1. FLUX-Reason-6M & PRISM-Bench: A Million-Scale Text-to-Image Reasoning Dataset and Comprehensive Benchmark
Authors: Rongyao Fang, Aldrich Yu, Chengqi Duan, Linjiang Huang, Shuai Bai, Yuxuan Cai, Kun Wang, Si Liu, Xihui Liu, Hongsheng Li β’
Published: 2025-09-11 β’
Source: arXiv
The advancement of open-source text-to-image (T2I) models has been hindered by the absence of large-scale, reasoning-focused datasets and comprehensive evaluation benchmarks, resulting in a performance gap compared to leading closed-source systems. To address this challenge, We introduce FLUX-Reason-6M and PRISM-Bench (Precise and Robust Image Synthesis Measurement Benchmark). FLUX-Reason-6M is a massive dataset consisting of 6 million high-quality FLUX-generated images and 20 million bilingual (English and Chinese) descriptions specifically designed to teach complex reasoning. The image are organized according to six key characteristics: Imagination, Entity, Text rendering, Style, Affection, and Composition, and design explicit Generation Chain-of-Thought (GCoT) to provide detailed breakdowns of image generation steps. The whole data curation takes 15,000 A100 GPU days, providing the community with a resource previously unattainable outside of large industrial labs. PRISM-Bench offers a novel evaluation standard with seven distinct tracks, including a formidable Long Text challenge using GCoT. Through carefully designed prompts, it utilizes advanced vision-language models for nuanced human-aligned assessment of prompt-image alignment and image aesthetics. Our extensive evaluation of 19 leading models on PRISM-Bench reveals critical performance gaps and highlights specific areas requiring improvement. Our dataset, benchmark, and evaluation code are released to catalyze the next wave of reasoning-oriented T2I generation. Project page: https://flux-reason-6m.github.io/ .
2. ButterflyQuant: Ultra-low-bit LLM Quantization through Learnable Orthogonal Butterfly Transforms
Authors: Bingxin Xu, Zhen Dong, Oussama Elachqar, Yuzhang Shang β’
Published: 2025-09-11 β’
Source: arXiv
Large language models require massive memory footprints, severely limiting deployment on consumer hardware. Quantization reduces memory through lower numerical precision, but extreme 2-bit quantization suffers from catastrophic performance loss due to outliers in activations. Rotation-based methods such as QuIP and QuaRot apply orthogonal transforms to eliminate outliers before quantization, using computational invariance: $\mathbf{y} = \mathbf{Wx} = (\mathbf{WQ}^T)(\mathbf{Qx})$ for orthogonal $\mathbf{Q}$. However, these methods use fixed transforms--Hadamard matrices achieving optimal worst-case coherence $\mu = 1/\sqrt{n}$--that cannot adapt to specific weight distributions. We identify that different transformer layers exhibit distinct outlier patterns, motivating layer-adaptive rotations rather than one-size-fits-all approaches. We propose ButterflyQuant, which replaces Hadamard rotations with learnable butterfly transforms parameterized by continuous Givens rotation angles. Unlike Hadamard's discrete $\{+1, -1\}$ entries that are non-differentiable and prohibit gradient-based learning, butterfly transforms' continuous parameterization enables smooth optimization while guaranteeing orthogonality by construction. This orthogonal constraint ensures theoretical guarantees in outlier suppression while achieving $O(n \log n)$ computational complexity with only $\frac{n \log n}{2}$ learnable parameters. We further introduce a uniformity regularization on post-transformation activations to promote smoother distributions amenable to quantization. Learning requires only 128 calibration samples and converges in minutes on a single GPU--a negligible one-time cost. On LLaMA-2-7B with 2-bit quantization, ButterflyQuant achieves 15.4 perplexity versus 22.1 for QuaRot.
3. Cosmic $Ο$ensions Indirectly Correlate with Reionization Optical Depth
Authors: Itamar J. Allali, Lingfeng Li, Praniti Singh, JiJi Fan β’
Published: 2025-09-11 β’
Source: arXiv
The reionization optical depth $\tau_{\rm reio}$ has interesting connections to existing cosmological anomalies. As first studied in the context of the Hubble tension in our previous paper, a larger $\tau_{\rm reio}$, which could be achieved by removing the Planck low-$\ell$ polarization data, could boost $H_0$ slightly, resulting in a mild reduction of the tension between the early- and late-universe determinations of $H_0$. It has been shown later that a larger $\tau_{\rm reio}$ could also relieve other anomalies including: the tension between BAO and CMB data, the neutrino mass tension, and the latest DESI plus supernovae data's tension with the standard cosmological constant scenario. In this paper, we systematically analyze the correlations between $\tau_{\rm reio}$ and relevant cosmological parameters in the existing cosmic observation anomalies. In addition to Pearson correlation coefficients extracted directly from the covariance matrix, we also study partial correlation coefficients which measure intrinsic relationships between pairs of parameters removing the influence of other parameters. We show that $\tau_{\rm reio}$ has weak intrinsic correlations with the parameters responsible for the tensions and anomalies discussed. The large direct Pearson correlations that allow larger $\tau_{\rm reio}$ inferences to alleviate the cosmological tensions each arise from complicated networks through multiple parameters. As a result, the relationships between $\tau_{\rm reio}$ and each anomaly are not independent of each other. We also employ our method of computing correlations to clarify the impact of large scale polarization data, and comment also on the effects of CMB observations from ACT and SPT.
4. SpatialVID: A Large-Scale Video Dataset with Spatial Annotations
Authors: Jiahao Wang, Yufeng Yuan, Rujie Zheng, Youtian Lin, Jian Gao, Lin-Zhuo Chen, Yajie Bao, Yi Zhang, Chang Zeng, Yanxi Zhou, Xiaoxiao Long, Hao Zhu, Zhaoxiang Zhang, Xun Cao, Yao Yao β’
Published: 2025-09-11 β’
Source: arXiv
Significant progress has been made in spatial intelligence, spanning both spatial reconstruction and world exploration. However, the scalability and real-world fidelity of current models remain severely constrained by the scarcity of large-scale, high-quality training data. While several datasets provide camera pose information, they are typically limited in scale, diversity, and annotation richness, particularly for real-world dynamic scenes with ground-truth camera motion. To this end, we collect \textbf{SpatialVID}, a dataset consists of a large corpus of in-the-wild videos with diverse scenes, camera movements and dense 3D annotations such as per-frame camera poses, depth, and motion instructions. Specifically, we collect more than 21,000 hours of raw video, and process them into 2.7 million clips through a hierarchical filtering pipeline, totaling 7,089 hours of dynamic content. A subsequent annotation pipeline enriches these clips with detailed spatial and semantic information, including camera poses, depth maps, dynamic masks, structured captions, and serialized motion instructions. Analysis of SpatialVID's data statistics reveals a richness and diversity that directly foster improved model generalization and performance, establishing it as a key asset for the video and 3D vision research community.
5. Cosmology inference with perturbative forward modeling at the field level: a comparison with joint power spectrum and bispectrum analyses
Authors: Kazuyuki Akitsu, Marko SimonoviΔ, Shi-Fan Chen, Giovanni Cabass, Matias Zaldarriaga β’
Published: 2025-09-11 β’
Source: arXiv
We extend field-level inference to jointly constrain the cosmological parameters $\{A,\omega_{\rm cdm},H_0\}$, in both real and redshift space. Our analyses are based on mock data generated using a perturbative forward model, with noise drawn from a Gaussian distribution with a constant power spectrum. This idealized setting, where the field-level likelihood is exactly Gaussian, allows us to precisely quantify the information content in the nonlinear field on large scales. We find that field-level inference accurately recovers all cosmological parameters in both real and redshift space, with uncertainties consistent with perturbation theory expectations. We show that these error bars are comparable to those obtained from a joint power spectrum and bispectrum analysis using the same perturbative model. Finally, we perform several tests using the Gaussian field-level likelihood to fit the mock data where the true noise model is non-Gaussian, and find significant biases in the inferred cosmological parameters. These results highlight that the success of field-level inference critically depends on using the correct likelihood, which may be the primary challenge for applying this method to smaller scales even in the perturbative regime.
6. Dexplore: Scalable Neural Control for Dexterous Manipulation from Reference-Scoped Exploration
Authors: Sirui Xu, Yu-Wei Chao, Liuyu Bian, Arsalan Mousavian, Yu-Xiong Wang, Liang-Yan Gui, Wei Yang β’
Published: 2025-09-11 β’
Source: arXiv
Hand-object motion-capture (MoCap) repositories offer large-scale, contact-rich demonstrations and hold promise for scaling dexterous robotic manipulation. Yet demonstration inaccuracies and embodiment gaps between human and robot hands limit the straightforward use of these data. Existing methods adopt a three-stage workflow, including retargeting, tracking, and residual correction, which often leaves demonstrations underused and compound errors across stages. We introduce Dexplore, a unified single-loop optimization that jointly performs retargeting and tracking to learn robot control policies directly from MoCap at scale. Rather than treating demonstrations as ground truth, we use them as soft guidance. From raw trajectories, we derive adaptive spatial scopes, and train with reinforcement learning to keep the policy in-scope while minimizing control effort and accomplishing the task. This unified formulation preserves demonstration intent, enables robot-specific strategies to emerge, improves robustness to noise, and scales to large demonstration corpora. We distill the scaled tracking policy into a vision-based, skill-conditioned generative controller that encodes diverse manipulation skills in a rich latent representation, supporting generalization across objects and real-world deployment. Taken together, these contributions position Dexplore as a principled bridge that transforms imperfect demonstrations into effective training signals for dexterous manipulation.
7. Measuring Epistemic Humility in Multimodal Large Language Models
Authors: Bingkui Tong, Jiaer Xia, Sifeng Shang, Kaiyang Zhou β’
Published: 2025-09-11 β’
Source: arXiv
Hallucinations in multimodal large language models (MLLMs) -- where the model generates content inconsistent with the input image -- pose significant risks in real-world applications, from misinformation in visual question answering to unsafe errors in decision-making. Existing benchmarks primarily test recognition accuracy, i.e., evaluating whether models can select the correct answer among distractors. This overlooks an equally critical capability for trustworthy AI: recognizing when none of the provided options are correct, a behavior reflecting epistemic humility. We present HumbleBench, a new hallucination benchmark designed to evaluate MLLMs' ability to reject plausible but incorrect answers across three hallucination types: object, relation, and attribute. Built from a panoptic scene graph dataset, we leverage fine-grained scene graph annotations to extract ground-truth entities and relations, and prompt GPT-4-Turbo to generate multiple-choice questions, followed by a rigorous manual filtering process. Each question includes a "None of the above" option, requiring models not only to recognize correct visual information but also to identify when no provided answer is valid. We evaluate a variety of state-of-the-art MLLMs -- including both general-purpose and specialized reasoning models -- on HumbleBench and share valuable findings and insights with the community. By incorporating explicit false-option rejection, HumbleBench fills a key gap in current evaluation suites, providing a more realistic measure of MLLM reliability in safety-critical settings. Our code and dataset are released publicly and can be accessed at https://github.com/maifoundations/HumbleBench.
8. Explaining the Reputational Risks of AI-Mediated Communication: Messages Labeled as AI-Assisted Are Viewed as Less Diagnostic of the Sender's Moral Character
Authors: Pranav Khadpe, Kimi Wenzel, George Loewenstein, Geoff Kaufman β’
Published: 2025-09-11 β’
Source: arXiv
When someone sends us a thoughtful message, we naturally form judgments about their character. But what happens when that message carries a label indicating it was written with the help of AI? This paper investigates how the appearance of AI assistance affects our perceptions of message senders. Adding nuance to previous research, through two studies (N=399) featuring vignette scenarios, we find that AI-assistance labels don't necessarily make people view senders negatively. Rather, they dampen the strength of character signals in communication. We show that when someone sends a warmth-signalling message (like thanking or apologizing) without AI help, people more strongly categorize the sender as warm. At the same time, when someone sends a coldness-signalling message (like bragging or blaming) without assistance, people more confidently categorize them as cold. Interestingly, AI labels weaken both these associations: An AI-assisted apology makes the sender appear less warm than if they had written it themselves, and an AI-assisted blame makes the sender appear less cold than if they had composed it independently. This supports our signal diagnosticity explanation: messages labeled as AI-assisted are viewed as less diagnostic than messages which seem unassisted. We discuss how our findings shed light on the causal origins of previously reported observations in AI-Mediated Communication.
9. Large-scale variability in macroturbulence driven by pulsations in the rapidly rotating massive star Zeta Oph from high-cadence ESPRESSO spectroscopy and TESS photometry
Authors: A. J. Kalita, D. M. Bowman, M. Abdul-Masih, S. SimΓ³n-DΓaz β’
Published: 2025-09-11 β’
Source: arXiv
Despite their importance, the dynamical properties of massive stars remain poorly understood. Rotation is a key driver of internal mixing and angular momentum transport, significantly influencing massive star evolution. However, constraining rotation from spectroscopy is challenging, as spectral lines often exhibit excess broadening beyond rotation. The origin of this additional broadening, typically attributed to large-scale velocity fields and commonly referred to as macroturbulence, remains uncertain and unconstrained. Here, we present the combined analysis of TESS photometry and rapid time-series spectroscopy using the high-resolution ESPRESSO instrument at the Very Large Telescope of the European Southern Observatory for the rapidly rotating and pulsating massive star Zeta Ophiuchi. Leveraging excellent temporal coverage, our analysis demonstrates that pulsation-induced variability leads to peak-to-peak scatter as large as 88 km/s in the observed macroturbulent velocity time series. We also demonstrate that time-averaged macroturbulent velocities are spectral line specific and can exceed 100 km/s . Furthermore, the macroturbulent velocities from shorter integration times are systematically lower than those derived from stacked spectra mimicking longer exposure times typically needed for fainter stars. These results highlight the role of pulsations in driving variable macroturbulence in massive stars, while also pointing to a potential bias in spectroscopic estimates of macroturbulence for fainter massive stars.
10. DiFlow-TTS: Discrete Flow Matching with Factorized Speech Tokens for Low-Latency Zero-Shot Text-To-Speech
Authors: Ngoc-Son Nguyen, Hieu-Nghia Huynh-Nguyen, Thanh V. T. Tran, Truong-Son Hy, Van Nguyen β’
Published: 2025-09-11 β’
Source: arXiv
Zero-shot Text-to-Speech (TTS) aims to synthesize high-quality speech that mimics the voice of an unseen speaker using only a short reference sample, requiring not only speaker adaptation but also accurate modeling of prosodic attributes. Recent approaches based on language models, diffusion, and flow matching have shown promising results in zero-shot TTS, but still suffer from slow inference and repetition artifacts. Discrete codec representations have been widely adopted for speech synthesis, and recent works have begun to explore diffusion models in purely discrete settings, suggesting the potential of discrete generative modeling for speech synthesis. However, existing flow-matching methods typically embed these discrete tokens into a continuous space and apply continuous flow matching, which may not fully leverage the advantages of discrete representations. To address these challenges, we introduce DiFlow-TTS, which, to the best of our knowledge, is the first model to explore purely Discrete Flow Matching for speech synthesis. DiFlow-TTS explicitly models factorized speech attributes within a compact and unified architecture. It leverages in-context learning by conditioning on textual content, along with prosodic and acoustic attributes extracted from a reference speech, enabling effective attribute cloning in a zero-shot setting. In addition, the model employs a factorized flow prediction mechanism with distinct heads for prosody and acoustic details, allowing it to learn aspect-specific distributions. Experimental results demonstrate that DiFlow-TTS achieves promising performance in several key metrics, including naturalness, prosody, preservation of speaker style, and energy control. It also maintains a compact model size and achieves low-latency inference, generating speech up to 25.8 times faster than the latest existing baselines.
11. Bridging the Capability Gap: Joint Alignment Tuning for Harmonizing LLM-based Multi-Agent Systems
Authors: Minghang Zhu, Zhengliang Shi, Zhiwei Xu, Shiguang Wu, Lingjie Wang, Pengjie Ren, Zhaochun Ren, Zhumin Chen β’
Published: 2025-09-11 β’
Source: arXiv
The advancement of large language models (LLMs) has enabled the construction of multi-agent systems to solve complex tasks by dividing responsibilities among specialized agents, such as a planning agent for subgoal generation and a grounding agent for executing tool-use actions. Most existing methods typically fine-tune these agents independently, leading to capability gaps among them with poor coordination. To address this, we propose MOAT, a Multi-Agent Joint Alignment Tuning framework that improves agents collaboration through iterative alignment. MOAT alternates between two key stages: (1) Planning Agent Alignment, which optimizes the planning agent to generate subgoal sequences that better guide the grounding agent; and (2) Grounding Agent Improving, which fine-tunes the grounding agent using diverse subgoal-action pairs generated by the agent itself to enhance its generalization capablity. Theoretical analysis proves that MOAT ensures a non-decreasing and progressively convergent training process. Experiments across six benchmarks demonstrate that MOAT outperforms state-of-the-art baselines, achieving average improvements of 3.1% on held-in tasks and 4.4% on held-out tasks.
12. Arctic Oscillation Modulation of Winter Air-Sea Coupling in the East/Japan Sea: Persistence, Timescales, and Extremes
Authors: Gyuchang Lim, JongJin Park β’
Published: 2025-09-11 β’
Source: arXiv
The winter climate of the East/Japan Sea (EJS) is strongly affected by the Arctic Oscillation (AO), yet how AO polarity reshapes the memory, coupling patterns, and predictability of sea-surface temperature anomalies (SSTA) remains poorly quantified. Using 30 winters (1993--2022) of daily OISST and ERA5 fields, we combine multivariate Maximum Covariance Analysis (MCA) with an Ornstein--Uhlenbeck (OU)-like integration of atmospheric principal components (PCs). The leading coupled mode explains 87% (+AO) and 75% (-AO) of squared covariance, with SSTA hot spots in East Korea Bay and along the subpolar front. Zero-lag correlations between the SSTA PC and OU-integrated atmospheric PCs reveal characteristic memory timescales ($\tau$) of $\sim$18--25 days for wind-stress curl (CurlTau), $\sim$15--30 days for near-surface air temperature (ATMP) and zonal winds, and $\sim$30--50 days for sea-level pressure (SLP) and meridional winds -- longer under -AO. Detrended Fluctuation Analysis (DFA) shows SSTA persistence $H \approx 1.3$--$1.4$ and that integrated atmospheric responses acquire ocean-like persistence, validating Hasselmann's stochastic framework for winter EJS. AO-phase contrasts align with a curl$\rightarrow$Ekman pumping$\rightarrow$eddy/SSH$\rightarrow$SST pathway: +AO favors anticyclonic/downwelling responses and warmer SSTA, whereas -AO favors cyclonic/upwelling and cooler SSTA. These diagnostics identify phase-specific predictor windows (e.g., 3-week OU-integrated CurlTau/ATMP; 4--7-week SLP/V-wind under -AO) to initialize subseasonal extremes prediction (marine heatwaves and cold-surge-impacted SST). The approach quantifies memory scales and spatial coupling that were not explicitly resolved by previous composite analyses, offering a tractable foundation for probabilistic forecast models.
13. Self-dual monopole loops, instantons and confinement
Authors: Mendel Nguyen, Mithat Γnsal β’
Published: 2025-09-11 β’
Source: arXiv
It is well-known that the standard instanton analysis in 4d Yang-Mills is plagued with the instanton size moduli problem, which renders the instanton contribution to vacuum energy density (or one-instanton partition function) infrared divergent. The formalism also ignores the implications of long range (magnetic dipole type) $1/r^4$ interaction between the small instantons, since it is weaker than Coulomb interaction. We show that in $U(1)$ lattice gauge theory, where finite action configurations are monopole loops, small loops at large separations also interact with the same type of $1/r^4$ interaction. If one ignores the classical interactions between monopoles, following the same idea as in Yang-Mills theory, the one-monopole partition function is also infrared divergent at strong coupling. However, $1/r^4$ interactions among small loops should be viewed as a consequence of multipole expansion, and emanate from $1/r^2$ interaction between current segments. Taking interactions into account, one can prove that the strongly coupled $U(1)$ lattice gauge theory is dual to a lattice abelian Higgs model, and more importantly, free of infrared divergences. The model exhibits mass gap and confinement by monopole condensation. We suggest that the structure of moduli space of instantons, ADHM data, and the long ranged classical interactions in pure Yang-Mills theory should be examined with this refined perspective. We conjecture that, in contradistinction to the current views on the subject, internal structure of instantons in Yang-Mills theory is responsible for confinement in $4d$ , similar to sigma model in $d=2$ dimensions.
14. Functional Groups are All you Need for Chemically Interpretable Molecular Property Prediction
Authors: Roshan Balaji, Joe Bobby, Nirav Pravinbhai Bhatt β’
Published: 2025-09-11 β’
Source: arXiv
Molecular property prediction using deep learning (DL) models has accelerated drug and materials discovery, but the resulting DL models often lack interpretability, hindering their adoption by chemists. This work proposes developing molecule representations using the concept of Functional Groups (FG) in chemistry. We introduce the Functional Group Representation (FGR) framework, a novel approach to encoding molecules based on their fundamental chemical substructures. Our method integrates two types of functional groups: those curated from established chemical knowledge (FG), and those mined from a large molecular corpus using sequential pattern mining (MFG). The resulting FGR framework encodes molecules into a lower-dimensional latent space by leveraging pre-training on a large dataset of unlabeled molecules. Furthermore, the proposed framework allows the inclusion of 2D structure-based descriptors of molecules. We demonstrate that the FGR framework achieves state-of-the-art performance on a diverse range of 33 benchmark datasets spanning physical chemistry, biophysics, quantum mechanics, biological activity, and pharmacokinetics while enabling chemical interpretability. Crucially, the model's representations are intrinsically aligned with established chemical principles, allowing chemists to directly link predicted properties to specific functional groups and facilitating novel insights into structure-property relationships. Our work presents a significant step toward developing high-performing, chemically interpretable DL models for molecular discovery.
15. Explaining Concept Drift through the Evolution of Group Counterfactuals
Authors: Ignacy StΔpka, Jerzy Stefanowski β’
Published: 2025-09-11 β’
Source: arXiv
Machine learning models in dynamic environments often suffer from concept drift, where changes in the data distribution degrade performance. While detecting this drift is a well-studied topic, explaining how and why the model's decision-making logic changes still remains a significant challenge. In this paper, we introduce a novel methodology to explain concept drift by analyzing the temporal evolution of group-based counterfactual explanations (GCEs). Our approach tracks shifts in the GCEs' cluster centroids and their associated counterfactual action vectors before and after a drift. These evolving GCEs act as an interpretable proxy, revealing structural changes in the model's decision boundary and its underlying rationale. We operationalize this analysis within a three-layer framework that synergistically combines insights from the data layer (distributional shifts), the model layer (prediction disagreement), and our proposed explanation layer. We show that such holistic view allows for a more comprehensive diagnosis of drift, making it possible to distinguish between different root causes, such as a spatial data shift versus a re-labeling of concepts.
16. Constraints on Ultra-heavy DM from TeV-PeV gamma-ray diffuse measurements
Authors: Manuel Rocamora, Pedro De La Torre Luque, Miguel A. SΓ‘nchez-Conde β’
Published: 2025-09-11 β’
Source: arXiv
Recent experiments have measured the Galactic $\gamma$-ray diffuse emission up to PeV energies, opening a window to study acceleration of Galactic cosmic rays and their propagation up to the cosmic-ray knee. Furthermore, these observations provide a powerful tool to set strong constraints into very-heavy dark matter particles, with masses in the TeV-PeV range. In this paper, we explore the potential of the newest observations of diffuse emissions at the Galactic plane from HAWC and LHAASO to probe this kind of dark matter over a wide mass range. Here, we model secondary emissions (inverse-Compton) from the electrons and positrons produced in the annihilation/decay of dark matter, on top of their prompt $\gamma$-ray emission, including the effects of absorption of high-energy photons via pair production. Furthermore, we show that including the astrophysical backgrounds (namely diffuse emission from cosmic-ray collisions or emission from unresolved sources) can significantly improve these limits. We find that the new measurements provided, specially by LHAASO with the combination of the WCDA and KM2A detectors, allow us to set strong constraints in decaying dark matter, being competitive and even improving the strongest constraints at the moment. We also highlight that these regions lead to constraints that are less affected by uncertainties from the dark matter distribution and discuss how CTA north and SWGO will be able to improve limits in this mass range.
17. A Multi-Scale Feature Extraction and Fusion UNet for Pathloss Prediction in UAV-Assisted mmWave Radio Networks
Authors: Sajjad Hussain β’
Published: 2025-09-11 β’
Source: arXiv
Accurate pathloss prediction is essential for the design and optimization of UAV-assisted millimeter-wave (mmWave) networks. While deep learning approaches have shown strong potential, their generalization across diverse environments, robustness to noisy inputs, and sensitivity to UAV altitude remain underexplored. To address these challenges, we propose a UNet-based deep learning architecture that combines multi-scale feature extraction, convolution-based feature fusion, and an atrous spatial pyramid pooling (ASPP) bottleneck for efficient context aggregation. The model predicts pathloss maps from log-distance, line-of-sight (LOS) mask, and building mask inputs. In addition, we develop a fully vectorized LOS mask computation algorithm that significantly accelerates pre-processing and enables large-scale dataset generation. Extensive evaluations on both in-house ray-tracing data and the RadioMapSeer benchmark demonstrate that the proposed model outperforms several state-of-the-art baselines in accuracy and efficiency. All source code is publicly released to support reproducibility and future research.
18. Multiwavelength observations of a new black-widow millisecond pulsar PSR J1544-2555
Authors: Sergio Belmonte Diaz, Tinn Thingmeearkom, Adipol Phosrisom, Rene Breton, Marta Burgay, Colin Clark, Lars Nieder, Martin Mayer, Werner Becker, Ewann Barr, Sarah Buchner, Kaustav Kashyap Das, Vik Dhillon, Oliver Dodge, Elizabeth Ferrara, Jean-Mathias Griessmeier, Ramesh Karuppusamy, Mark Kennedy, Michael Kramer, Prajwal Padmanabh, John Paice, Antonio Rodriguez, Ben Stappers β’
Published: 2025-09-11 β’
Source: arXiv
We report the discovery of a new black-widow millisecond pulsar, PSR J1544-2555, associated with the Fermi-LAT source 4FGL J1544.2-2554. Optical, radio, and gamma-ray observations confirmed its nature as a compact spider binary system. Optical photometry from ULTRACAM revealed a \(\sim\)2.7-hour orbital period, guiding MeerKAT observations that detected \(\sim\)2.4-ms radio pulsations. Subsequent timing campaigns using the Murriyang Parkes Telescope, the Effelsberg 100-m Radio Telescope, and the Nan\c{c}ay Radio Telescope allowed us to obtain a preliminary timing solution, which enabled us to find gamma-ray pulsations. The final timing solution, spanning 16 years of Fermi-LAT gamma-ray data, also displays orbital period variations typical of spider pulsars. X-ray observations from eROSITA indicate non-thermal emission, but the relatively low count rate prohibits the search for X-ray pulsations. Optical light curve modelling using Icarus suggests the asymmetry is best explained by a spot model, where uneven heating creates localised temperature variations on the companion. While the optical spectra we obtained are compatible with the physical properties we infer for the companion star, they were not of sufficient signal-to-noise to allow for radial velocity measurements, thus limiting constraints on the neutron star's mass. The observed bluer colour near the light curve minimum suggests possible non-thermal emission from intra-binary shocks, supported by the presence of an X-ray source. This discovery exemplifies the proven capability of the Fermi-LAT catalogue in identifying millisecond pulsar candidates and highlights the role of optical surveys in detecting variable sources suitable for radio follow-up.
19. Fault-tolerant transformations of spacetime codes
Authors: Arthur Pesah, Austin K. Daniel, Ilan Tzitrin, Michael Vasmer β’
Published: 2025-09-11 β’
Source: arXiv
Recent advances in quantum error-correction (QEC) have shown that it is often beneficial to understand fault-tolerance as a dynamical process, a circuit with redundant measurements that help correct errors, rather than as a static code equipped with a syndrome extraction circuit. Spacetime codes have emerged as a natural framework to understand error correction at the circuit level while leveraging the traditional QEC toolbox. Here, we introduce a framework based on chain complexes and chain maps to model spacetime codes and transformations between them. We show that stabilizer codes, quantum circuits, and decoding problems can all be described using chain complexes, and that the equivalence of two spacetime codes can be characterized by specific maps between chain complexes, the fault-tolerant maps, that preserve the number of encoded qubits, fault distance, and minimum-weight decoding problem. As an application of this framework, we extend the foliated cluster state construction from stabilizer codes to any spacetime code, showing that any Clifford circuit can be transformed into a measurement-based protocol with the same fault-tolerant properties. To this protocol, we associate a chain complex which encodes the underlying decoding problem, generalizing previous cluster state complex constructions. Our method enables the construction of cluster states from non-CSS, subsystem, and Floquet codes, as well as from logical Clifford operations on a given code.
20. Graph Alignment via Dual-Pass Spectral Encoding and Latent Space Communication
Authors: Maysam Behmanesh, Erkan Turan, Maks Ovsjanikov β’
Published: 2025-09-11 β’
Source: arXiv
Graph alignment-the problem of identifying corresponding nodes across multiple graphs-is fundamental to numerous applications. Most existing unsupervised methods embed node features into latent representations to enable cross-graph comparison without ground-truth correspondences. However, these methods suffer from two critical limitations: the degradation of node distinctiveness due to oversmoothing in GNN-based embeddings, and the misalignment of latent spaces across graphs caused by structural noise, feature heterogeneity, and training instability, ultimately leading to unreliable node correspondences. We propose a novel graph alignment framework that simultaneously enhances node distinctiveness and enforces geometric consistency across latent spaces. Our approach introduces a dual-pass encoder that combines low-pass and high-pass spectral filters to generate embeddings that are both structure-aware and highly discriminative. To address latent space misalignment, we incorporate a geometry-aware functional map module that learns bijective and isometric transformations between graph embeddings, ensuring consistent geometric relationships across different representations. Extensive experiments on graph benchmarks demonstrate that our method consistently outperforms existing unsupervised alignment baselines, exhibiting superior robustness to structural inconsistencies and challenging alignment scenarios. Additionally, comprehensive evaluation on vision-language benchmarks using diverse pretrained models shows that our framework effectively generalizes beyond graph domains, enabling unsupervised alignment of vision and language representations.
21. Fluent but Unfeeling: The Emotional Blind Spots of Language Models
Authors: Bangzhao Shu, Isha Joshi, Melissa Karnaze, Anh C. Pham, Ishita Kakkar, Sindhu Kothe, Arpine Hovasapian, Mai ElSherief β’
Published: 2025-09-11 β’
Source: arXiv
The versatility of Large Language Models (LLMs) in natural language understanding has made them increasingly popular in mental health research. While many studies explore LLMs' capabilities in emotion recognition, a critical gap remains in evaluating whether LLMs align with human emotions at a fine-grained level. Existing research typically focuses on classifying emotions into predefined, limited categories, overlooking more nuanced expressions. To address this gap, we introduce EXPRESS, a benchmark dataset curated from Reddit communities featuring 251 fine-grained, self-disclosed emotion labels. Our comprehensive evaluation framework examines predicted emotion terms and decomposes them into eight basic emotions using established emotion theories, enabling a fine-grained comparison. Systematic testing of prevalent LLMs under various prompt settings reveals that accurately predicting emotions that align with human self-disclosed emotions remains challenging. Qualitative analysis further shows that while certain LLMs generate emotion terms consistent with established emotion theories and definitions, they sometimes fail to capture contextual cues as effectively as human self-disclosures. These findings highlight the limitations of LLMs in fine-grained emotion alignment and offer insights for future research aimed at enhancing their contextual understanding.
22. Bridging the Gap in Phishing Detection: A Comprehensive Phishing Dataset Collector
Authors: Aditya Kulkarni, Shahil Manishbhai Patel, Shivam Pradip Tirmare, Vivek Balachandran, Tamal Das β’
Published: 2025-09-11 β’
Source: arXiv
To combat phishing attacks -- aimed at luring web users to divulge their sensitive information -- various phishing detection approaches have been proposed. As attackers focus on devising new tactics to bypass existing detection solutions, researchers have adapted by integrating machine learning and deep learning into phishing detection. Phishing dataset collection is vital to developing effective phishing detection approaches, which highly depend on the diversity of the gathered datasets. The lack of diversity in the dataset results in a biased model. Since phishing websites are often short-lived, collecting them is also a challenge. Consequently, very few phishing webpage dataset repositories exist to date. No single repository comprehensively consolidates all phishing elements corresponding to a phishing webpage, namely, URL, webpage source code, screenshot, and related webpage resources. This paper introduces a resource collection tool designed to gather various resources associated with a URL, such as CSS, Javascript, favicons, webpage images, and screenshots. Our tool leverages PhishTank as the primary source for obtaining active phishing URLs. Our tool fetches several additional webpage resources compared to PyWebCopy Python library, which provides webpage content for a given URL. Additionally, we share a sample dataset generated using our tool comprising 4,056 legitimate and 5,666 phishing URLs along with their associated resources. We also remark on the top correlated phishing features with their associated class label found in our dataset. Our tool offers a comprehensive resource set that can aid researchers in developing effective phishing detection approaches.
23. Numerical modelling of a partially loaded intermodal container freight train passing through a tunnel
Authors: Zhen Liu, David Soper, Hassan Hemida, Boyang Chen β’
Published: 2025-09-11 β’
Source: arXiv
The bluff nature of a freight train locomotive, coupled with large gaps created between different wagon formations and loaded goods, influence the overall pressure wave pattern generated as the train passes through a tunnel. Typically, 1D models are used to predict the patterns and properties of tunnel pressure wave formations. However, accurate modelling of regions of separation at the head of the blunted containers and at unloaded gap sections is essential for precise predictions of pressure magnitudes. This has traditionally been difficult to capture with 1D models. Furthermore, achieving this accuracy through 3D computational methods demands exceptional mesh quality, significant computational resources, and the careful selection of numerical models. This paper evaluates various numerical models to capture these complexities within regions of flow separation. Findings have supported the development of a new 1D programme to calculate the pressure wave generated by a freight locomotive entering a tunnel, and is here further extended to consider the discontinuities of the train body created by intermodal container loading patterns, by implementing new mesh system and boundary conditions into the 1D programme. A parameterisation study for different loading configurations is also presented to improve the overall programme adaptability, and the relationship between predetermined parameters and gap length is investigated. We validate the effectiveness of the improved 1D model through comprehensive Large Eddy Simulation (LES) results and conduct an extensive parameterisation study to enhance its applicability across various loading configurations. Consequently, this research bridges the gap in freight train tunnel aerodynamics, offering a versatile 1D numerical tool for accurate pressure wave prediction.
24. Visual Grounding from Event Cameras
Authors: Lingdong Kong, Dongyue Lu, Ao Liang, Rong Li, Yuhao Dong, Tianshuai Hu, Lai Xing Ng, Wei Tsang Ooi, Benoit R. Cottereau β’
Published: 2025-09-11 β’
Source: arXiv
Event cameras capture changes in brightness with microsecond precision and remain reliable under motion blur and challenging illumination, offering clear advantages for modeling highly dynamic scenes. Yet, their integration with natural language understanding has received little attention, leaving a gap in multimodal perception. To address this, we introduce Talk2Event, the first large-scale benchmark for language-driven object grounding using event data. Built on real-world driving scenarios, Talk2Event comprises 5,567 scenes, 13,458 annotated objects, and more than 30,000 carefully validated referring expressions. Each expression is enriched with four structured attributes -- appearance, status, relation to the viewer, and relation to surrounding objects -- that explicitly capture spatial, temporal, and relational cues. This attribute-centric design supports interpretable and compositional grounding, enabling analysis that moves beyond simple object recognition to contextual reasoning in dynamic environments. We envision Talk2Event as a foundation for advancing multimodal and temporally-aware perception, with applications spanning robotics, human-AI interaction, and so on.
25. Mechanism Design with Outliers and Predictions
Authors: Argyrios Deligkas, Eduard Eiben, Sophie Klumper, Guido SchΓ€fer, Artem Tsikiridis β’
Published: 2025-09-11 β’
Source: arXiv
We initiate the study of mechanism design with outliers, where the designer can discard $z$ agents from the social cost objective. This setting is particularly relevant when some agents exhibit extreme or atypical preferences. As a natural case study, we consider facility location on the line: $n$ strategic agents report their preferred locations, and a mechanism places a facility to minimize a social cost function. In our setting, the $z$ agents farthest from the chosen facility are excluded from the social cost. While it may seem intuitive that discarding outliers improves efficiency, our results reveal that the opposite can hold. We derive tight bounds for deterministic strategyproof mechanisms under the two most-studied objectives: utilitarian and egalitarian social cost. Our results offer a comprehensive view of the impact of outliers. We first show that when $z \ge n/2$, no strategyproof mechanism can achieve a bounded approximation for either objective. For egalitarian cost, selecting the $(z + 1)$-th order statistic is strategyproof and 2-approximate. In fact, we show that this is best possible by providing a matching lower bound. Notably, this lower bound of 2 persists even when the mechanism has access to a prediction of the optimal location, in stark contrast to the setting without outliers. For utilitarian cost, we show that strategyproof mechanisms cannot effectively exploit outliers, leading to the counterintuitive outcome that approximation guarantees worsen as the number of outliers increases. However, in this case, access to a prediction allows us to design a strategyproof mechanism achieving the best possible trade-off between consistency and robustness. Finally, we also establish lower bounds for randomized mechanisms that are truthful in expectation.
26. Boosting Embodied AI Agents through Perception-Generation Disaggregation and Asynchronous Pipeline Execution
Authors: Shulai Zhang, Ao Xu, Quan Chen, Han Zhao, Weihao Cui, Ningxin Zheng, Haibin Lin, Xin Liu, Minyi Guo β’
Published: 2025-09-11 β’
Source: arXiv
Embodied AI systems operate in dynamic environments, requiring seamless integration of perception and generation modules to process high-frequency input and output demands. Traditional sequential computation patterns, while effective in ensuring accuracy, face significant limitations in achieving the necessary "thinking" frequency for real-world applications. In this work, we present Auras, an algorithm-system co-designed inference framework to optimize the inference frequency of embodied AI agents. Auras disaggregates the perception and generation and provides controlled pipeline parallelism for them to achieve high and stable throughput. Faced with the data staleness problem that appears when the parallelism is increased, Auras establishes a public context for perception and generation to share, thereby promising the accuracy of embodied agents. Experimental results show that Auras improves throughput by 2.54x on average while achieving 102.7% of the original accuracy, demonstrating its efficacy in overcoming the constraints of sequential computation and providing high throughput.
27. Exploring the magnetic landscape of easily-exfoliable two-dimensional materials
Authors: Fatemeh Haddadi, Davide Campi, Flaviano dos Santos, Nicolas Mounet, Louis Ponet, Nicola Marzari, Marco Gibertini β’
Published: 2025-09-11 β’
Source: arXiv
Magnetic materials often exhibit complex energy landscapes with multiple local minima, each corresponding to a self-consistent electronic structure solution. Finding the global minimum is challenging, and heuristic methods are not always guaranteed to succeed. Here, we apply a recently developed automated workflow to systematically explore the energy landscape of 194 magnetic monolayers obtained from the Materials Cloud 2D crystals database and determine their ground-state magnetic order. Our approach enables effective control and sampling of orbital occupation matrices, allowing rapid identification of local minima. We find a diverse set of self-consistent collinear metastable states, further enriched by Hubbard-corrected energy functionals, when the $U$ parameters have been computed from first principles using linear-response theory. We categorise the monolayers by their magnetic ordering and highlight promising candidates. Our results include 109 ferromagnetic, 83 antiferromagnetic, and 2 altermagnetic monolayers, along with 12 novel ferromagnetic half-metals with potential for spintronics technologies.
28. Incorporating AI Incident Reporting into Telecommunications Law and Policy: Insights from India
Authors: Avinash Agarwal, Manisha J. Nene β’
Published: 2025-09-11 β’
Source: arXiv
The integration of artificial intelligence (AI) into telecommunications infrastructure introduces novel risks, such as algorithmic bias and unpredictable system behavior, that fall outside the scope of traditional cybersecurity and data protection frameworks. This paper introduces a precise definition and a detailed typology of telecommunications AI incidents, establishing them as a distinct category of risk that extends beyond conventional cybersecurity and data protection breaches. It argues for their recognition as a distinct regulatory concern. Using India as a case study for jurisdictions that lack a horizontal AI law, the paper analyzes the country's key digital regulations. The analysis reveals that India's existing legal instruments, including the Telecommunications Act, 2023, the CERT-In Rules, and the Digital Personal Data Protection Act, 2023, focus on cybersecurity and data breaches, creating a significant regulatory gap for AI-specific operational incidents, such as performance degradation and algorithmic bias. The paper also examines structural barriers to disclosure and the limitations of existing AI incident repositories. Based on these findings, the paper proposes targeted policy recommendations centered on integrating AI incident reporting into India's existing telecom governance. Key proposals include mandating reporting for high-risk AI failures, designating an existing government body as a nodal agency to manage incident data, and developing standardized reporting frameworks. These recommendations aim to enhance regulatory clarity and strengthen long-term resilience, offering a pragmatic and replicable blueprint for other nations seeking to govern AI risks within their existing sectoral frameworks.
29. SEDM: Scalable Self-Evolving Distributed Memory for Agents
Authors: Haoran Xu, Jiacong Hu, Ke Zhang, Lei Yu, Yuxin Tang, Xinyuan Song, Yiqun Duan, Lynn Ai, Bill Shi β’
Published: 2025-09-11 β’
Source: arXiv
Long-term multi-agent systems inevitably generate vast amounts of trajectories and historical interactions, which makes efficient memory management essential for both performance and scalability. Existing methods typically depend on vector retrieval and hierarchical storage, yet they are prone to noise accumulation, uncontrolled memory expansion, and limited generalization across domains. To address these challenges, we present SEDM, Self-Evolving Distributed Memory, a verifiable and adaptive framework that transforms memory from a passive repository into an active, self-optimizing component. SEDM integrates verifiable write admission based on reproducible replay, a self-scheduling memory controller that dynamically ranks and consolidates entries according to empirical utility, and cross-domain knowledge diffusion that abstracts reusable insights to support transfer across heterogeneous tasks. Evaluations on benchmark datasets demonstrate that SEDM improves reasoning accuracy while reducing token overhead compared with strong memory baselines, and further enables knowledge distilled from fact verification to enhance multi-hop reasoning. The results highlight SEDM as a scalable and sustainable memory mechanism for open-ended multi-agent collaboration. The code will be released in the later stage of this project.
30. In-Loop Filtering Using Learned Look-Up Tables for Video Coding
Authors: Zhuoyuan Li, Jiacheng Li, Yao Li, Jialin Li, Li Li, Dong Liu, Feng Wu β’
Published: 2025-09-11 β’
Source: arXiv
In-loop filtering (ILF) is a key technology in video coding standards to reduce artifacts and enhance visual quality. Recently, neural network-based ILF schemes have achieved remarkable coding gains, emerging as a powerful candidate for next-generation video coding standards. However, the use of deep neural networks (DNN) brings significant computational and time complexity or high demands for dedicated hardware, making it challenging for general use. To address this limitation, we study a practical ILF solution by adopting look-up tables (LUTs). After training a DNN with a restricted reference range for ILF, all possible inputs are traversed, and the output values of the DNN are cached into LUTs. During the coding process, the filtering process is performed by simply retrieving the filtered pixel through locating the input pixels and interpolating between the cached values, instead of relying on heavy inference computations. In this paper, we propose a universal LUT-based ILF framework, termed LUT-ILF++. First, we introduce the cooperation of multiple kinds of filtering LUTs and propose a series of customized indexing mechanisms to enable better filtering reference perception with limited storage consumption. Second, we propose the cross-component indexing mechanism to enable the filtering of different color components jointly. Third, in order to make our solution practical for coding uses, we propose the LUT compaction scheme to enable the LUT pruning, achieving a lower storage cost of the entire solution. The proposed framework is implemented in the VVC reference software. Experimental results show that the proposed framework achieves on average 0.82%/2.97%/1.63% and 0.85%/4.11%/2.06% bitrate reduction for common test sequences, under the AI and RA configurations, respectively. Compared to DNN-based solutions, our proposed solution has much lower time complexity and storage cost.
31. GrACE: A Generative Approach to Better Confidence Elicitation in Large Language Models
Authors: Zhaohan Zhang, Ziquan Liu, Ioannis Patras β’
Published: 2025-09-11 β’
Source: arXiv
Assessing the reliability of Large Language Models (LLMs) by confidence elicitation is a prominent approach to AI safety in high-stakes applications, such as healthcare and finance. Existing methods either require expensive computational overhead or suffer from poor calibration, making them impractical and unreliable for real-world deployment. In this work, we propose GrACE, a Generative Approach to Confidence Elicitation that enables scalable and reliable confidence elicitation for LLMs. GrACE adopts a novel mechanism in which the model expresses confidence by the similarity between the last hidden state and the embedding of a special token appended to the vocabulary, in real-time. We fine-tune the model for calibrating the confidence with calibration targets associated with accuracy. Experiments with three LLMs and two benchmark datasets show that the confidence produced by GrACE achieves the best discriminative capacity and calibration on open-ended generation tasks, outperforming six competing methods without resorting to additional sampling or an auxiliary model. Moreover, we propose two strategies for improving test-time scaling based on confidence induced by GrACE. Experimental results show that using GrACE not only improves the accuracy of the final decision but also significantly reduces the number of required samples in the test-time scaling scheme, indicating the potential of GrACE as a practical solution for deploying LLMs with scalable, reliable, and real-time confidence estimation.
32. Joint Optimisation of Load Balancing and Energy Efficiency for O-RAN Deployments
Authors: Mohammed M. H. Qazzaz, Abdelaziz Salama, Maryam Hafeez, Syed A. R. Zaidi β’
Published: 2025-09-11 β’
Source: arXiv
Open Radio Access Network (O-RAN) architecture provides an intrinsic capability to exploit key performance monitoring (KPM) within Radio Intelligence Controller (RIC) to derive network optimisation through xApps. These xApps can leverage KPM knowledge to dynamically switch on/off the associated RUs where such a function is supported over the E2 interface. Several existing studies employ artificial intelligence (AI)/Machine Learning (ML) based approaches to realise such dynamic sleeping for increased energy efficiency (EE). Nevertheless, most of these approaches rely upon offloading user equipment (UE) to carve out a sleeping opportunity. Such an approach inherently creates load imbalance across the network. Such load imbalance may impact the throughput performance of offloaded UEs as they might be allocated a lower number of physical resource blocks (PRBs). Maintaining the same PRB allocation while addressing the EE at the network level is a challenging task. To that end, in this article, we present a comprehensive ML-based framework for joint optimisation of load balancing and EE for ORAN deployments. We formulate the problem as a multi-class classification system that predictively evaluates potential RU configurations before optimising the EE, mapping network conditions to three load balance categories (Well Balanced, Moderately Balanced, Imbalanced). Our multi-threshold approach (Conservative, Moderate, Aggressive) accommodates different operational priorities between energy savings and performance assurance. Experimental evaluation using 4.26 million real network measurements from simulations demonstrates that our Random Forest model achieves 98.3% F1-macro performance, representing 195% improvement over traditional baseline strategies.
33. Explaining Tournament Solutions with Minimal Supports
Authors: ClΓ©ment Contet, Umberto Grandi, JΓ©rΓ΄me Mengin β’
Published: 2025-09-11 β’
Source: arXiv
Tournaments are widely used models to represent pairwise dominance between candidates, alternatives, or teams. We study the problem of providing certified explanations for why a candidate appears among the winners under various tournament rules. To this end, we identify minimal supports, minimal sub-tournaments in which the candidate is guaranteed to win regardless of how the rest of the tournament is completed (that is, the candidate is a necessary winner of the sub-tournament). This notion corresponds to an abductive explanation for the question,"Why does the winner win the tournament", a central concept in formal explainable AI. We focus on common tournament solutions: the top cycle, the uncovered set, the Copeland rule, the Borda rule, the maximin rule, and the weighted uncovered set. For each rule we determine the size of the smallest minimal supports, and we present polynomial-time algorithms to compute them for all but the weighted uncovered set, for which the problem is NP-complete. Finally, we show how minimal supports can serve to produce compact, certified, and intuitive explanations.
34. Determination of CKM matrix element and axial vector form factors from weak decays of quantum-entangled strange baryons
Authors: BESIII Collaboration, M. Ablikim, M. N. Achasov, P. Adlarson, X. C. Ai, R. Aliberti, A. Amoroso, Q. An, Y. Bai, O. Bakina, Y. Ban, H. -R. Bao, V. Batozskaya, K. Begzsuren, N. Berger, M. Berlowski, M. Bertani, D. Bettoni, F. Bianchi, E. Bianco, A. Bortone, I. Boyko, R. A. Briere, A. Brueggemann, H. Cai, M. H. Cai, X. Cai, A. Calcaterra, G. F. Cao, N. Cao, S. A. Cetin, X. Y. Chai, J. F. Chang, G. R. Che, Y. Z. Che, C. H. Chen, Chao Chen, G. Chen, H. S. Chen, H. Y. Chen, M. L. Chen, S. J. Chen, S. L. Chen, S. M. Chen, T. Chen, X. R. Chen, X. T. Chen, X. Y. Chen, Y. B. Chen, Y. Q. Chen, Y. Q. Chen, Z. Chen, Z. J. Chen, Z. K. Chen, S. K. Choi, X. Chu, G. Cibinetto, F. Cossio, J. Cottee-Meldrum, J. J. Cui, H. L. Dai, J. P. Dai, A. Dbeyssi, R. E. de Boer, D. Dedovich, C. Q. Deng, Z. Y. Deng, A. Denig, I. Denysenko, M. Destefanis, F. De Mori, B. Ding, X. X. Ding, Y. Ding, Y. Ding, Y. X. Ding, J. Dong, L. Y. Dong, M. Y. Dong, X. Dong, M. C. Du, S. X. Du, S. X. Du, Y. Y. Duan, Z. H. Duan, P. Egorov, G. F. Fan, J. J. Fan, Y. H. Fan, J. Fang, J. Fang, S. S. Fang, W. X. Fang, Y. Q. Fang, L. Fava, F. Feldbauer, G. Felici, C. Q. Feng, J. H. Feng, L. Feng, Q. X. Feng, Y. T. Feng, M. Fritsch, C. D. Fu, J. L. Fu, Y. W. Fu, H. Gao, X. B. Gao, Y. Gao, Y. N. Gao, Y. N. Gao, Y. Y. Gao, S. Garbolino, I. Garzia, L. Ge, P. T. Ge, Z. W. Ge, C. Geng, E. M. Gersabeck, A. Gilman, K. Goetzen, J. D. Gong, L. Gong, W. X. Gong, W. Gradl, S. Gramigna, M. Greco, M. H. Gu, Y. T. Gu, C. Y. Guan, A. Q. Guo, L. B. Guo, M. J. Guo, R. P. Guo, Y. P. Guo, A. Guskov, J. Gutierrez, K. L. Han, T. T. Han, F. Hanisch, K. D. Hao, X. Q. Hao, F. A. Harris, K. K. He, K. L. He, F. H. Heinsius, C. H. Heinz, Y. K. Heng, C. Herold, P. C. Hong, G. Y. Hou, X. T. Hou, Y. R. Hou, Z. L. Hou, H. M. Hu, J. F. Hu, Q. P. Hu, S. L. Hu, T. Hu, Y. Hu, Z. M. Hu, G. S. Huang, K. X. Huang, L. Q. Huang, P. Huang, X. T. Huang, Y. P. Huang, Y. S. Huang, T. Hussain, N. HΓΌsken, N. in der Wiesche, J. Jackson, Q. Ji, Q. P. Ji, W. Ji, X. B. Ji, X. L. Ji, Y. Y. Ji, Z. K. Jia, D. Jiang, H. B. Jiang, P. C. Jiang, S. J. Jiang, T. J. Jiang, X. S. Jiang, Y. Jiang, J. B. Jiao, J. K. Jiao, Z. Jiao, S. Jin, Y. Jin, M. Q. Jing, X. M. Jing, T. Johansson, S. Kabana, N. Kalantar-Nayestanaki, X. L. Kang, X. S. Kang, M. Kavatsyuk, B. C. Ke, V. Khachatryan, A. Khoukaz, R. Kiuchi, O. B. Kolcu, B. Kopf, M. Kuessner, X. Kui, N. Kumar, A. Kupsc, W. KΓΌhn, Q. Lan, W. N. Lan, T. T. Lei, M. Lellmann, T. Lenz, C. Li, C. Li, C. H. Li, C. K. Li, D. M. Li, F. Li, G. Li, H. B. Li, H. J. Li, H. N. Li, Hui Li, J. R. Li, J. S. Li, K. Li, K. L. Li, K. L. Li, L. J. Li, Lei Li, M. H. Li, M. R. Li, P. L. Li, P. R. Li, Q. M. Li, Q. X. Li, R. Li, S. X. Li, T. Li, T. Y. Li, W. D. Li, W. G. Li, X. Li, X. H. Li, X. L. Li, X. Y. Li, X. Z. Li, Y. Li, Y. G. Li, Y. P. Li, Z. J. Li, Z. Y. Li, C. Liang, H. Liang, Y. F. Liang, Y. T. Liang, G. R. Liao, L. B. Liao, M. H. Liao, Y. P. Liao, J. Libby, A. Limphirat, C. C. Lin, D. X. Lin, L. Q. Lin, T. Lin, B. J. Liu, B. X. Liu, C. Liu, C. X. Liu, F. Liu, F. H. Liu, Feng Liu, G. M. Liu, H. Liu, H. B. Liu, H. H. Liu, H. M. Liu, Huihui Liu, J. B. Liu, J. J. Liu, K. Liu, K. Liu, K. Y. Liu, Ke Liu, L. C. Liu, Lu Liu, M. H. Liu, M. H. Liu, P. L. Liu, Q. Liu, S. B. Liu, T. Liu, W. K. Liu, W. M. Liu, W. T. Liu, X. Liu, X. Liu, X. K. Liu, X. L. Liu, X. Y. Liu, Y. Liu, Y. Liu, Y. Liu, Y. B. Liu, Z. A. Liu, Z. D. Liu, Z. Q. Liu, X. C. Lou, F. X. Lu, H. J. Lu, J. G. Lu, X. L. Lu, Y. Lu, Y. H. Lu, Y. P. Lu, Z. H. Lu, C. L. Luo, J. R. Luo, J. S. Luo, M. X. Luo, T. Luo, X. L. Luo, Z. Y. Lv, X. R. Lyu, Y. F. Lyu, Y. H. Lyu, F. C. Ma, H. L. Ma, Heng Ma, J. L. Ma, L. L. Ma, L. R. Ma, Q. M. Ma, R. Q. Ma, R. Y. Ma, T. Ma, X. T. Ma, X. Y. Ma, Y. M. Ma, F. E. Maas, I. MacKay, M. Maggiora, S. Malde, Q. A. Malik, H. X. Mao, Y. J. Mao, Z. P. Mao, S. Marcello, A. Marshall, F. M. Melendi, Y. H. Meng, Z. X. Meng, G. Mezzadri, H. Miao, T. J. Min, R. E. Mitchell, X. H. Mo, B. Moses, N. Yu. Muchnoi, J. Muskalla, Y. Nefedov, F. Nerling, L. S. Nie, I. B. Nikolaev, Z. Ning, S. Nisar, Q. L. Niu, W. D. Niu, C. Normand, S. L. Olsen, Q. Ouyang, S. Pacetti, X. Pan, Y. Pan, A. Pathak, Y. P. Pei, M. Pelizaeus, H. P. Peng, X. J. Peng, Y. Y. Peng, K. Peters, K. Petridis, J. L. Ping, R. G. Ping, S. Plura, V. Prasad, F. Z. Qi, H. R. Qi, M. Qi, S. Qian, W. B. Qian, C. F. Qiao, J. H. Qiao, J. J. Qin, J. L. Qin, L. Q. Qin, L. Y. Qin, P. B. Qin, X. P. Qin, X. S. Qin, Z. H. Qin, J. F. Qiu, Z. H. Qu, J. Rademacker, C. F. Redmer, A. Rivetti, M. Rolo, G. Rong, S. S. Rong, F. Rosini, Ch. Rosner, M. Q. Ruan, N. Salone, A. Sarantsev, Y. Schelhaas, K. Schoenning, M. Scodeggio, K. Y. Shan, W. Shan, X. Y. Shan, Z. J. Shang, J. F. Shangguan, L. G. Shao, M. Shao, C. P. Shen, H. F. Shen, W. H. Shen, X. Y. Shen, B. A. Shi, H. Shi, J. L. Shi, J. Y. Shi, S. Y. Shi, X. Shi, H. L. Song, J. J. Song, T. Z. Song, W. M. Song, Y. J. Song, Y. X. Song, Zirong Song, S. Sosio, S. Spataro, S Stansilaus, F. Stieler, S. S Su, Y. J. Su, G. B. Sun, G. X. Sun, H. Sun, H. K. Sun, J. F. Sun, K. Sun, L. Sun, S. S. Sun, T. Sun, Y. C. Sun, Y. H. Sun, Y. J. Sun, Y. Z. Sun, Z. Q. Sun, Z. T. Sun, C. J. Tang, G. Y. Tang, J. Tang, J. J. Tang, L. F. Tang, Y. A. Tang, L. Y. Tao, M. Tat, J. X. Teng, J. Y. Tian, W. H. Tian, Y. Tian, Z. F. Tian, I. Uman, B. Wang, B. Wang, Bo Wang, C. Wang, C. Wang, Cong Wang, D. Y. Wang, H. J. Wang, J. J. Wang, K. Wang, L. L. Wang, L. W. Wang, M. Wang, M. Wang, N. Y. Wang, Shun Wang, T. Wang, T. J. Wang, W. Wang, W. Wang, W. P. Wang, X. Wang, X. F. Wang, X. J. Wang, X. L. Wang, X. N. Wang, Y. Wang, Y. D. Wang, Y. F. Wang, Y. H. Wang, Y. J. Wang, Y. L. Wang, Y. N. Wang, Y. Q. Wang, Yaqian Wang, Yi Wang, Yuan Wang, Z. Wang, Z. L. Wang, Z. L. Wang, Z. Q. Wang, Z. Y. Wang, D. H. Wei, H. R. Wei, F. Weidner, S. P. Wen, Y. R. Wen, U. Wiedner, G. Wilkinson, M. Wolke, C. Wu, J. F. Wu, L. H. Wu, L. J. Wu, L. J. Wu, Lianjie Wu, S. G. Wu, S. M. Wu, X. Wu, X. H. Wu, Y. J. Wu, Z. Wu, L. Xia, X. M. Xian, B. H. Xiang, D. Xiao, G. Y. Xiao, H. Xiao, Y. L. Xiao, Z. J. Xiao, C. Xie, K. J. Xie, X. H. Xie, Y. Xie, Y. G. Xie, Y. H. Xie, Z. P. Xie, T. Y. Xing, C. F. Xu, C. J. Xu, G. F. Xu, H. Y. Xu, H. Y. Xu, M. Xu, Q. J. Xu, Q. N. Xu, T. D. Xu, W. Xu, W. L. Xu, X. P. Xu, Y. Xu, Y. Xu, Y. C. Xu, Z. S. Xu, F. Yan, H. Y. Yan, L. Yan, W. B. Yan, W. C. Yan, W. H. Yan, W. P. Yan, X. Q. Yan, H. J. Yang, H. L. Yang, H. X. Yang, J. H. Yang, R. J. Yang, T. Yang, Y. Yang, Y. F. Yang, Y. H. Yang, Y. Q. Yang, Y. X. Yang, Y. Z. Yang, M. Ye, M. H. Ye, Z. J. Ye, Junhao Yin, Z. Y. You, B. X. Yu, C. X. Yu, G. Yu, J. S. Yu, L. Q. Yu, M. C. Yu, T. Yu, X. D. Yu, Y. C. Yu, C. Z. Yuan, H. Yuan, J. Yuan, J. Yuan, L. Yuan, S. C. Yuan, S. H. Yuan, X. Q. Yuan, Y. Yuan, Z. Y. Yuan, C. X. Yue, Ying Yue, A. A. Zafar, S. H. Zeng, X. Zeng, Y. Zeng, Y. J. Zeng, Y. J. Zeng, X. Y. Zhai, Y. H. Zhan, Zhang, A. Q. Zhang, B. L. Zhang, B. X. Zhang, D. H. Zhang, G. Y. Zhang, G. Y. Zhang, H. Zhang, H. Zhang, H. C. Zhang, H. H. Zhang, H. Q. Zhang, H. R. Zhang, H. Y. Zhang, J. Zhang, J. Zhang, J. J. Zhang, J. L. Zhang, J. Q. Zhang, J. S. Zhang, J. W. Zhang, J. X. Zhang, J. Y. Zhang, J. Z. Zhang, Jianyu Zhang, L. M. Zhang, Lei Zhang, N. Zhang, P. Zhang, Q. Zhang, Q. Y. Zhang, R. Y. Zhang, S. H. Zhang, Shulei Zhang, X. M. Zhang, X. Y Zhang, X. Y. Zhang, Y. Zhang, Y. Zhang, Y. T. Zhang, Y. H. Zhang, Y. M. Zhang, Y. P. Zhang, Z. D. Zhang, Z. H. Zhang, Z. L. Zhang, Z. L. Zhang, Z. X. Zhang, Z. Y. Zhang, Z. Y. Zhang, Z. Z. Zhang, Zh. Zh. Zhang, G. Zhao, J. Y. Zhao, J. Z. Zhao, L. Zhao, L. Zhao, M. G. Zhao, N. Zhao, R. P. Zhao, S. J. Zhao, Y. B. Zhao, Y. L. Zhao, Y. X. Zhao, Z. G. Zhao, A. Zhemchugov, B. Zheng, B. M. Zheng, J. P. Zheng, W. J. Zheng, X. R. Zheng, Y. H. Zheng, B. Zhong, C. Zhong, H. Zhou, J. Q. Zhou, J. Y. Zhou, S. Zhou, X. Zhou, X. K. Zhou, X. R. Zhou, X. Y. Zhou, Y. X. Zhou, Y. Z. Zhou, A. N. Zhu, J. Zhu, K. Zhu, K. J. Zhu, K. S. Zhu, L. Zhu, L. X. Zhu, S. H. Zhu, T. J. Zhu, W. D. Zhu, W. D. Zhu, W. J. Zhu, W. Z. Zhu, Y. C. Zhu, Z. A. Zhu, X. Y. Zhuang, J. H. Zou, J. Zu β’
Published: 2025-09-11 β’
Source: arXiv
The electromagnetic structure of the nucleon can be determined from the scattering of electrons off a nucleon target. However, to study its axial structure, neutrino beams are required. The results from these experiments should be extrapolated to zero energy-momentum transfers to access the static properties of the nucleon. For baryons with strange quarks, hyperons, the static limit can instead be approached in semi-leptonic decays, which give direct access to the weak magnetism and axial-vector coupling strengths that are inaccessible in electromagnetic interactions. The axial-vector coupling as while weak magnetism coupling and the overall normalization, given by form factor $f_1$, are being determined with increased precision from the theory of strong interactions using a first principles formulation on the space--time lattice. Furthermore, the probability of the semi-leptonic hyperon decay is approximately proportional to $|V_{us}|^2\cdot (f_1^2+3g_1^2)$, where $V_{us}$ is the CKM matrix element responsible for the transition between an $s$ and a $u$ quark. Current determinations of $|V_{us}|$ come from kaon decays, but the results are not consistent and could indicate a deviation from CKM matrix unitarity, a tell-tale sign of physics beyond the Standard Model (SM) of elementary particles. Here we determine the absolute branching fraction and weak coupling strengths for $\Lambda\to p e^-\bar\nu_e$, and $\bar \Lambda\to \bar p e^+\nu_e$. These observables combined with form factors determined from first-principle lattice QCD calculations allow for the extraction of the $|V_{us}|$ value. We demonstrate how $|V_{us}|$ can be extracted with increasing sensitivity using polarized hyperons from entangled, baryon-antibaryon pairs, thus enabling a complementary road to that of meson decays. In addition, the presented experimental method can be used for other semileptonic decays of baryons.
35. CoAtNeXt:An Attention-Enhanced ConvNeXtV2-Transformer Hybrid Model for Gastric Tissue Classification
Authors: Mustafa Yurdakul, Sakir Tasdemir β’
Published: 2025-09-11 β’
Source: arXiv
Background and objective Early diagnosis of gastric diseases is crucial to prevent fatal outcomes. Although histopathologic examination remains the diagnostic gold standard, it is performed entirely manually, making evaluations labor-intensive and prone to variability among pathologists. Critical findings may be missed, and lack of standard procedures reduces consistency. These limitations highlight the need for automated, reliable, and efficient methods for gastric tissue analysis. Methods In this study, a novel hybrid model named CoAtNeXt was proposed for the classification of gastric tissue images. The model is built upon the CoAtNet architecture by replacing its MBConv layers with enhanced ConvNeXtV2 blocks. Additionally, the Convolutional Block Attention Module (CBAM) is integrated to improve local feature extraction through channel and spatial attention mechanisms. The architecture was scaled to achieve a balance between computational efficiency and classification performance. CoAtNeXt was evaluated on two publicly available datasets, HMU-GC-HE-30K for eight-class classification and GasHisSDB for binary classification, and was compared against 10 Convolutional Neural Networks (CNNs) and ten Vision Transformer (ViT) models. Results CoAtNeXt achieved 96.47% accuracy, 96.60% precision, 96.47% recall, 96.45% F1 score, and 99.89% AUC on HMU-GC-HE-30K. On GasHisSDB, it reached 98.29% accuracy, 98.07% precision, 98.41% recall, 98.23% F1 score, and 99.90% AUC. It outperformed all CNN and ViT models tested and surpassed previous studies in the literature. Conclusion Experimental results show that CoAtNeXt is a robust architecture for histopathological classification of gastric tissue images, providing performance on binary and multiclass. Its highlights its potential to assist pathologists by enhancing diagnostic accuracy and reducing workload.