a measure of model uncertainty, is strongly correlated with the reliability of VLM. Prior entropy-based attacks maximize uncertainty at all decoding steps, implicitly assuming that every token contributes equally to generation instability. We show instead that a small fraction (about 20%) of high-entropy tokens, i.e., critical decision points in autoregressive generation, disproportionately governs output trajectories. By concentrating adversarial perturbations on these positions, we achieve semantic degradation comparable to global methods while using substantially smaller budgets. More importantly, across multiple representative VLMs, such selective attacks convert 35-49% of benign outputs into harmful ones, exposing a more critical safety risk. Remarkably, these vulnerable high-entropy forks recur across architecturally diverse VLMs, enabling feasible transferability (17-26% harmful rates on unseen targets). Motivated by these findings, we propose Entropy-bank Guided Adversarial attacks (EGA), which achieves competitive attack success rates (93-95%) alongside high harmful conversion, thereby revealing new weaknesses in current VLM safety mechanisms.
| 2025-12-26
| 2025-12-29
|
[
"cs.CV",
"cs.LG"
] |
Mengqi He, Xinyu Tian, Xin Shen, Jinhong Ni, Shu Zou, Zhaoyuan Yang, Jing Zhang
|
2512.21916
|
Patch as Node: Human-Centric Graph Representation Learning for Multimodal Action Recognition
|
While human action recognition has witnessed notable achievements, multimodal methods fusing RGB and skeleton modalities still suffer from their inherent heterogeneity and fail to fully exploit the complementary potential between them. In this paper, we propose PAN, the first human-centric graph representation learning framework for multimodal action recognition, in which token embeddings of RGB patches containing human joints are represented as spatiotemporal graphs. The human-centric graph modeling paradigm suppresses the redundancy in RGB frames and aligns well with skeleton-based methods, thus enabling a more effective and semantically coherent fusion of multimodal features. Since the sampling of token embeddings heavily relies on 2D skeletal data, we further propose attention-based post calibration to reduce the dependency on high-quality skeletal data at a minimal cost interms of model performance. To explore the potential of PAN in integrating with skeleton-based methods, we present two variants: PAN-Ensemble, which employs dual-path graph convolution networks followed by late fusion, and PAN-Unified, which performs unified graph representation learning within a single network. On three widely used multimodal action recognition datasets, both PAN-Ensemble and PAN-Unified achieve state-of-the-art (SOTA) performance in their respective settings of multimodal fusion: separate and unified modeling, respectively.
| 2025-12-26
| 2025-12-29
|
[
"cs.CV"
] |
Zeyu Liang, Hailun Xia, Naichuan Zheng
|
2512.22005
|
Asymptotics for the spectrum of the Laplacian in thin bars with varying cross sections
|
We consider spectral problems for Laplace operator in 3D rod structures with a small cross section of diameter $O(\varepsilon)$, $\varepsilon$ being a positive parameter. The boundary conditions are Dirichlet (Neumann, respectively) on the bases of this structure and Neumann on the lateral boundary. As $\varepsilon\to 0$, we show the convergence of the spectrum with conservation of the multiplicity towards that of a 1D spectral model with Dirichlet (Neumann, respectively) boundary conditions. This 1D model may arise in diffusion or vibrations models of nonhomogeneous media with different physical characteristics and it takes into account the geometry of the 3D domain. We deal with the low frequencies and the approach to eigenfunctions in the suitable Sobolev spaces is also outlined.
| 2025-12-26
| 2025-12-29
|
[
"math.AP"
] |
Pablo Benavent-Ocejo, Delfina Gómez, Maria-Eugenia Pérez-MartÃnez
|
2512.21879
|
A Communication-Efficient Distributed Algorithm for Learning with Heterogeneous and Structurally Incomplete Multi-Site Data
|
In multicenter biomedical research, integrating data from multiple decentralized sites provides more robust and generalizable findings due to its larger sample size and the ability to account for the between-site heterogeneity. However, sharing individual-level data across sites is often difficult due to patient privacy concerns and regulatory restrictions. To overcome this challenge, many distributed algorithms, that fit a global model by only communicating aggregated information across sites, have been proposed. A major challenge in applying existing distributed algorithms to real-world data is that their validity often relies on the assumption that data across sites are independently and identically distributed, which is frequently violated in practice. In biomedical applications, data distributions across clinical sites can be heterogeneous. Additionally, the set of covariates available at each site may vary due to different data collection protocols. We propose a distributed inference framework for data integration in the presence of both distribution heterogeneity and data structural heterogeneity. By modeling heterogeneous and structurally missing data using density-tilted generalized method of moments, we developed a general aggregated data-based distributed algorithm that is communication-efficient and heterogeneity-aware. We establish the asymptotic properties of our estimator and demonstrate the validity of our method via simulation studies.
| 2025-12-26
| 2025-12-29
|
[
"stat.ME"
] |
Xiaokang Liu, Yuchen Yang, Yifei Sun, Jiang Bian, Yanyuan Ma, Raymond J. Carroll, Yong Chen
|
2512.22086
|
Unlocking klockmannite: formation of colloidal quasi-2D CuSe nanocrystals and photo-physical properties arising from crystal anisotropy
|
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 tuning the injection temperature and precursor concentrations without any additional ligands. This approach produces large klockmannite nanosheets with lateral sizes from 200 nm to several micrometres, as well as uniform triangular nanoplatelets with sizes of 12-25 nm that are monocrystalline and display strong NIR plasmonic absorption. The spectral features of the anisotropic klockmannite phase in the NIR have been analysed using complex-scaled discrete dipole approximation (CSDDA) calculations, which reveal pronounced optical anisotropy and the emergence of hyperbolic regime. The combined effect of propagating and evanescent fields is regarded as the underlying reason of such modes in the hyperbolic domain. Finally, the ultrafast photophysical behaviour of the material in klockmannite phase is examined, including hot-hole cooling, trapping, and coherent phonons generation. Our findings emphasize the important role of the intrinsic crystal anisotropy in governing the physical properties of nanoscale klockmannite.
| 2025-12-26
| 2025-12-29
|
[
"cond-mat.mtrl-sci"
] |
Urvi Parekh, Nadiia Didukh, Samira Dabelstein, Ronja Piehler, Eugen Klein, Jivesh Kaushal, Tobias Korn, Stefan Lochbrunner, Christian Klinke, Stefan Scheel, Rostyslav Lesyuk
|
2601.00838
|
School Transport Electrification -- Adoption, Strategies, Methods and Policy: A Comprehensive Review
|
The move towards electric school buses (ESBs) marks a critical step in creating a healthier and more sustainable future for students. To meet the ambitious goal of zero-emission school buses by 2035,this review focuses on the need assessment, practices, gaps, challenges, and way forward. We conducted a comprehensive assessment of more than 100 relevant sources, resulting in a final investigation. In-depth, systematic, and qualitative content analysis with SWOT analysis produced critical insights into school transport electrification. The results showed that 1.8% of the total buses in the US have already been converted to electric, where California alone owning 29% of the buses. Subsidies from various agencies and programs have contributed to the rapid growth of electrification. However, challenges in cost, technology, and policies must be mitigated through innovation and stakeholder partnerships. Policy support is boosting subsidies, industry investment and market readiness. Equitable policy is important to support underserved and disadvantaged populations, which can be addressed through four key dimensions of equity: procedural, recognition, distributive, and reparative equity. Furthermore, the traditional bus deployment model is still the most common, whereas Transportation-as-a-Service (TaaS) is an innovative ESB deployment model with the potential to accelerate ESB adoption by integrating vehicle-to-grid. SWOT analysis indicated that the achievement of the zero-emission goal, autonomous driving, and repowered vehicle technology are the greatest opportunities. Dynamic electrification strategies, V2G technology and system resiliency are yet to be discovered, which could be crucial for mass electrification.
| 2025-12-26
| 2026-01-06
|
[
"physics.soc-ph"
] |
Megh Bahadur KC, Ziqi Song
|
2512.22325
|
Quadratic-Phase Dunkl Transform: Fundamental properties, translation operators, convolution product and HUP
|
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 $μ\geq -1/2$. We define the transform and establish its fundamental properties, including continuity, a Riemann--Lebesgue lemma, linearity, scaling, and most importantly, a reversibility theorem and an associated Parseval formula. We show that this novel quadratic-phase integral type transform generalizes a wide class of known transforms, such as the quadratic-phase Fourier-Bessel transform, the quadratic-phase Fourier transform, the linear canonical Dunkl transform, the fractional Dunkl transform, and the classical Dunkl transform, by choosing the appropriate specialization of its parameters. Furthermore, we introduce and investigate a corresponding quadratic-phase Dunkl translation operator and a convolution structure, proving their basic properties and a Young's inequality. Finally, we establish a new Heisenberg-type uncertainty principle for the quadratic-phase Dunkl transform, which extends the classical uncertainty principle for a large class of integral type transforms.
| 2025-12-26
| 2025-12-30
|
[
"math.GM"
] |
Ahmed Saoudi
|
2512.22046
|
Backdoor Attacks on Prompt-Driven Video Segmentation Foundation Models
|
Prompt-driven Video Segmentation Foundation Models (VSFMs) such as SAM2 are increasingly deployed in applications like autonomous driving and digital pathology, raising concerns about backdoor threats. Surprisingly, we find that directly transferring classic backdoor attacks (e.g., BadNet) to VSFMs is almost ineffective, with ASR below 5\%. To understand this, we study encoder gradients and attention maps and observe that conventional training keeps gradients for clean and triggered samples largely aligned, while attention still focuses on the true object, preventing the encoder from learning a distinct trigger-related representation. To address this challenge, we propose BadVSFM, the first backdoor framework tailored to prompt-driven VSFMs. BadVSFM uses a two-stage strategy: (1) steer the image encoder so triggered frames map to a designated target embedding while clean frames remain aligned with a clean reference encoder; (2) train the mask decoder so that, across prompt types, triggered frame-prompt pairs produce a shared target mask, while clean outputs stay close to a reference decoder. Extensive experiments on two datasets and five VSFMs show that BadVSFM achieves strong, controllable backdoor effects under diverse triggers and prompts while preserving clean segmentation quality. Ablations over losses, stages, targets, trigger settings, and poisoning rates demonstrate robustness to reasonable hyperparameter changes and confirm the necessity of the two-stage design. Finally, gradient-conflict analysis and attention visualizations show that BadVSFM separates triggered and clean representations and shifts attention to trigger regions, while four representative defenses remain largely ineffective, revealing an underexplored vulnerability in current VSFMs.
| 2025-12-26
| 2025-12-29
|
[
"cs.CV",
"cs.CR"
] |
Zongmin Zhang, Zhen Sun, Yifan Liao, Wenhan Dong, Xinlei He, Xingshuo Han, Shengmin Xu, Xinyi Huang
|
2512.23750
|
Machine Learning Invariants of Tensors
|
We propose a data-driven approach to identifying the functionally independent invariants that can be constructed from a tensor with a given symmetry structure. Our algorithm proceeds by first enumerating graphs, or tensor networks, that represent inequivalent contractions of a product of tensors, computing instances of these scalars using randomly generated data, and then seeking linear relations between invariants using numerical linear algebra. Such relations yield syzygies, or functional dependencies relating different invariants. We apply this approach in an extended case study of the independent invariants that can be constructed from an antisymmetric $3$-form $H_{μνÏ}$ in six dimensions, finding five independent invariants. This result confirms that the most general Lagrangian for such a $3$-form, which depends on $H_{μνÏ}$ but not its derivatives, is an arbitrary function of five variables, and we give explicit formulas relating other invariants to the five independent scalars in this generating set.
| 2025-12-26
| 2026-01-01
|
[
"hep-th",
"math.AC"
] |
Athithan Elamaran, Christian Ferko, Sterling Scarlett
|
2507.04584
|
S$^2$Edit: Text-Guided Image Editing with Precise Semantic and Spatial Control
|
Recent advances in diffusion models have enabled high-quality generation and manipulation of images guided by texts, as well as concept learning from images. However, naive applications of existing methods to editing tasks that require fine-grained control, e.g., face editing, often lead to suboptimal solutions with identity information and high-frequency details lost during the editing process, or irrelevant image regions altered due to entangled concepts. In this work, we propose S$^2$Edit, a novel method based on a pre-trained text-to-image diffusion model that enables personalized editing with precise semantic and spatial control. We first fine-tune our model to embed the identity information into a learnable text token. During fine-tuning, we disentangle the learned identity token from attributes to be edited by enforcing an orthogonality constraint in the textual feature space. To ensure that the identity token only affects regions of interest, we apply object masks to guide the cross-attention maps. At inference time, our method performs localized editing while faithfully preserving the original identity with semantically disentangled and spatially focused identity token learned. Extensive experiments demonstrate the superiority of S$^2$Edit over state-of-the-art methods both quantitatively and qualitatively. Additionally, we showcase several compositional image editing applications of S$^2$Edit such as makeup transfer.
| 2025-12-26
| 2025-12-29
|
[
"cs.CV"
] |
Xudong Liu, Zikun Chen, Ruowei Jiang, Ziyi Wu, Kejia Yin, Han Zhao, Parham Aarabi, Igor Gilitschenski
|
2512.21911
|
Accelerate Speculative Decoding with Sparse Computation in Verification
|
Speculative decoding accelerates autoregressive language model inference by verifying multiple draft tokens in parallel. However, the verification stage often becomes the dominant computational bottleneck, especially for long-context inputs and mixture-of-experts (MoE) models. Existing sparsification methods are designed primarily for standard token-by-token autoregressive decoding to remove substantial computational redundancy in LLMs. This work systematically adopts different sparse methods on the verification stage of the speculative decoding and identifies structured redundancy across multiple dimensions. Based on these observations, we propose a sparse verification framework that jointly sparsifies attention, FFN, and MoE components during the verification stage to reduce the dominant computation cost. The framework further incorporates an inter-draft token and inter-layer retrieval reuse strategy to further reduce redundant computation without introducing additional training. Extensive experiments across summarization, question answering, and mathematical reasoning datasets demonstrate that the proposed methods achieve favorable efficiency-accuracy trade-offs, while maintaining stable acceptance length.
| 2025-12-26
| 2025-12-29
|
[
"cs.CL"
] |
Jikai Wang, Jianchao Tan, Yuxuan Hu, Jiayu Qin, Yerui Sun, Yuchen Xie, Xunliang Cai, Juntao Li, Min Zhang
|
2512.22054
|
Proceedings First Workshop on Adaptable Cloud Architectures
|
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.
| 2025-12-26
| 2025-12-29
|
[
"cs.SE",
"cs.DC"
] |
Giuseppe De Palma, Saverio Giallorenzo
|
2512.14693
|
Universal Reasoning Model
|
Universal transformers (UTs) have been widely used for complex reasoning tasks such as ARC-AGI and Sudoku, yet the specific sources of their performance gains remain underexplored. In this work, we systematically analyze UTs variants and show that improvements on ARC-AGI primarily arise from the recurrent inductive bias and strong nonlinear components of Transformer, rather than from elaborate architectural designs. Motivated by this finding, we propose the Universal Reasoning Model (URM), which enhances the UT with short convolution and truncated backpropagation. Our approach substantially improves reasoning performance, achieving state-of-the-art 53.8% pass@1 on ARC-AGI 1 and 16.0% pass@1 on ARC-AGI 2. Our code is avaliable at https://github.com/UbiquantAI/URM.
| 2025-12-26
| 2025-12-29
|
[
"cs.AI"
] |
Zitian Gao, Lynx Chen, Yihao Xiao, He Xing, Ran Tao, Haoming Luo, Joey Zhou, Bryan Dai
|
2512.21919
|
SWE-RM: Execution-free Feedback For Software Engineering Agents
|
Execution-based feedback like unit testing is widely used in the development of coding agents through test-time scaling (TTS) and reinforcement learning (RL). This paradigm requires scalable and reliable collection of unit test cases to provide accurate feedback, and the resulting feedback is often sparse and cannot effectively distinguish between trajectories that are both successful or both unsuccessful. In contrast, execution-free feedback from reward models can provide more fine-grained signals without depending on unit test cases. Despite this potential, execution-free feedback for realistic software engineering (SWE) agents remains underexplored. Aiming to develop versatile reward models that are effective across TTS and RL, however, we observe that two verifiers with nearly identical TTS performance can nevertheless yield very different results in RL. Intuitively, TTS primarily reflects the model's ability to select the best trajectory, but this ability does not necessarily generalize to RL. To address this limitation, we identify two additional aspects that are crucial for RL training: classification accuracy and calibration. We then conduct comprehensive controlled experiments to investigate how to train a robust reward model that performs well across these metrics. In particular, we analyze the impact of various factors such as training data scale, policy mixtures, and data source composition. Guided by these investigations, we introduce SWE-RM, an accurate and robust reward model adopting a mixture-of-experts architecture with 30B total parameters and 3B activated during inference. SWE-RM substantially improves SWE agents on both TTS and RL performance. For example, it increases the accuracy of Qwen3-Coder-Flash from 51.6% to 62.0%, and Qwen3-Coder-Max from 67.0% to 74.6% on SWE-Bench Verified using TTS, achieving new state-of-the-art performance among open-source models.
| 2025-12-26
| 2025-12-29
|
[
"cs.CL"
] |
KaShun Shum, Binyuan Hui, Jiawei Chen, Lei Zhang, X. W., Jiaxi Yang, Yuzhen Huang, Junyang Lin, Junxian He
|
2512.22405
|
Thick brane in Palatini formalism with a non-minimally coupled bulk scalar field
|
We study a thick brane scenario within the Palatini formulation of gravity, where the metric and affine connection are treated as independent variables. By introducing a non-minimal coupling between a bulk scalar field and the Ricci scalar, we obtain analytic solutions under a flat, four-dimensional Poincaré-invariant metric with a kink-like scalar configuration. The warp factor exhibits a bell-shaped profile, while the scalar potential forms a symmetric volcano-like structure, characteristic of a finite-thickness brane. The corresponding energy density is regular and localized, featuring a central peak with symmetrically placed negative minima. Through the analysis of linear tensor perturbations, we derive a Schrödinger-like equation with supersymmetric factorization, ensuring the absence of tachyonic modes and thus the stability of the background configuration. The effective potential also takes a volcano-like form that supports a localized graviton zero mode, confirming the recovery of four-dimensional gravity on the brane. A numerical study of the massive Kaluza--Klein spectrum reveals the progressive delocalization of massive modes into the bulk.
Our results demonstrate a stable and physically consistent thick brane configuration within the Palatini gravity framework, offering new insights into gravity localization and braneworld phenomenology.
| 2025-12-26
| 2026-01-06
|
[
"gr-qc",
"hep-th"
] |
Tahereh Azizi, Mojtaba Alimoradi
|
2112.05255
|
The dual approach to the $K(Ï, 1)$ conjecture
|
Dual presentations of Coxeter groups have recently led to breakthroughs in our understanding of affine Artin groups. In particular, they led to the proof of the $K(Ï, 1)$ conjecture and to the solution of the word problem. Will the "dual approach" extend to more general classes of Coxeter and Artin groups? In this paper, we describe the techniques used to prove the $K(Ï, 1)$ conjecture for affine Artin groups and we ask a series of questions that are mostly open beyond the spherical and affine cases.
| 2025-12-26
| 2025-12-30
|
[
"math.GR",
"math.AT",
"math.CO"
] |
Giovanni Paolini
|
2512.22119
|
Charge-Informed Quantum Error Correction
|
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 continuous phase transition in a disordered two-dimensional integer loop model on the Nishimori line. Using an effective replica field theory analysis and Monte Carlo numerics, we show that the optimal decoding transition exhibits Berezinskii-Kosterlitz-Thouless universality with a modified universal jump in winding number variance. We further generalize the model beyond the Nishimori line, which defines a large class of suboptimal decoders. At low nonzero temperatures and strong disorder, we find numerical evidence of a disorder-dominated loop-glass phase which corresponds to a "confidently incorrect" decoder. The zero-temperature limit defines the minimum-cost flow decoder, which serves as the ${\rm U}(1)$ analog of minimum-weight perfect matching in $\mathbb{Z}_2$ topological codes. Both the optimal and minimum-cost flow decoders are shown to dramatically outperform the charge-agnostic optimal decoder in symmetry-enriched topological codes.
| 2025-12-26
| 2025-12-29
|
[
"quant-ph",
"cond-mat.stat-mech",
"cond-mat.str-el"
] |
Vlad Temkin, Zack Weinstein, Ruihua Fan, Daniel Podolsky, Ehud Altman
|
2512.22333
|
Emotion classification using EEG headset signals and Random Forest
|
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 the ability of computers to understand the communication between humans. This study proposes the creation of a model that allows the classification of people's emotions based on their EEG signals, for which the brain-computer interface EMOTIV EPOC was used. This allowed the collection of electroencephalographic information from 50 people, all of whom were shown audiovisual resources that helped to provoke the desired mood. The information obtained was stored in a database for the generation of the model and the corresponding classification analysis. Random Forest model was created for emotion prediction (happiness, sadness and relaxation), based on the signals of any person. The results obtained were 97.21% accurate for happiness, 76% for relaxation and 76% for sadness. Finally, the model was used to generate a real-time emotion prediction algorithm; it captures the person's EEG signals, executes the generated algorithm and displays the result on the screen with the help of images representative of each emotion.
| 2025-12-26
| 2025-12-30
|
[
"cs.HC",
"cs.LG"
] |
Ricardo Vasquez, Diego RiofrÃo-Luzcando, Joe Carrion-Jumbo, Cesar Guevara
|
2510.10313
|
Low-cost Pyranometer-Based ANN Approach for MPPT in Solar PV Systems
|
This article presents a study on the application of artificial neural networks (ANNs) for maximum power point tracking (MPPT) in photovoltaic (PV) systems using low-cost pyranometer sensors. The proposed approach integrates pyranometers, temperature sensors, and an ANN to estimate the duty cycle of a DC/DC converter, enabling the system to consistently operate at its maximum power point. The strategy was implemented in the local control of a Cuk converter and experimentally validated against the conventional Perturb and Observe (P&O) method. Results demonstrate that the ANN-based technique, leveraging affordable sensor technology, achieves accurate MPPT performance with reduced fluctuations, enhancing the responsiveness and efficiency of PV tracking systems.
| 2025-12-26
| 2025-12-29
|
[
"eess.SY",
"cs.SY"
] |
Luiz Fernando M. Arruda, Moises Ferber, Diego Greff
|
2502.09990
|
X-Boundary: Establishing Exact Safety Boundary to Shield LLMs from Multi-Turn Jailbreaks without Compromising Usability
|
Despite the rapid development of safety alignment techniques for LLMs, defending against multi-turn jailbreaks is still a challenging task. In this paper, we conduct a comprehensive comparison, revealing that some existing defense methods can improve the robustness of LLMs against multi-turn jailbreaks but compromise usability, i.e., reducing general capabilities or causing the over-refusal problem. From the perspective of mechanism interpretability of LLMs, we discover that these methods fail to establish a boundary that exactly distinguishes safe and harmful feature representations. Therefore, boundary-safe representations close to harmful representations are inevitably disrupted, leading to a decline in usability. To address this issue, we propose X-Boundary to push harmful representations away from boundary-safe representations and obtain an exact distinction boundary. In this way, harmful representations can be precisely erased without disrupting safe ones. Experimental results show that X-Boundary achieves state-of-the-art defense performance against multi-turn jailbreaks, while reducing the over-refusal rate by about 20% and maintaining nearly complete general capability. Furthermore, we theoretically prove and empirically verify that X-Boundary can accelerate the convergence process during training. Please see our code at: https://github.com/AI45Lab/X-Boundary.
| 2025-12-26
| 2025-12-29
|
[
"cs.CR",
"cs.AI",
"cs.CL",
"cs.CV",
"cs.LG"
] |
Xiaoya Lu, Dongrui Liu, Yi Yu, Luxin Xu, Jing Shao
|
2511.18905
|
Congruences Modulo Powers of 7 for the Reciprocal Crank Parity Function
|
Amdeberhan and Merca recently studied arithmetic properties of the sequence $a(n)$, the reciprocal of the crank parity function, which counts the number of integer partitions of weight $n$ whose even parts are monochromatic and whose odd parts may appear in one of three colors (OEIS A298311). A key result of their work was the congruence $a(7n + 2) \equiv 0 \pmod{7}$ for all $n \geq 0$. We prove new congruences for the reciprocal crank parity function modulo powers of $7$.
| 2025-12-26
| 2025-12-29
|
[
"math.CO",
"math.NT"
] |
Dandan Chen
|
2512.22042
|
Esakia order-compactifications and locally Esakia spaces
|
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 taking the largest Esakia order-compacification is functorial, thus obtaining an analog of Banaschewski's theorem.
| 2025-12-26
| 2025-12-29
|
[
"math.LO",
"math.GN"
] |
Rodrigo Nicolau Almeida, Guram Bezhanishvili, Nick Bezhanishvili
|
2512.22012
|
Generalized binomial edge ideals are Cartwright-Sturmfels
|
Binomial edge ideals associated to a simple graph G were introduced by Herzog and collaborators and, independently, by Ohtani. They became an ``instant classic" in combinatorial commutative algebra with more than 100 papers devoted to their investigation over the past 15 years. They exhibit many striking properties, including being radical and, moreover, Cartwright-Sturmfels. Using the fact that binomial edge ideals can be seen as ideals of 2-minors of a matrix of variables with two rows, generalized binomial edge ideals of 2-minors of matrices of m rows were introduced by Rauh and proved to be radical. The goal of this paper is to prove that generalized binomial edge ideals are Cartwright-Sturmfels. On the way we provide results on ideal constructions preserving the Cartwright-Sturmfels property. We also give examples and counterexamples to the Cartwright-Sturmfels property for higher minors.
| 2025-12-26
| 2025-12-29
|
[
"math.AC",
"math.CO"
] |
Aldo Conca, Emanuela De Negri, Volkmar Welker
|
2507.15898
|
A Generative Model for Disentangling Galaxy Photometric Parameters
|
Ongoing and future photometric surveys will produce unprecedented volumes of galaxy images, necessitating robust, efficient methods for deriving galaxy morphological parameters at scale. Traditional approaches, such as parametric light-profile fitting, offer valuable insights but become computationally prohibitive when applied to billions of sources. In this work, we propose a Conditional AutoEncoder (CAE) framework to simultaneously model and characterize galaxy morphology. Our CAE is trained on a suite of realistic mock galaxy images generated via GalSim, encompassing a broad range of galaxy types, photometric parameters (e.g., flux, half-light radius, Sersic index, ellipticity), and observational conditions. By encoding each galaxy image into a low-dimensional latent representation conditioned on key parameters, our model effectively recovers these morphological features in a disentangled manner, while also reconstructing the original image. The results demonstrate that the CAE approach can accurately and efficiently infer complex structural properties, offering a powerful alternative to existing methods.
| 2025-12-26
| 2025-12-30
|
[
"astro-ph.IM",
"astro-ph.GA",
"cs.AI"
] |
Keen Leung, Colen Yan, Jun Yin
|
2512.22026
|
Proton therapy range uncertainty reduction using vendor-agnostic tissue characterization on a virtual photon-counting CT head scan
|
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 approach and a prototype software, TissueXplorer. A validated CT simulator (DukeSim) was used to generate photon-counting CT projections of a computational head model, which were reconstructed with an open-source toolbox (ASTRA). The dose of 2 Gy was delivered through protons in a single fraction to target two different cases of nasal and brain tumors with a single lateral beam angle. Ground truth treatment plan was made directly on the computational head model using clinical treatment planning software (RayStation). This plan was then recalculated on the corresponding CT images for which SPR values were estimated using both the conventional method and the prototype software TissueXplorer. The mean percentage difference in estimating the stopping power ratio with TissueXplorer in all head tissues inside the scanned volume was 0.28%. Stopping power ratios obtained with this method showed smaller dose distribution differences from the ground truth plan than the conventional stoichiometric calibration method on the computational head model. Virtual imaging offers an alternative approach to validation of the SPR prediction from CT imaging, as well as its effect on the dose distribution and thus downstream clinical outcomes. According to this simulation study, software solutions that utilize spectral information, such as TissueXplorer, hold promise for more accurate prediction of the stopping power ratio than the conventional stoichiometric approach.
| 2025-12-26
| 2025-12-29
|
[
"physics.med-ph"
] |
S. VrbaÅ¡ki, G. StaniÄ, S. Mollineli, M. Bhattarai, E. Abadi, M. Ciocca, E. Samei
|
2512.20308
|
SpidR: Learning Fast and Stable Linguistic Units for Spoken Language Models Without Supervision
|
The parallel advances in language modeling and speech representation learning have raised the prospect of learning language directly from speech without textual intermediates. This requires extracting semantic representations directly from speech. Our contributions are threefold. First, we introduce SpidR, a self-supervised speech representation model that efficiently learns representations with highly accessible phonetic information, which makes it particularly suited for textless spoken language modeling. It is trained on raw waveforms using a masked prediction objective combined with self-distillation and online clustering. The intermediate layers of the student model learn to predict assignments derived from the teacher's intermediate layers. This learning objective stabilizes the online clustering procedure compared to previous approaches, resulting in higher quality codebooks. SpidR outperforms wav2vec 2.0, HuBERT, WavLM, and DinoSR on downstream language modeling benchmarks (sWUGGY, sBLIMP, tSC). Second, we systematically evaluate across models and layers the correlation between speech unit quality (ABX, PNMI) and language modeling performance, validating these metrics as reliable proxies. Finally, SpidR significantly reduces pretraining time compared to HuBERT, requiring only one day of pretraining on 16 GPUs, instead of a week. This speedup is enabled by the pretraining method and an efficient codebase, which allows faster iteration and easier experimentation. We open-source the training code and model checkpoints at https://github.com/facebookresearch/spidr.
| 2025-12-26
| 2025-12-29
|
[
"cs.CL",
"cs.SD",
"eess.AS"
] |
Maxime Poli, Mahi Luthra, Youssef Benchekroun, Yosuke Higuchi, Martin Gleize, Jiayi Shen, Robin Algayres, Yu-An Chung, Mido Assran, Juan Pino, Emmanuel Dupoux
|
2512.21898
|
Flexible Multitask Learning with Factorized Diffusion Policy
|
Multitask learning poses significant challenges due to the highly multimodal and diverse nature of robot action distributions. However, effectively fitting policies to these complex task distributions is often difficult, and existing monolithic models often underfit the action distribution and lack the flexibility required for efficient adaptation. We introduce a novel modular diffusion policy framework that factorizes complex action distributions into a composition of specialized diffusion models, each capturing a distinct sub-mode of the behavior space for a more effective overall policy. In addition, this modular structure enables flexible policy adaptation to new tasks by adding or fine-tuning components, which inherently mitigates catastrophic forgetting. Empirically, across both simulation and real-world robotic manipulation settings, we illustrate how our method consistently outperforms strong modular and monolithic baselines.
| 2025-12-26
| 2025-12-29
|
[
"cs.RO",
"cs.AI"
] |
Chaoqi Liu, Haonan Chen, Sigmund H. Høeg, Shaoxiong Yao, Yunzhu Li, Kris Hauser, Yilun Du
|
2411.07103
|
On a connection between total positivity and Bernoulli stopping problems
|
Consider a discrete-time optimal selection problem where one observes a sequence of independent Bernoulli trials and receives a nonnegative reward upon stopping on a success. The aim is to find a single-choice strategy that maximises the expected payoff. These Bernoulli stopping problems are characterised by two key properties: (i) a recurrence relation connecting the reward sequence to the continuation payoff sequence, and (ii) the total positivity of the Markov chain embedded in success epochs of the trials. The recurrence is fundamental in proving the optimality of the myopic strategy under unimodal continuation payoff sequence, while the total positivity ensures that the expectation of a quasi-unimodal function of the chain remains quasi-unimodal with respect to the initial state. In particular, if the number of successes is finite almost surely, the quasi-unimodality of the reward sequence is sufficient for the myopic rule to be optimal. Illustrative examples are given in various last-success settings.
| 2025-12-26
| 2025-12-30
|
[
"math.PR"
] |
Zakaria Derbazi
|
2512.22362
|
On the number of words of $N=3 \,M$ letters with a three-letter alphabet
|
In this paper we address the well-known problem of counting the number of $3M$-letter words that can be formed from a three-letter alphabet by decomposing it into four possible cases based on its remainder when divided by three. The solution to the problem also gives us some sums of trinomial coefficients.
| 2025-12-26
| 2025-12-30
|
[
"math.CO"
] |
Pablo Serra
|
2512.21949
|
WST spectroscopic variability alerts: discovery space, data flow system requirements
|
The enormous multiplexity of the WST opens up the possibility to trigger alerts for variable objects - an option that has been reserved so far only for imaging surveys. WST can go further by detecting spectroscopic line profile and line strength variations. I review previous alert-issuing surveys that are limited to imaging, and describe some of the new research possibilities that this feature of the data flow system (DFS) would open up. The latter range from variability of emission line stars, such as Bes, WRs and LBVs to variability of active galaxies and quasars, including the so-called changing look objects that shift between Type 1 and Type 2. Furthermore, I describe the requirements that the WST DFS must meet to make this feasible. The most critical aspect is the rapid data processing for timely follow-up. Next, the alert system is tightly connected with the data reduction and archive, because it will need an extensive and continuously updated spectral reference database. The new spectra will have to be compared against these reference spectra to identify variations. The reference spectra can either be "native" from the WST itself or they can originate from other spectroscopic surveys. Two options for the DFS are considered: one is to conduct an automated search of the WST's own archive, and potentially of other spectroscopic archives and a second option is to allow the users to submit reference spectra on their own. The spectroscopic alert system will open up a completely new discovery space that is not accessible to the existing or planned near-future surveys. Finally, I discuss the advantages of moving the variability detection to physical parameters by modeling the observed and reference spectra and comparing the derived fitting parameters. This strategy offers a robust method for alert ranking.
| 2025-12-26
| 2025-12-29
|
[
"astro-ph.IM",
"physics.ins-det"
] |
Valenitn D. Ivanov
|
2507.12453
|
Cost-aware Stopping for Bayesian Optimization
|
In automated machine learning, scientific discovery, and other applications of Bayesian optimization, deciding when to stop evaluating expensive black-box functions in a cost-aware manner is an important but underexplored practical consideration. A natural performance metric for this purpose is the cost-adjusted simple regret, which captures the trade-off between solution quality and cumulative evaluation cost. While several heuristic or adaptive stopping rules have been proposed, they lack guarantees ensuring stopping before incurring excessive function evaluation costs. We propose a principled cost-aware stopping rule for Bayesian optimization that adapts to varying evaluation costs without heuristic tuning. Our rule is grounded in a theoretical connection to state-of-the-art cost-aware acquisition functions, namely the Pandora's Box Gittins Index (PBGI) and log expected improvement per cost (LogEIPC). We prove a theoretical guarantee bounding the expected cost-adjusted simple regret incurred by our stopping rule when paired with either acquisition function. Across synthetic and empirical tasks, including hyperparameter optimization and neural architecture size search, pairing our stopping rule with PBGI or LogEIPC usually matches or outperforms other acquisition-function--stopping-rule pairs in terms of cost-adjusted simple regret.
| 2025-12-26
| 2025-12-29
|
[
"cs.LG"
] |
Qian Xie, Linda Cai, Alexander Terenin, Peter I. Frazier, Ziv Scully
|
2512.23747
|
State-of-the-art Small Language Coder Model: Mify-Coder
|
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 frontier-grade models in code generation and agent-driven workflows. Our training pipeline combines high-quality curated sources with synthetic data generated through agentically designed prompts, refined iteratively using enterprise-grade evaluation datasets. LLM-based quality filtering further enhances data density, enabling frugal yet effective training. Through disciplined exploration of CPT-SFT objectives, data mixtures, and sampling dynamics, we deliver frontier-grade code intelligence within a single continuous training trajectory. Empirical evidence shows that principled data and compute discipline allow smaller models to achieve competitive accuracy, efficiency, and safety compliance. Quantized variants of Mify-Coder enable deployment on standard desktop environments without requiring specialized hardware.
| 2025-12-26
| 2026-01-01
|
[
"cs.SE",
"cs.AI",
"cs.CL"
] |
Abhinav Parmar, Abhisek Panigrahi, Abhishek Kumar Dwivedi, Abhishek Bhattacharya, Adarsh Ramachandra, Aditya Choudhary, Aditya Garg, Aditya Raj, Alankrit Bhatt, Alpesh Yadav, Anant Vishnu, Ananthu Pillai, Ankush Kumar, Aryan Patnaik, Aswatha Narayanan S, Avanish Raj Singh, Bhavya Shree Gadda, Brijesh Pankajbhai Kachhadiya, Buggala Jahnavi, Chidurala Nithin Krishna, Chintan Shah, Chunduru Akshaya, Debarshi Banerjee, Debrup Dey, Deepa R., Deepika B G, Faiz ur Rahman, Gagan Gayari, Gudhi Jagadeesh Kumar Naidu, Gursimar Singh, Harshal Tyagi, Harshini K, James Mani Vathalloor, Jayarama Nettar, Jayashree Gajjam, Joe Walter Sugil George, Kamalakara Sri Krishna Tadepalli, Kamalkumar Rathinasamy, Karan Chaurasia, Karthikeyan S, Kashish Arora, Kaushal Desai, Khushboo Buwade, Kiran Manjrekar, Malikireddy Venkata Sai Likhitha, Manjunath A, Mitali Mahavir Bedmutha, Mohammed Rafee Tarafdar, Nikhil Tiwari, Nikitha K Gigi, Pavan Ravikumar, Pendyala Swarnanjali, Piyush Anand, Prakash Chandrasekar, Prasanna Bhalchandra Gawade, Prasanth Sivan, Preeti Khurana, Priyanshi Babbar, Rajab Ali Mondal, Rajesh Kumar Vissapragada, Rajeshwari Ganesan, Rajeswari Koppisetti, Ramjee R., Ramkumar Thiruppathisamy, Rani G. S., S Reka, Samarth Gupta, Sandeep Reddy Kothakota, Sarathy K, Sathyanarayana Sampath Kumar, Saurabh Kumar, Shashank Khasare, Shenbaga Devi Venkatesh Kumar, Shiva Rama Krishna Parvatham, Shoeb Shaikh, Shrishanmathi A, Shubham Pathak, Sree Samhita Koppaka, Sreenivasa Raghavan K S, Sreeram Venkatasubramanian, Suprabha Desai Bojja, Swetha R, Syed Ahmed, Chinmai Harshitha Thota, Tushar Yadav, Veeravelly Kusumitha, V V S S Prasanth Patnaik, Vidya Sri Sesetti, Vijayakeerthi K, Vikram Raj Bakshi, Vinay K K, Vinoth Kumar Loganathan, Vipin Tiwari, Vivek Kumar Shrivastav, V Venkata Sri Datta Charan, Wasim Akhtar Khan
|
2512.20653
|
A gauge identity for interscale transfer in inhomogeneous turbulence
|
The local definition of interscale energy transfer is missing in inhomogeneous turbulence research. This manifests as a discrepancy between the subgrid-scale production $Î ^{\mathrm{SGS}}$ and the increment-based transfer density $Î ^{\mathrm{KHMH}}$. Here, this missing definition is found by identifying a gauge freedom in the spatial transport of energy, yielding the identity: $Î ^{\mathrm{SGS}} = \int G_\ell Î ^{\mathrm{KHMH}} \, d\boldsymbol{r} + \nabla \cdot \boldsymbol{J}_{\mathrm{gauge}}$. The formulations are proven to differ strictly by the divergence of the current $\boldsymbol{J}_{\mathrm{gauge}}$. Validation against the analytical Womersley solution confirms the identity to within machine precision ($<10^{-14}$). The current $\boldsymbol{J}_{\mathrm{gauge}}$ is identified as the mechanism for redistribution toward compliant boundaries. Both measures are shown to converge to the unique Duchon--Robert dissipation $D(u)$, unifying the theoretical framework for non-stationary turbulence.
| 2025-12-26
| 2025-12-29
|
[
"physics.flu-dyn"
] |
Khalid M. Saqr
|
2502.21022
|
When Unsupervised Domain Adaptation meets One-class Anomaly Detection: Addressing the Two-fold Unsupervised Curse by Leveraging Anomaly Scarcity
|
This paper introduces the first fully unsupervised domain adaptation (UDA) framework for unsupervised anomaly detection (UAD). The performance of UAD techniques degrades significantly in the presence of a domain shift, difficult to avoid in a real-world setting. While UDA has contributed to solving this issue in binary and multi-class classification, such a strategy is ill-posed in UAD. This might be explained by the unsupervised nature of the two tasks, namely, domain adaptation and anomaly detection. Herein, we first formulate this problem that we call the two-fold unsupervised curse. Then, we propose a pioneering solution to this curse, considered intractable so far, by assuming that anomalies are rare. Specifically, we leverage clustering techniques to identify a dominant cluster in the target feature space. Posed as the normal cluster, the latter is aligned with the source normal features. Concretely, given a one-class source set and an unlabeled target set composed mostly of normal data and some anomalies, we fit the source features within a hypersphere while jointly aligning them with the features of the dominant cluster from the target set. The paper provides extensive experiments and analysis on common adaptation benchmarks for anomaly detection, demonstrating the relevance of both the newly introduced paradigm and the proposed approach. The code will be made publicly available.
| 2025-12-26
| 2025-12-29
|
[
"cs.LG",
"cs.CV"
] |
Nesryne Mejri, Enjie Ghorbel, Anis Kacem, Pavel Chernakov, Niki Foteinopoulou, Djamila Aouada
|
2505.12717
|
ToTRL: Unlock LLM Tree-of-Thoughts Reasoning Potential through Puzzles Solving
|
Large language models (LLMs) demonstrate significant reasoning capabilities, particularly through long chain-of-thought (CoT) processes, which can be elicited by reinforcement learning (RL). However, prolonged CoT reasoning presents limitations, primarily verbose outputs due to excessive introspection. The reasoning process in these LLMs often appears to follow a trial-and-error methodology rather than a systematic, logical deduction. In contrast, tree-of-thoughts (ToT) offers a conceptually more advanced approach by modeling reasoning as an exploration within a tree structure. This reasoning structure facilitates the parallel generation and evaluation of multiple reasoning branches, allowing for the active identification, assessment, and pruning of unproductive paths. This process can potentially lead to improved performance and reduced token costs. Building upon the long CoT capability of LLMs, we introduce tree-of-thoughts RL (ToTRL), a novel on-policy RL framework with a rule-based reward. ToTRL is designed to guide LLMs in developing the parallel ToT strategy based on the sequential CoT strategy. Furthermore, we employ LLMs as players in a puzzle game during the ToTRL training process. Solving puzzle games inherently necessitates exploring interdependent choices and managing multiple constraints, which requires the construction and exploration of a thought tree, providing challenging tasks for cultivating the ToT reasoning capability. Our empirical evaluations demonstrate that our ToTQwen3-8B model, trained with our ToTRL, achieves significant improvement in performance and reasoning efficiency on complex reasoning tasks.
| 2025-12-26
| 2025-12-29
|
[
"cs.CL"
] |
Haoyuan Wu, Xueyi Chen, Rui Ming, Jilong Gao, Shoubo Hu, Zhuolun He, Bei Yu
|
2512.16038
|
LOG.io: Unified Rollback Recovery and Data Lineage Capture for Distributed Data Pipelines
|
This paper introduces LOG.io, a comprehensive solution designed for correct rollback recovery and fine-grain data lineage capture in distributed data pipelines. It is tailored for serverless scalable architectures and uses a log-based rollback recovery protocol. LOG.io supports a general programming model, accommodating non-deterministic operators, interactions with external systems, and arbitrary custom code. It is non-blocking, allowing failed operators to recover independently without interrupting other active operators, thereby leveraging data parallelization, and it facilitates dynamic scaling of operators during pipeline execution. Performance evaluations, conducted within the SAP Data Intelligence system, compare LOG.io with the Asynchronous Barrier Snapshotting (ABS) protocol, originally implemented in Flink. Our experiments show that when there are straggler operators in a data pipeline and the throughput of events is moderate (e.g., 1 event every 100 ms), LOG.io performs as well as ABS during normal processing and outperforms ABS during recovery. Otherwise, ABS performs better than LOG.io for both normal processing and recovery. However, we show that in these cases, data parallelization can largely reduce the overhead of LOG.io while ABS does not improve. Finally, we show that the overhead of data lineage capture, at the granularity of the event and between any two operators in a pipeline, is marginal, with less than 1.5% in all our experiments.
| 2025-12-26
| 2025-12-29
|
[
"cs.DC"
] |
Eric Simon, Renato B. Hoffmann, Lucas Alf, Dalvan Griebler
|
2512.06670
|
Local structure of the Hilbert scheme of conics in quintic del Pezzo varieties
|
Let $X$ be the quintic del Pezzo $4$-fold. It is very well-known that $X$ is realized by a smooth linear section of Grassmannian $\mathrm{Gr}(2,5)$. In this paper, we prove that the Hilbert scheme of conics in $X$ is a smooth variety of dimension $7$ by using a torus action on $X$, which provides a more direct proof about the first named author's previous result.
| 2025-12-26
| 2025-12-29
|
[
"math.AG"
] |
Kiryong Chung, Bomyeong Kim, Minseong Kwon
|
2512.21905
|
High-Fidelity and Long-Duration Human Image Animation with Diffusion Transformer
|
Recent progress in diffusion models has significantly advanced the field of human image animation. While existing methods can generate temporally consistent results for short or regular motions, significant challenges remain, particularly in generating long-duration videos. Furthermore, the synthesis of fine-grained facial and hand details remains under-explored, limiting the applicability of current approaches in real-world, high-quality applications. To address these limitations, we propose a diffusion transformer (DiT)-based framework which focuses on generating high-fidelity and long-duration human animation videos. First, we design a set of hybrid implicit guidance signals and a sharpness guidance factor, enabling our framework to additionally incorporate detailed facial and hand features as guidance. Next, we incorporate the time-aware position shift fusion module, modify the input format within the DiT backbone, and refer to this mechanism as the Position Shift Adaptive Module, which enables video generation of arbitrary length. Finally, we introduce a novel data augmentation strategy and a skeleton alignment model to reduce the impact of human shape variations across different identities. Experimental results demonstrate that our method outperforms existing state-of-the-art approaches, achieving superior performance in both high-fidelity and long-duration human image animation.
| 2025-12-26
| 2025-12-29
|
[
"cs.CV"
] |
Shen Zheng, Jiaran Cai, Yuansheng Guan, Shenneng Huang, Xingpei Ma, Junjie Cao, Hanfeng Zhao, Qiang Zhang, Shunsi Zhang, Xiao-Ping Zhang
|
2509.06384
|
On Bott-Chern and Aeppli cohomologies of two-dimensional toroidal groups
|
A toroidal group is a generalization of a complex torus, and is obtained as the quotient of the complex Euclidean space $\mathbb{C}^n$ by a discrete subgroup. Toroidal groups with finite-dimensional cohomology, called theta toroidal groups, are known to exhibit behavior analogous to that of complex tori. We compute Bott--Chern and Aeppli cohomologies for two-dimensional non-compact theta toroidal groups.
| 2025-12-26
| 2025-12-29
|
[
"math.CV"
] |
Jinichiro Tanaka
|
2512.22346
|
Higher Order Dualities between Prime Ideals
|
Extending the works of Alladi and Sweeting and Woo, we state and prove the general higher order duality between prime ideals in number rings. We then use the second order duality to obtain the a new formula for the Chebotarev Density involving sums of the generalized Möbius function and the prime ideal counting function. We also provide two estimates of such sums as an application of the duality identity. A discussion of the duality in a slightly more general setting is done at the end.
| 2025-12-26
| 2025-12-30
|
[
"math.NT"
] |
Sroyon Sengupta
|
2512.22309
|
LLMBoost: Make Large Language Models Stronger with Boosting
|
Ensemble learning of LLMs has emerged as a promising alternative to enhance performance, but existing approaches typically treat models as black boxes, combining the inputs or final outputs while overlooking the rich internal representations and interactions across models.In this work, we introduce LLMBoost, a novel ensemble fine-tuning framework that breaks this barrier by explicitly leveraging intermediate states of LLMs. Inspired by the boosting paradigm, LLMBoost incorporates three key innovations. First, a cross-model attention mechanism enables successor models to access and fuse hidden states from predecessors, facilitating hierarchical error correction and knowledge transfer. Second, a chain training paradigm progressively fine-tunes connected models with an error-suppression objective, ensuring that each model rectifies the mispredictions of its predecessor with minimal additional computation. Third, a near-parallel inference paradigm design pipelines hidden states across models layer by layer, achieving inference efficiency approaching single-model decoding. We further establish the theoretical foundations of LLMBoost, proving that sequential integration guarantees monotonic improvements under bounded correction assumptions. Extensive experiments on commonsense reasoning and arithmetic reasoning tasks demonstrate that LLMBoost consistently boosts accuracy while reducing inference latency.
| 2025-12-26
| 2025-12-30
|
[
"cs.LG",
"cs.AI"
] |
Zehao Chen, Tianxiang Ai, Yifei Li, Gongxun Li, Yuyang Wei, Wang Zhou, Guanghui Li, Bin Yu, Zhijun Chen, Hailong Sun, Fuzhen Zhuang, Jianxin Li, Deqing Wang, Yikun Ban
|
2508.04084
|
Convolutional autoencoders for the reconstruction of three-dimensional interfacial multiphase flows
|
We present a systematic investigation of convolutional autoencoders for the reduced-order representation of three-dimensional interfacial multiphase flows. Focusing on the reconstruction of phase indicators, we examine how the choice of interface representation, including sharp, diffuse, and level-set formulations, impacts reconstruction accuracy across a range of interface complexities. Training and validation are performed using both synthetic datasets with controlled geometric complexity and high-fidelity simulations of multiphase homogeneous isotropic turbulence. We show that the interface representation plays a critical role in autoencoder performance. Excessively sharp interfaces lead to the loss of small-scale features, while overly diffuse interfaces degrade overall accuracy. Across all datasets and metrics considered, a moderately diffuse interface provides the best balance between preserving fine-scale structures and achieving accurate reconstructions. These findings elucidate key limitations and best practices for dimensionality reduction of multiphase flows using autoencoders. By clarifying how interface representations interact with the inductive biases of convolutional neural networks, this work lays the foundation for decoupling the training of autoencoders for accurate state compression from the training of surrogate models for temporal forecasting or input-output prediction in latent space.
| 2025-12-26
| 2025-12-29
|
[
"cs.CE",
"cs.LG",
"physics.flu-dyn"
] |
Murray Cutforth, Shahab Mirjalili
|
2512.22299
|
Experimental study on the wall-pressure fluctuations of flow over an axisymmetric hull
|
Wall pressure fluctuations beneath the turbulent boundary layer of high-speed underwater vehicles are crucial for hydro-acoustics and acoustic stealth. However, a comprehensive understanding remains limited due to a lack of high-quality experimental data, particularly under realistic operational conditions. To address this gap, this study establishes the first high-fidelity experimental database of wall-pressure fluctuations on an axisymmetric hull at high Reynolds numbers. The dataset's primary innovation is its systematic inclusion of complex maneuvering (yaw and pitch) conditions, providing a benchmark for validating flow noise prediction models. Analysis of this dataset yields key physical insights. The study quantifies systematic Reynolds number effects, including a spectral energy shift toward lower frequencies, and spectral scaling laws by revealing the critical influence of pressure-gradient effects. These findings provide fundamental insights into non-equilibrium 3D turbulent flows and establish an essential dataset to support the design of quieter and more effective underwater vehicles.
| 2025-12-26
| 2025-12-30
|
[
"physics.flu-dyn"
] |
Peng Jiang, Haoyu Zhang, Yi Dai, Tao Peng, Bin Xie, Shijun Liao
|
2512.21953
|
Phase-Coherent D-MIMO ISAC: Multi-Target Estimation and Spectral Efficiency Trade-Offs
|
We investigate distributed multiple-input multiple-output (D-MIMO) integrated sensing and communication (ISAC) systems, in which multiple phase-synchronized access points (APs) jointly serve user equipments (UEs) while cooperatively detecting and estimating multiple static targets. To achieve high-accuracy multi-target estimation, we propose a two-stage sensing framework combining non-coherent and coherent maximum-likelihood (ML) estimation. In parallel, adaptive AP mode-selection strategies are introduced to balance communication and sensing performance: a communication-centric scheme that maximizes downlink spectral efficiency (SE) and a sensing-centric scheme that selects geometrically diverse receive APs to enhance sensing coverage. Simulation results confirm the SE-sensing trade-off, where appropriate power allocation between communication and sensing and larger array apertures alleviate performance degradation, achieving high SE with millimeter-level sensing precision. We further demonstrate that the proposed AP-selection strategy reveals an optimal number of receive APs that maximizes sensing coverage without significantly sacrificing SE.
| 2025-12-26
| 2025-12-29
|
[
"eess.SP"
] |
Venkatesh Tentu, Henk Wymeersch, Musa Furkan Keskin, Sauradeep Dey, Tommy Svensson
|
2511.14732
|
Quantum State Preparation with Resolution Refinement
|
We introduce a method called resolution refinement that allows one to bootstrap eigenstate preparation on a quantum computer. We first prepare an eigenstate of a low-resolution Hamiltonian using any method of choice. The eigenstate is then lifted to higher resolution and adiabatically evolved to produce the corresponding eigenstate of a higher-fidelity Hamiltonian. We give examples of resolution refinement applied to both single-particle basis states as well as a spatial lattice grid. For basis refinement, we compute few-body ground states of the Busch model for interacting particles in a harmonic trap in one dimension. For lattice refinement, we compute Hartree-Fock nuclear states for a central Woods-Saxon potential in three dimensions, and we compute bound states and continuum states in a multi-species Hubbard model of fermions in one dimension. In all cases, the method is efficient and requires an adiabatic evolution time that scales with the inverse of the energy gap times the square root of the system size. We show that this very favorable scaling arises from the fact that resolution refinement does not make large changes to the structure or energies of the low-energy eigenstates.
| 2025-12-26
| 2025-12-30
|
[
"quant-ph",
"hep-lat",
"nucl-th"
] |
Scott Bogner, Heiko Hergert, Morten Hjorth-Jensen, Ryan LaRose, Dean Lee, Matthew Patkowski
|
2512.21871
|
Bridging the Copyright Gap: Do Large Vision-Language Models Recognize and Respect Copyrighted Content?
|
Large vision-language models (LVLMs) have achieved remarkable advancements in multimodal reasoning tasks. However, their widespread accessibility raises critical concerns about potential copyright infringement. Will LVLMs accurately recognize and comply with copyright regulations when encountering copyrighted content (i.e., user input, retrieved documents) in the context? Failure to comply with copyright regulations may lead to serious legal and ethical consequences, particularly when LVLMs generate responses based on copyrighted materials (e.g., retrieved book experts, news reports). In this paper, we present a comprehensive evaluation of various LVLMs, examining how they handle copyrighted content -- such as book excerpts, news articles, music lyrics, and code documentation when they are presented as visual inputs. To systematically measure copyright compliance, we introduce a large-scale benchmark dataset comprising 50,000 multimodal query-content pairs designed to evaluate how effectively LVLMs handle queries that could lead to copyright infringement. Given that real-world copyrighted content may or may not include a copyright notice, the dataset includes query-content pairs in two distinct scenarios: with and without a copyright notice. For the former, we extensively cover four types of copyright notices to account for different cases. Our evaluation reveals that even state-of-the-art closed-source LVLMs exhibit significant deficiencies in recognizing and respecting the copyrighted content, even when presented with the copyright notice. To solve this limitation, we introduce a novel tool-augmented defense framework for copyright compliance, which reduces infringement risks in all scenarios. Our findings underscore the importance of developing copyright-aware LVLMs to ensure the responsible and lawful use of copyrighted content.
| 2025-12-26
| 2025-12-29
|
[
"cs.CL",
"cs.AI",
"cs.CR",
"cs.CY"
] |
Naen Xu, Jinghuai Zhang, Changjiang Li, Hengyu An, Chunyi Zhou, Jun Wang, Boyu Xu, Yuyuan Li, Tianyu Du, Shouling Ji
|
2512.18192
|
Multi-Part Object Representations via Graph Structures and Co-Part Discovery
|
Discovering object-centric representations from images can significantly enhance the robustness, sample efficiency and generalizability of vision models. Works on images with multi-part objects typically follow an implicit object representation approach, which fail to recognize these learned objects in occluded or out-of-distribution contexts. This is due to the assumption that object part-whole relations are implicitly encoded into the representations through indirect training objectives. We address this limitation by proposing a novel method that leverages on explicit graph representations for parts and present a co-part object discovery algorithm. We then introduce three benchmarks to evaluate the robustness of object-centric methods in recognizing multi-part objects within occluded and out-of-distribution settings. Experimental results on simulated, realistic, and real-world images show marked improvements in the quality of discovered objects compared to state-of-the-art methods, as well as the accurate recognition of multi-part objects in occluded and out-of-distribution contexts. We also show that the discovered object-centric representations can more accurately predict key object properties in a downstream task, highlighting the potential of our method to advance the field of object-centric representations.
| 2025-12-26
| 2025-12-29
|
[
"cs.CV"
] |
Alex Foo, Wynne Hsu, Mong Li Lee
|
2512.21963
|
Topological properties of generalized Markoff mod $p$ graphs
|
The generalized Markoff mod $p$ graph is defined via the equation $x^2+y^2+z^2=xyz+κ$ over the finite field $\mathbb{F}_p$ of prime order $p$. In this paper, we investigate the topological properties of the graph such as non-planarity, surface embeddability, and the existence of short cycles. Our approach is based on a systematic construction of $K_{3,3}$-subdivisions, integrating techniques from graph theory, computer algebra, and number theory.
| 2025-12-26
| 2025-12-29
|
[
"math.NT",
"math.CO"
] |
Shohei Satake, Yoshinori Yamasaki
|
2512.22403
|
Active Nonparametric Two-Sample Testing by Betting on Heterogeneous Data Sources
|
We study the problem of active nonparametric sequential two-sample testing over multiple heterogeneous data sources. In each time slot, a decision-maker adaptively selects one of $K$ data sources and receives a paired sample generated from that source for testing. The goal is to decide as quickly as possible whether the pairs are generated from the same distribution or not. The gain achieved by such adaptive sampling (in terms of smaller expected stopping time or larger error exponents) has been well-characterized for parametric models via Chernoff's adaptive MLE selection rule [1]. However, analogous results are not known for the case of nonparametric problems, such as two-sample testing, where we place no restrictions on the distributions.
Our main contribution is a general active nonparametric testing procedure that combines an adaptive source-selecting strategy within the testing-by-betting framework of [2] that works under minimal distributional assumptions. In each time slot, our scheme proceeds by selecting a source according to a probability that mixes exploitation, favoring sources with the largest empirical distinguishability, and exploration via a vanishing greedy strategy. The (paired) observations so collected are then used to update the "betting-wealth process", which is a stochastic process guaranteed to be a nonnegative martingale under the null. The procedure stops and rejects the null when the wealth process exceeds an appropriate threshold; an event that is unlikely under the null. We show that our test controls the type-I error at a prespecified level-$α$ under the null, and establish its power-one property and a bound on its expected sample size under the alternative. Our results provide a precise characterization of the improvements achievable by a principled adaptive sampling strategy over its passive analog.
| 2025-12-26
| 2025-12-30
|
[
"math.ST",
"cs.IT",
"math.IT",
"stat.TH"
] |
Chia-Yu Hsu, Shubhanshu Shekhar
|
2512.19337
|
Real-time propagators resummed with nontrivial boundary wavefunctions in a constant electric field
|
We present the derivation of an alternative representation of the real-time in-in formalism under a spatially homogeneous and time independent electric field. Because the system exhibits instability associated with pair production of particles and antiparticles, the perturbation theory should be reorganized depending on the choice of the reference vacuum. We recast the boundary wavefunctions into the quadratic self-energy-like terms in the functional integration formalism. The resulting generating functional in the modified in-in formalism leads to the propagators that resum infinite diagrams necessary to capture the vacuum-instability effects. The proper-time representations of the propagators reproduce the known expressions from the canonical operator formalism, but our derivation based on the generating functional along the closed-time path clarifies the origin of the additional proper-time contour and provides a better physical understanding. Finally, as a concrete example of the application, we compute the in-in expectation value of the vector current in a constant electric field, and find that the simple one-loop calculation captures the pair production effect.
| 2025-12-26
| 2025-12-29
|
[
"hep-ph",
"hep-th"
] |
Kenji Fukushima, Shuhei Minato
|
2505.04891
|
Clustering with Communication: A Variational Framework for Single Cell Representation Learning
|
Single-cell RNA sequencing (scRNA-seq) has revealed complex cellular heterogeneity, but recent studies emphasize that understanding biological function also requires modeling cell-cell communication (CCC), the signaling interactions mediated by ligand-receptor pairs that coordinate cellular behavior. Tools like CellChat have demonstrated that CCC plays a critical role in processes such as cell differentiation, tissue regeneration, and immune response, and that transcriptomic data inherently encodes rich information about intercellular signaling. We propose CCCVAE, a novel variational autoencoder framework that incorporates CCC signals into single-cell representation learning. By leveraging a communication-aware kernel derived from ligand-receptor interactions and a sparse Gaussian process, CCCVAE encodes biologically informed priors into the latent space. Unlike conventional VAEs that treat each cell independently, CCCVAE encourages latent embeddings to reflect both transcriptional similarity and intercellular signaling context. Empirical results across four scRNA-seq datasets show that CCCVAE improves clustering performance, achieving higher evaluation scores than standard VAE baselines. This work demonstrates the value of embedding biological priors into deep generative models for unsupervised single-cell analysis.
| 2025-12-26
| 2025-12-29
|
[
"cs.LG",
"cs.AI",
"stat.ML"
] |
Cong Qi, Yeqing Chen, Zhi Wei
|
2512.21836
|
Spectroscopic Characterization of Metallocene Single Crystals Grown by Physical Vapor Transport Method
|
High-quality metallocene single crystals with a low density of impurities and high homogeneity were prepared using the physical vapor transport method. These crystals were then characterized using various spectroscopic tools and X-ray diffraction. Laser-induced breakdown spectroscopy confirmed the presence of metal ions in each freshly grown sample despite all these crystals undergoing physical deformation with different lifetimes. X-ray diffraction analysis confirmed that all our metallocene single crystals retained a monoclinic structure at room temperature. The vibrational properties of our metallocene crystals were examined using Raman and Fourier-transform infrared spectroscopy. The inter- and intra-ring vibrational modes, along with additional modes associated with the crystalline form, were identified as inherent vibrational properties of our metallocene single crystals. Given the increasingly important role of metallocene in organic solar cells, organic light-emitting displays and molecular quantum systems, this research will enhance our understanding of the intrinsic physical properties of cleaner, more crystalline metallocene single crystals.
| 2025-12-26
| 2025-12-29
|
[
"cond-mat.mtrl-sci",
"cond-mat.soft"
] |
Ian B. Logue, Sandaruka Jayasooriya Arachchilage, Lance M. Griswold, Moses B. Gaither-Ganim, Lincoln W. Weber, Robyn Cook, Stephen Hofer, Praveena Satkunam, Dipanjan Mazumdar, Poopalasingam Sivakumar, Bumsu Lee
|
2501.06793
|
Differentially Private Gradient-Tracking-Based Distributed Stochastic Optimization over Directed Graphs
|
This paper proposes a differentially private gradient-tracking-based distributed stochastic optimization algorithm over directed graphs. In particular, privacy noises are incorporated into each agent's state and tracking variable to mitigate information leakage, after which the perturbed states and tracking variables are transmitted to neighbors. We design two novel schemes for the step-sizes and the sampling number within the algorithm. The sampling parameter-controlled subsampling method employed by both schemes enhances the differential privacy level, and ensures a finite cumulative privacy budget even over infinite iterations. The algorithm achieves both almost sure and mean square convergence for nonconvex objectives. Furthermore, when nonconvex objectives satisfy the Polyak-Lojasiewicz condition, Scheme (S1) achieves a polynomial mean square convergence rate, and Scheme (S2) achieves an exponential mean square convergence rate. The trade-off between privacy and convergence is presented. The effectiveness of the algorithm and its superior performance compared to existing works are illustrated through numerical examples of distributed training on the benchmark datasets "MNIST" and "CIFAR-10".
| 2025-12-26
| 2025-12-29
|
[
"eess.SY",
"cs.SY"
] |
Jialong Chen, Jimin Wang, Ji-Feng Zhang
|
2505.00920
|
Giant exciton binding energy in bulk CrCl3
|
Van der Waals (vdW) materials, with their unique combination of electronic, optical, and magnetic properties, are emerging as promising platforms for exploring excitonic phenomena. Thus far, the choice of materials with exceptional excitonic response has been limited to two-dimensional (2D) configurations of vdW materials. At the same time, large interlayer distance and the possibility to create a variety of heterostructures offers an opportunity to control the dielectric screening in van der Waals heterostructures and van der Waals 3D materials, thus engineering the excitonic properties. Here, we reveal that bulk vdW crystal CrCl3 answers this quest with a record exciton binding energy of 1.64 eV owing to a delicate interplay of quasi-2D electronic confinement and local magnetic correlations. We also suggest that the non-local magnetic correlations play an important role in the temperature dependence of photoluminescence intensity. Furthermore, we observe colossal binding energies in vdW crystals NbOCl2 (0.66 eV) and MoCl3 (0.35 eV) and formulate a universal exciton binding energy dependence on bandgap for 2D and 3D vdW materials. Hence, our findings establish a fundamental link between the layered structure of vdW materials and their excitonic properties.
| 2025-12-26
| 2025-12-29
|
[
"cond-mat.mtrl-sci",
"physics.optics"
] |
Georgy Ermolaev, Tagir Mazitov, Anton Minnekhanov, Arslan Mazitov, Gleb Tselikov, Aleksandr Slavich, Mikhail Tatmyshevskiy, Mikhail Kashchenko, Nikolay Pak, Andrey Vyshnevyy, Valentin Solovei, Ivan Kazantsev, Gleb Tikhonowski, Alexander Melentev, Elena Zhukova, Dmitriy Grudinin, Junhua Luo, Ivan Kruglov, Aleksey Arsenin, Sangen Zhao, Kostya S. Novoselov, Andrey Katanin, Valentyn S. Volkov
|
2512.22008
|
Rethinking photonic nanojets: a new definition and design paradigm
|
We propose a rigorous, physically interpretable, and quantifiable definition of the photonic nanojet (PNJ). This framework resolves longstanding ambiguities in measuring PNJ dimensions and leverages an optimal mass transport-based metric to quantify PNJ quality. Building on this metric, we develop a PNJ steering methodology that requires no opto-mechanical intervention, relying solely on phase-only illumination modulation.
| 2025-12-26
| 2025-12-29
|
[
"physics.optics",
"math-ph",
"math.MP"
] |
Mirza KaramehmedoviÄ, Kristoffer Linder-Steinlein, Jesper Glückstad
|
2512.21999
|
Look Closer! An Adversarial Parametric Editing Framework for Hallucination Mitigation in VLMs
|
While Vision-Language Models (VLMs) have garnered increasing attention in the AI community due to their promising practical applications, they exhibit persistent hallucination issues, generating outputs misaligned with visual inputs. Recent studies attribute these hallucinations to VLMs' over-reliance on linguistic priors and insufficient visual feature integration, proposing heuristic decoding calibration strategies to mitigate them. However, the non-trainable nature of these strategies inherently limits their optimization potential. To this end, we propose an adversarial parametric editing framework for Hallucination mitigation in VLMs, which follows an \textbf{A}ctivate-\textbf{L}ocate-\textbf{E}dit \textbf{A}dversarially paradigm. Specifically, we first construct an activation dataset that comprises grounded responses (positive samples attentively anchored in visual features) and hallucinatory responses (negative samples reflecting LLM prior bias and internal knowledge artifacts). Next, we identify critical hallucination-prone parameter clusters by analyzing differential hidden states of response pairs. Then, these clusters are fine-tuned using prompts injected with adversarial tuned prefixes that are optimized to maximize visual neglect, thereby forcing the model to prioritize visual evidence over inherent parametric biases. Evaluations on both generative and discriminative VLM tasks demonstrate the significant effectiveness of ALEAHallu in alleviating hallucinations. Our code is available at https://github.com/hujiayu1223/ALEAHallu.
| 2025-12-26
| 2025-12-29
|
[
"cs.CV",
"cs.LG"
] |
Jiayu Hu, Beibei Li, Jiangwei Xia, Yanjun Qin, Bing Ji, Zhongshi He
|
2512.21987
|
Optimal Placement of Data Centers to Support Power Distribution Networks Using Intelligent Algorithms with Economic Indicators
|
Data centers are among the fastest growing electricity consumers and can impose severe voltage drops and feeder losses when connected to weak distribution networks. This paper formulates a techno economic siting problem in which each candidate data center site is mapped to a bus of the distribution network and is assumed to deploy on site renewable generation and power electronic interfaces, resulting in a controllable net active power injection equivalent to distributed generation. A mixed integer nonlinear optimization model is developed to jointly select the connection bus and size the DG capacity while respecting network operating limits. The objective combines three normalized terms including active power losses, a voltage deviation index capturing profile quality, and investment cost derived from location dependent land price and unit DG cost. To address the discrete continuous search space, an intelligent genetic algorithm is embedded in a multi scenario decision framework with adaptive weight tuning. Three stakeholder scenarios prioritize losses, voltage quality, or techno economic balance, and additional balanced scenarios are generated automatically until the optimal bus decision converges. A case study on the IEEE 33 bus radial system demonstrates the effectiveness of the approach. The converged design selects bus 14 with 1.10 MW DG, reducing total losses from 202.67 kW to 129.37 kW while improving the minimum bus voltage to 0.933 per unit at a moderate investment cost of 1.33 MUSD. The proposed framework provides an interpretable pathway to integrate economic indicators into distribution aware data center siting.
| 2025-12-26
| 2025-12-29
|
[
"eess.SY",
"cs.SY"
] |
Amin Hajihasani, Mahmoud Modaresi
|
2512.22354
|
Water Production of Interstellar Comet 3I/ATLAS from SOHO/SWAN Observations after Perihelion
|
The Solar Wind ANisotropies (SWAN) all-sky hydrogen Lyman-alpha camera on the Solar and Heliosphere Observatory (SOHO) observed the hydrogen coma of interstellar comet 3I/ATLAS, also called C/2025 N1 (ATLAS), beginning on November 6, 2025, 9 days after perihelion. Water production rates were calculated from each image of 3I/ATLAS using the methodology of Mäkinen & Combi (2005, Icarus 177, 217) and fluorescence rates calculated using the daily solar Lyman-alpha fluxes from the LASP database corrected for solar rotation. The method has been used for over 90 comet apparitions (Combi 2022; Combi et al 2019). A water production rate of 3.17 x 10^29 s^-1 was found on November 6 when the comet was at a heliocentric distance of 1.40 au and at a sufficient solar elongation angle. It decreased over time after that, down to 1-2 x 10^28 s^-1 around 40 days post-perihelion (December 8).
| 2025-12-26
| 2025-12-30
|
[
"astro-ph.EP",
"astro-ph.GA"
] |
M. R. Combi, T. Mâkinen, J. -L. Bertaux, E. Quemerais, S. Ferron, R. Lallement, W. Schmidt
|
2305.07328
|
Enhance Multi-Scale Spatial-Temporal Coherence for Configurable Video Anomaly Detection
|
The development of unsupervised Video Anomaly Detection (VAD) relies on technologies in the field of signal processing. Since the anomaly is quite ambiguous and unbounded, different detection demands may often be raised even in one scenario. Thus, we propose to design the configurable VAD with flexible solutions targeting to solve the issue that previous methods have to train their models from scratch and waste resources when detection demands even change slightly. Moreover, we also design a dataset with good compatibility to evaluate the VAD performance when changes happen in detection demands. Besides, videos contain important information regarding continuous changes in the object's appearance and motion. Thus, we also propose a module to establish the multi-scale spatial-temporal coherence, which improves the accuracy and has the ability to dynamically adjust and accurately capture spatial-temporal normal patterns. Experiments show that our method not only models coherence effectively but also has better configurable ability.
| 2025-12-26
| 2025-12-30
|
[
"cs.CV"
] |
Kai Cheng, Xinzhe Li, Lijuan Che
|
2512.22391
|
Derived Gamma Geometry II: Stable $\infty$-Categories of Gamma-Modules, Derived Monoidal Structures, and Obstructions to Binary Shadows
|
Let \(\T\) be a commutative ternary \(\Gm\)-semiring in the sense of the triadic, \(\Gm\)-parametrized multiplication \(\{a,b,c\}_γ\). Building on the affine \(\Gm\)-spectrum \(\SpecG(\T)\), the structure sheaf, and the equivalence between \(\Gm\)-modules and quasi-coherent \(\Gm\)-sheaves on affine \(\Gm\)-schemes, we construct and organize the derived formalism at the level of stable \(\infty\)-categories.
Our first contribution is a technically explicit construction of a stable \(\infty\)-category \(\Dinfty(\T,\Gm)\) enhancing the unbounded derived category of \(\Gm\)-modules, obtained by dg-nerve and \(\infty\)-localization of chain complexes. We further explain the derived monoidal structure induced by the ternary \(\Gm\)-tensor product and the corresponding internal \(\RHom\), under standard exactness/projectivity hypotheses.
Our second contribution is an obstruction theory to \emph{binary reduction}: we formalize the nonexistence of any conservative ``binary module shadow'' compatible with the cubic localization calculus intrinsic to ternary \(\Gm\)-semirings. In particular, any attempt to represent the triadic \(\Gm\)-action by binary scalars forces \(\Gm\)-mode data to be absorbed into the scalars, hence ceases to be a genuine reduction.
Finally, we give a detailed affine derived equivalence between derived quasi-coherent \(\Gm\)-sheaves on \(X=\SpecG(\T)\) and \(\Dinfty(\T,\Gm)\), and we include worked examples illustrating the cubic localization relation and its derived consequences.
| 2025-12-26
| 2025-12-30
|
[
"math.RA"
] |
Chandrasekhar Gokavarapu
|
2205.04316
|
Gamow Temperature in Tsallis and Kaniadakis Statistics
|
Relying on the quantum tunnelling concept and Maxwell-Boltzmann-Gibbs statistics, Gamow shows that the star-burning process happens at temperatures comparable to a critical value, called the Gamow temperature ({\tt T}) and less than the prediction of the classical framework. In order to highlight the role of the equipartition theorem in the Gamow argument, a thermal length scale is defined, and then the effects of non-extensivity on the Gamow temperature have been investigated by focusing on the Tsallis and Kaniadakis statistics. The results attest that while the Gamow temperature decreases in the framework of Kaniadakis statistics, it can be bigger or smaller than {\tt T} when Tsallis statistics are employed.
| 2025-12-26
| 2025-12-29
|
[
"cond-mat.stat-mech"
] |
Hooman Moradpour, Mohsen Javaherian, Ebrahim Namvar, Amir Hadi Ziaie
|
2512.22034
|
Combinatorial characterzations of $T$-designs in the nonbinary Johnson scheme
|
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 designs in the nonbinary Hamming scheme. In this paper, we obtain a characterization that applies to a strictly larger class of index sets $T$, based on a methodological extension of Delsarte's original framework (1973). This new characterization naturally recovers classical block designs and orthogonal arrays as special cases. To describe these designs uniformly, we introduce $(r,s)$-designs, a new family of combinatorial objects that arise naturally from our characterization. We also derive absolute lower bounds on the cardinality of $(r,s)$-designs from the multiplicities of the primitive idempotents of the nonbinary Johnson scheme, and construct examples with index $λ=1$ that attain certain natural lower bounds.
| 2025-12-26
| 2025-12-29
|
[
"math.CO"
] |
Hiroshi Nozaki, Yuta Watanabe
|
2505.07982
|
Perfect state transfer on graphs with clusters
|
Using graphs with clusters, we provide a unified approach for constructing graphs with pair state transfer-relative to the adjacency, Laplacian, and signless Laplacian matrix-between the same pair of states at the same time, despite being non-regular. We show that for each $k\geq 5$, there are infinitely many connected graphs with maximum valency $k$ admitting this property. This framework also aids in establishing sufficient conditions for pair state transfer in edge-perturbed graphs, including complete graphs and complete bipartite graphs. Furthermore, we utilize graph products to generate new infinite families of graphs with the above property.
| 2025-12-26
| 2025-12-29
|
[
"math.CO",
"quant-ph"
] |
Hermie Monterde, Hiranmoy Pal
|
2512.22356
|
Clustering Confuses Spectro-photometry: An Investigation of 2D and 3D Forced Profile Matching for Stacking Line-intensity Mapping Data on Source Catalogues
|
Line-intensity mapping (LIM) is an emerging observational technique that is used to observe the universe on large scales at low resolution through spectral line emission. Stacking analyses coadd cutouts of LIM data on positions of known signal emitters, robustly detecting signal otherwise hidden in a noisy map. In this article, we present two augmentations of a stacking pipeline, both aiming to refine the sensitivity of the stack by assuming a specific observed signal shape in 2D spatial axes or 3D spatial and spectral axes, as well as stacking on source coordinates more precise than the resolution of the LIM data cube. We test these methods on a series of simplistic and complex simulations mimicking observations with the CO Mapping Array Project (COMAP) Pathfinder. We find that these fitting methods provide up to a 25% advantage in detection significance over the original stack method in realistic COMAP-like simulations. We also find that the optimal fitting profile, given our CO line model and galaxy model, is larger than the 5' width of the COMAP beam and takes on a Lorentzian shape in the spectral dimension. Our findings suggest a nuanced dependence of the optimal profile size and shape on the LIM signal itself, including redshift-space clustering and fingers-of-God effects that depend on the tracer luminosity function and bias.
| 2025-12-26
| 2025-12-30
|
[
"astro-ph.CO",
"astro-ph.IM"
] |
Ella M. Mansfield, Delaney A. Dunne, Dongwoo T. Chung
|
2512.22094
|
Gauge Coupling Unification in Gauge-Higgs GUT: Theory and Phenomenology
|
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 hierarchies within fermion masses and mixings, as well as provide a common origin of the gauge symmetries and the sector that spontaneously breaks them. At the same time, they furnish interesting LHC signatures in the form of TeV-scale resonances of the $X,Y$-like bosons, providing a trace of the grand-unified group, accessible at low energies. Using the method of Planck-brane correlators allows us to evolve the couplings consistently from the electroweak scale up to the Planck scale, avoiding shortcomings of other frequently-used approaches and including the effects of bulk scalars, fermions, and gauge-bosons within a common framework. We thereby revisit, contrast, and supplement results in the literature, the latter for example by including brane masses and the gauge-Higgs vacuum expectation value. Moreover, in a phenomenology section, we apply our results to the concrete case of Georgi-Glashow-like unification with a SU(6) $\supset$ SU(5) symmetry in the 5D bulk, presenting a quantitative survey of the quality of unification. We find that grand unification is possible in such models in the presence of moderately large brane kinetic terms.
| 2025-12-26
| 2025-12-29
|
[
"hep-ph"
] |
Andrei Angelescu, Andreas Bally, Florian Goertz, Sascha Weber
|
2512.21837
|
Knowledge Reasoning of Large Language Models Integrating Graph-Structured Information for Pest and Disease Control in Tobacco
|
This paper proposes a large language model (LLM) approach that integrates graph-structured information for knowledge reasoning in tobacco pest and disease control. Built upon the GraphRAG framework, the proposed method enhances knowledge retrieval and reasoning by explicitly incorporating structured information from a domain-specific knowledge graph. Specifically, LLMs are first leveraged to assist in the construction of a tobacco pest and disease knowledge graph, which organizes key entities such as diseases, symptoms, control methods, and their relationships. Based on this graph, relevant knowledge is retrieved and integrated into the reasoning process to support accurate answer generation. The Transformer architecture is adopted as the core inference model, while a graph neural network (GNN) is employed to learn expressive node representations that capture both local and global relational information within the knowledge graph. A ChatGLM-based model serves as the backbone LLM and is fine-tuned using LoRA to achieve parameter-efficient adaptation. Extensive experimental results demonstrate that the proposed approach consistently outperforms baseline methods across multiple evaluation metrics, significantly improving both the accuracy and depth of reasoning, particularly in complex multi-hop and comparative reasoning scenarios.
| 2025-12-26
| 2025-12-29
|
[
"cs.CL"
] |
Siyu Li, Chenwei Song, Wan Zhou, Xinyi Liu
|
2212.11580
|
A Theory of Conversion Relations for Prefixed Units of Measure
|
Units of measure with prefixes and conversion rules are given a formal semantic model in terms of categorial group theory. Basic structures and both natural and contingent semantic operations are defined. Conversion rules are represented as a class of ternary relations with both group-like and category-like properties. A hierarchy of subclasses is explored, each satisfying stronger useful algebraic properties than the preceding, culminating in a direct efficient conversion-by-rewriting algorithm.
| 2025-12-26
| 2025-12-31
|
[
"cs.PL",
"cs.DM"
] |
Baltasar Trancón y Widemann, Markus Lepper
|
2512.22389
|
Unveiling the CO2 Hydrate Phase Diagram from Computer Simulation: Locating the Hydrate-Liquid-Vapor Coexistence and its Upper Quadruple Point
|
Carbon dioxide (CO2) hydrates hold promising applications in capturing and separating CO2 for climate change mitigation. Understanding their behavior at the molecular level is therefore essential, and computer simulations have become powerful tools for exploring their formation and stability, providing valuable insights into their underlying mechanisms. In this work, we perform molecular dynamics simulations to compute the three-phase coexistence line involving the stability region where CO2 is in the vapor phase: CO2 hydrate - liquid water - vapor. This computation was previously inaccessible using the traditional three-phase direct coexistence technique. To achieve this, we employ a novel solubility-based method, which allows us to accurately evaluate the coexistence line. Finally, we have determined the upper quadruple point (Q2) where the four phases, namely hydrate, liquid water, liquid CO2, and vapor, coexist. Our pioneering result for the Q2 value shows remarkable agreement with experimental observations, validating the accuracy of our findings and representing a significant milestone in the field of gas hydrate research.
| 2025-12-26
| 2025-12-30
|
[
"cond-mat.soft"
] |
Jesús Algaba, Samuel Blazquez, Cristóbal Romero-Guzmán, Carlos Vega, MarÃa M. Conde, Felipe J. Blas
|
2512.22404
|
Mining the Gold: Student-AI Chat Logs as Rich Sources for Automated Knowledge Gap Detection
|
With the significant increase in enrollment in computing-related programs over the past 20 years, lecture sizes have grown correspondingly. In large lectures, instructors face challenges on identifying students' knowledge gaps timely, which is critical for effective teaching. Existing classroom response systems rely on instructor-initiated interactions, which limits their ability to capture the spontaneous knowledge gaps that naturally emerge during lectures. With the widespread adoption of LLMs among students, we recognize these student-AI dialogues as a valuable, student-centered data source for identifying knowledge gaps. In this idea paper, we propose QueryQuilt, a multi-agent LLM framework that automatically detects common knowledge gaps in large-scale lectures by analyzing students' chat logs with AI assistants. QueryQuilt consists of two key components: (1) a Dialogue Agent that responds to student questions while employing probing questions to reveal underlying knowledge gaps, and (2) a Knowledge Gap Identification Agent that systematically analyzes these dialogues to identify knowledge gaps across the student population. By generating frequency distributions of identified gaps, instructors can gain comprehensive insights into class-wide understanding. Our evaluation demonstrates promising results, with QueryQuilt achieving 100% accuracy in identifying knowledge gaps among simulated students and 95% completeness when tested on real student-AI dialogue data. These initial findings indicate the system's potential for facilitate teaching in authentic learning environments. We plan to deploy QueryQuilt in actual classroom settings for comprehensive evaluation, measuring its detection accuracy and impact on instruction.
| 2025-12-26
| 2025-12-30
|
[
"cs.HC",
"cs.SE"
] |
Quanzhi Fu, Qiyu Wu, Dan Williams
|
2511.07944
|
A Method for On-Orbit Calibration of the VLAST-P Electromagnetic Calorimeter
|
The Very Large Area Gamma-ray Space Telescope Pathfinder (VLAST-P), as the technology validation satellite for the VLAST mission, is designed to observe high-energy solar bursts on orbit. The CsI electromagnetic calorimeter (ECAL) is one of the key sub-detectors of VLAST-P. To investigate the on-orbit energy calibration method of the ECAL, a Geant4-based simulation of VLAST-P was carried out. The results show an energy resolution better than 10% in the 0.1 to 5 GeV range and a linearity deviation below 2%. A dedicated minimum-ionization-particle (MIP) calibration method was developed to ensure accurate energy reconstruction and to monitor detector stability throughout the in-orbit calibration period.
| 2025-12-26
| 2025-12-29
|
[
"hep-ex"
] |
Jiaxuan Wang, Zhen Wang, Borong Peng, Renjun Wang, Yunlong Zhang, Zhongtao Shen, Yifeng Wei, Dengyi Chen, Xiang Li, Yiming Hu, Jianhua Guo
|
2601.00837
|
Pediatric Pneumonia Detection from Chest X-Rays:A Comparative Study of Transfer Learning and Custom CNNs
|
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, evaluating frozen-backbone and fine-tuning regimes.
Methods: A dataset of 5,216 pediatric chest X-rays was split 80/10/10 for training, validation, and testing. Seven models were trained and assessed using accuracy, F1-score, and AUC. Grad-CAM visualizations provided explainability.
Results: Fine-tuned ResNet50 achieved the best performance: 99.43\% accuracy, 99.61\% F1-score, and 99.93\% AUC, with only 3 misclassifications. Fine-tuning outperformed frozen-backbone models by 5.5 percentage points on average. Grad-CAM confirmed clinically relevant lung regions guided predictions.
Conclusions: Transfer learning with fine-tuning substantially outperforms CNNs trained from scratch for pediatric pneumonia detection, showing near-perfect accuracy. This system has strong potential as a screening tool in resource-limited settings. Future work should validate these findings on multi-center and adult datasets.
Keywords: Pneumonia detection, deep learning, transfer learning, CNN, chest X-ray, pediatric diagnosis, ResNet, DenseNet, EfficientNet, Grad-CAM.
| 2025-12-26
| 2026-01-06
|
[
"cs.CV",
"cs.AI"
] |
Agniv Roy Choudhury
|
2512.21867
|
DPAR: Dynamic Patchification for Efficient Autoregressive Visual Generation
|
Decoder-only autoregressive image generation typically relies on fixed-length tokenization schemes whose token counts grow quadratically with resolution, substantially increasing the computational and memory demands of attention. We present DPAR, a novel decoder-only autoregressive model that dynamically aggregates image tokens into a variable number of patches for efficient image generation. Our work is the first to demonstrate that next-token prediction entropy from a lightweight and unsupervised autoregressive model provides a reliable criterion for merging tokens into larger patches based on information content. DPAR makes minimal modifications to the standard decoder architecture, ensuring compatibility with multimodal generation frameworks and allocating more compute to generation of high-information image regions. Further, we demonstrate that training with dynamically sized patches yields representations that are robust to patch boundaries, allowing DPAR to scale to larger patch sizes at inference. DPAR reduces token count by 1.81x and 2.06x on Imagenet 256 and 384 generation resolution respectively, leading to a reduction of up to 40% FLOPs in training costs. Further, our method exhibits faster convergence and improves FID by up to 27.1% relative to baseline models.
| 2025-12-26
| 2025-12-29
|
[
"cs.CV"
] |
Divyansh Srivastava, Akshay Mehra, Pranav Maneriker, Debopam Sanyal, Vishnu Raj, Vijay Kamarshi, Fan Du, Joshua Kimball
|
2507.03603
|
Selection bias effects on high-$p_\mathrm{T}$ yield and correlation measurements in Oxygen+Oxygen collisions
|
Oxygen+Oxygen (O+O) collisions at RHIC and the LHC offer a unique experimental opportunity to observe the onset of jet quenching in intermediate relativistic collision systems. As with the smaller proton-nucleus or larger nucleus-nucleus systems, measurements of centrality-selected high-$p_\mathrm{T}$ processes in O+O collisions are expected to be sensitive to selection bias effects, which will be necessary to quantify or mitigate before a definitive conclusion on the presence of jet quenching. Using two Monte Carlo heavy-ion event generators, we provide a survey of centrality bias effects on high-$p_\mathrm{T}$ yield and correlation measurements. Some highlights of our findings include that (1) bias factors for the accessible kinematic range at RHIC show a non-trivial $p_\mathrm{T}$ dependence, compared to a negligible one at the LHC given the smaller accessible Bjorken-$x$ range, (2) centrality definitions based on multiplicity are less sensitive to bias effects than those based on the transverse energy, (3) the Angantyr generator gives qualitatively similar but larger-magnitude bias factors than HIJING, and (4) correlation measurements have a much smaller sensitivity to bias effects than do yield measurements. The findings here are intended to guide the experimental design and interpretation of O+O jet quenching and other hard-process measurements.
| 2025-12-26
| 2025-12-29
|
[
"nucl-ex"
] |
JaeBeom Park, J. L. Nagle, Dennis V. Perepelitsa, Sanghoon Lim, Constantin Loizides
|
2512.22049
|
Quantum Secret Sharing Rates
|
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. Following the approach of Zou et al. (2015) on classical secret sharing, we introduce an information-theoretic model for the rate analysis of QSS and its relation to compound quantum channels. We establish a regularized characterization for the QSS capacity, and determine the capacity for QSS with dephasing noise.
| 2025-12-26
| 2025-12-29
|
[
"quant-ph",
"cs.IT",
"math.IT"
] |
Gabrielle Lalou, Husein Natur, Uzi Pereg
|
2512.21988
|
The Color-Clinical Decoupling: Why Perceptual Calibration Fails Clinical Biomarkers in Smartphone Dermatology
|
Smartphone-based tele-dermatology assumes that colorimetric calibration ensures clinical reliability, yet this remains untested for underrepresented skin phototypes. We investigated whether standard calibration translates to reliable clinical biomarkers using 43,425 images from 965 Korean subjects (Fitzpatrick III-IV) across DSLR, tablet, and smartphone devices. While Linear Color Correction Matrix (CCM) normalization reduced color error by 67-77% -- achieving near-clinical accuracy (Delta E < 2.3) -- this success did not translate to biomarker reliability.
We identify a phenomenon termed "color-clinical decoupling": despite perceptual accuracy, the Individual Typology Angle (ITA) showed poor inter-device agreement (ICC = 0.40), while the Melanin Index achieved good agreement (ICC = 0.77). This decoupling is driven by the ITA formula's sensitivity to b* channel noise and is further compounded by anatomical variance. Facial region accounts for 25.2% of color variance -- 3.6x greater than device effects (7.0%) -- challenging the efficacy of single-patch calibration. Our results demonstrate that current colorimetric standards are insufficient for clinical-grade biomarker extraction, necessitating region-aware protocols for mobile dermatology.
| 2025-12-26
| 2025-12-29
|
[
"eess.IV",
"cs.CV",
"q-bio.QM"
] |
Sungwoo Kang
|
2512.22009
|
iSHIFT: Lightweight Slow-Fast GUI Agent with Adaptive Perception
|
Multimodal Large Language Models (MLLMs) show strong potential for interpreting and interacting with complex, pixel-rich Graphical User Interface (GUI) environments. However, building agents that are both efficient for high-level tasks and precise for fine-grained interactions remains challenging. GUI agents must perform routine actions efficiently while also handling tasks that demand exact visual grounding, yet existing approaches struggle when accuracy depends on identifying specific interface elements. These MLLMs also remain large and cannot adapt their reasoning depth to the task at hand. In this work, we introduce iSHIFT: Implicit Slow-fast Hybrid Inference with Flexible Tokens, a lightweight agent that integrates latent thinking (implicit chain-of-thought) with a perception control module. iSHIFT enables an MLLM to switch between a slow mode, which leverages detailed visual grounding for high precision and a fast mode that uses global cues for efficiency. Special perception tokens guide attention to relevant screen regions, allowing the model to decide both how to reason and where to focus. Despite its compact 2.5B size, iSHIFT matches state-of-the-art performance on multiple benchmark datasets.
| 2025-12-26
| 2025-12-29
|
[
"cs.CV"
] |
Sarthak Mehrotra, Sairam V C Rebbapragada, Mani Hemanth Reddy Bonthu, Vineeth N Balasubramanian
|
2510.23483
|
Towards a Functionally Complete and Parameterizable TFHE Processor
|
Fully homomorphic encryption allows the evaluation of arbitrary functions on encrypted data. It can be leveraged to secure outsourced and multiparty computation. TFHE is a fast torus-based fully homomorphic encryption scheme that allows both linear operations, as well as the evaluation of arbitrary non-linear functions. It currently provides the fastest bootstrapping operation performance of any other FHE scheme. Despite its fast performance, TFHE suffers from a considerably higher computational overhead for the evaluation of homomorphic circuits. Computations in the encrypted domain are orders of magnitude slower than their unencrypted equivalents. This bottleneck hinders the widespread adoption of (T)FHE for the protection of sensitive data. While state-of-the-art implementations focused on accelerating and outsourcing single operations, their scalability and practicality are constrained by high memory bandwidth costs. In order to overcome this, we propose an FPGA-based hardware accelerator for the evaluation of homomorphic circuits. Specifically, we design a functionally complete TFHE processor for FPGA hardware capable of processing instructions on the data completely on the FPGA. In order to achieve a higher throughput from our TFHE processor, we implement an improved programmable bootstrapping module, which outperforms the current state-of-the-art by 240% to 480% more bootstrappings per second. Our efficient, compact, and scalable design lays the foundation for implementing complete FPGA-based TFHE processor architectures.
| 2025-12-26
| 2025-12-29
|
[
"cs.CR"
] |
Valentin Reyes Häusler, Gabriel Ott, Aruna Jayasena, Andreas Peter
|
2512.22052
|
Representing in Low Rank I: conjugacy, topological and homological aspects
|
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 $\mathcal{O}$ such that $\mathcal{O}$ has finitely many units, i.e. such that $\operatorname{SL}_2(\mathcal{O})$ has arithmetic rank $1$. In this first part, the focus is on two aspects.
One aspect concerns characterisations of such representing spaces in terms of Serre's homological goodness property, small virtual cohomological dimension and higher Kleinian-type embeddings. As an application, we obtain several characterisations of the finite groups $G$ whose irreducible representations are of the mentioned form. In particular, such groups $G$ are precisely those such that $\mathcal{U}(R G)$, with $R$ the ring of integers of $F$, can be constructed from groups which virtually map onto a non-abelian free group. Along the way we investigate the latter property for congruence subgroups of higher modular groups and its implications for the congruence kernel. This is used to obtain new information on the congruence kernel of the unit group of a group ring.
The second aspect concerns the conjugacy classes of the images of finite subgroups of $\mathcal{U}(R G)$ under the irreducible representations of $G$. More precisely, we initiate the study of a blockwise variant of the Zassenhaus conjectures and the subgroup isomorphism problem. Moreover, we contribute to them for the low rank representations above.
| 2025-12-26
| 2025-12-29
|
[
"math.RT",
"math.GR",
"math.RA"
] |
Robynn Corveleyn, Geoffrey Janssens, Doryan Temmerman
|
2511.13540
|
Fairness-Aware Graph Representation Learning with Limited Demographic Information
|
Ensuring fairness in Graph Neural Networks is fundamental to promoting trustworthy and socially responsible machine learning systems. In response, numerous fair graph learning methods have been proposed in recent years. However, most of them assume full access to demographic information, a requirement rarely met in practice due to privacy, legal, or regulatory restrictions. To this end, this paper introduces a novel fair graph learning framework that mitigates bias in graph learning under limited demographic information. Specifically, we propose a mechanism guided by partial demographic data to generate proxies for demographic information and design a strategy that enforces consistent node embeddings across demographic groups. In addition, we develop an adaptive confidence strategy that dynamically adjusts each node's contribution to fairness and utility based on prediction confidence. We further provide theoretical analysis demonstrating that our framework, FairGLite, achieves provable upper bounds on group fairness metrics, offering formal guarantees for bias mitigation. Through extensive experiments on multiple datasets and fair graph learning frameworks, we demonstrate the framework's effectiveness in both mitigating bias and maintaining model utility.
| 2025-12-26
| 2025-12-29
|
[
"cs.LG",
"cs.CY"
] |
Zichong Wang, Zhipeng Yin, Liping Yang, Jun Zhuang, Rui Yu, Qingzhao Kong, Wenbin Zhang
|
2512.22074
|
Semiperfect rings with a Nakayama permutation: A survey of Double annihilator property and Size condition
|
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 rings, respectively. However, their shared origins date back to the work of T. Nakayama.
We study these concepts through the lens of the Nakayama permutation, an invariant initially used to define (quasi-)Frobenius rings. We propose semiperfect rings as the setting for this study, treating them as the natural generalisation of finite rings, because they possess the characteristic decomposition of unity preserved by projection onto a semisimple top. This allows us to extend the utility of the Nakayama permutation beyond the classical Artinian setting.
By analysing the Nakayama permutation in this broader context, we show that many classical properties of (quasi-)Frobenius rings are not exclusive to the finite case, but are special cases of the general behaviour of semiperfect rings with essential socles.
We illustrate these results using B. J. Müller's representation of semiperfect rings as rings of formal matrices. The clear description of socles and tops in this setting provides a straightforward method for constructing counterexamples, such as quasi-Frobenius rings that are not Frobenius.
| 2025-12-26
| 2025-12-29
|
[
"math.RA"
] |
Dominik Krasula
|
2512.21839
|
Fano compactifications of mutation algebras
|
In this article, we introduce the notion of mutation semigroup algebras. This concept simultaneously generalizes cluster algebras and semigroup algebras. We show that, under some mild conditions on the singularities, the spectrum $U={\rm Spec}(R)$ of a mutation semigroup algebra $R$ admits a log Fano compactification $U\hookrightarrow X$. The compactification $X$ can be chosen to be a $\mathbb{Q}$-factorial log Fano variety whenever $U$ is $\mathbb{Q}$-factorial. Furthermore, we prove that a $\mathbb{Q}$-factorial klt Fano variety $X$ is of cluster type if and only if its Cox ring ${\rm Cox}(X)$ is a ${\rm Cl}(X)$-graded mutation semigroup algebra. In order to enlighten the previous theorems, we provide several explicit examples motivated by birational geometry, representation theory, and combinatorics.
| 2025-12-26
| 2025-12-29
|
[
"math.AG",
"math.CO",
"math.RT"
] |
Joshua Enwright, Luca Francone, JoaquÃn Moraga, Hunter Spink
|
2204.01210
|
Co-Teaching for Unsupervised Domain Adaptation and Expansion
|
Unsupervised Domain Adaptation (UDA) essentially trades a model's performance on a source domain for improving its performance on a target domain. To overcome this, Unsupervised Domain Expansion (UDE) has been introduced, which adapts the model to the target domain while preserving its performance in the source domain. In both UDA and UDE, a model tailored to a given domain is assumed to well handle samples from the given domain. We question the assumption by reporting the existence of cross-domain visual ambiguity: Due to the unclear boundary between the two domains, samples from one domain can be visually close to the other domain. Such sorts of samples are typically in the minority in their host domain, so they tend to be overlooked by the domain-specific model, but can be better handled by a model from the other domain. We exploit this finding by proposing Co-Teaching (CT), which is instantiated with knowledge distillation based CT (kdCT) plus mixup based CT (miCT). Specifically, kdCT leverages a dual-teacher architecture to enhance the student network's ability to handle cross-domain ambiguity. Meanwhile, miCT further enhances the generalization ability of the student. Extensive experiments on image classification and driving-scene segmentation show the viability of CT for UDE.
| 2025-12-26
| 2025-12-29
|
[
"cs.CV"
] |
Hailan Lin, Qijie Wei, Kaibin Tian, Ruixiang Zhao, Xirong Li
|
2509.01971
|
Maps from knots to $2$-links, chord diagrams, and a way to enhance Vassiliev invariants
|
In the present paper, we discuss a way of generalising Vassiliev knot invariants and weight systems to framed chord diagrams having framing 0 and 1.
| 2025-12-26
| 2025-12-29
|
[
"math.GT"
] |
Vassily Olegovich Manturov
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.