Fusion

arXiv:2508.16358v1 Announce Type: new Abstract: Fusion power systems can in principle be used to make significant amounts of fissile material. To do so, an operator would have to introduce fertile material, such as uranium-238, in a suitable region of the reactor where it is exposed to an intense neutron flux. The possibility of using a fusion reactor for this purpose has raised the question of how these facilities can be monitored to ensure their peaceful use. This study examines whether covert production of fissile material in a declared fusion plant could be detected with an onsite antineutrino detector. We find that even a relatively small detector should be able to confirm production rates of a few kilograms of plutonium over 30 days, despite the cosmogenic background and the antineutrino emissions associated with neutron activation of reactor components.

arXiv:2508.15967v1 Announce Type: new Abstract: Modern supercomputers are increasingly relying on Graphic Processing Units (GPUs) and other accelerators to achieve exa-scale performance at reasonable energy usage. The challenge of exploiting these accelerators is the incompatibility between different vendors. A scientific code written using CUDA will not operate on a AMD gpu. Frameworks that can abstract the physics from the accelerator kernel code are needed to exploit the current and future hardware. In the world of machine learning, several auto differentiation frameworks have been developed that have the promise of abstracting the math from the compute hardware. However in practice, these framework often lag in supporting non-CUDA platforms. Their reliance on python makes them challenging to embed within non python based applications. In this paper we present the development of a graph computation framework which compiles physics equations to optimized kernel code for the central

arXiv:2508.16204v1 Announce Type: new Abstract: Model merging is a powerful technique for integrating the specialized knowledge of multiple machine learning models into a single model. However, existing methods require manually partitioning model parameters into fixed groups for merging, which restricts the exploration of potential combinations and limits performance. To overcome these limitations, we propose Model Merging of Natural Niches (M2N2), an evolutionary algorithm with three key features: (1) dynamic adjustment of merging boundaries to progressively explore a broader range of parameter combinations; (2) a diversity preservation mechanism inspired by the competition for resources in nature, to maintain a population of diverse, high-performing models that are particularly well-suited for merging; and (3) a heuristicbased attraction metric to identify the most promising pairs of models for fusion. Our experimental results demonstrate, for the first time, that model merging can

arXiv:2508.15852v1 Announce Type: new Abstract: We introduce PGF-Net (Progressive Gated-Fusion Network), a novel deep learning framework designed for efficient and interpretable multimodal sentiment analysis. Our framework incorporates three primary innovations. Firstly, we propose a Progressive Intra-Layer Fusion paradigm, where a Cross-Attention mechanism empowers the textual representation to dynamically query and integrate non-linguistic features from audio and visual streams within the deep layers of a Transformer encoder. This enables a deeper, context-dependent fusion process. Secondly, the model incorporates an Adaptive Gated Arbitration mechanism, which acts as a dynamic controller to balance the original linguistic information against the newly fused multimodal context, ensuring stable and meaningful integration while preventing noise from overwhelming the signal. Lastly, a hybrid Parameter-Efficient Fine-Tuning (PEFT) strategy is employed, synergistically combining global

Researchers at the University of British Columbia have shown that a small bench-top reactor can enhance nuclear fusion rates by electrochemically loading a metal with deuterium fuel. Unlike massive magnetic confinement reactors, their experiment uses a room-temperature setup that packs deuterium into palladium like a sponge, boosting the likelihood of fusion events.

arXiv:2508.15653v1 Announce Type: new Abstract: Online HD map construction is a fundamental task in autonomous driving systems, aiming to acquire semantic information of map elements around the ego vehicle based on real-time sensor inputs. Recently, several approaches have achieved promising results by incorporating offline priors such as SD maps and HD maps or by fusing multi-modal data. However, these methods depend on stale offline maps and multi-modal sensor suites, resulting in avoidable computational overhead at inference. To address these limitations, we employ a knowledge distillation strategy to transfer knowledge from multimodal models with prior knowledge to an efficient, low-cost, and vision-centric student model. Specifically, we propose MapKD, a novel multi-level cross-modal knowledge distillation framework with an innovative Teacher-Coach-Student (TCS) paradigm. This framework consists of: (1) a camera-LiDAR fusion model with SD/HD map priors serving as the teacher; (2)

arXiv:2508.15537v1 Announce Type: new Abstract: Extracting narrow roads from high-resolution remote sensing imagery remains a significant challenge due to their limited width, fragmented topology, and frequent occlusions. To address these issues, we propose D3FNet, a Dilated Dual-Stream Differential Attention Fusion Network designed for fine-grained road structure segmentation in remote perception systems. Built upon the encoder-decoder backbone of D-LinkNet, D3FNet introduces three key innovations:(1) a Differential Attention Dilation Extraction (DADE) module that enhances subtle road features while suppressing background noise at the bottleneck; (2) a Dual-stream Decoding Fusion Mechanism (DDFM) that integrates original and attention-modulated features to balance spatial precision with semantic context; and (3) a multi-scale dilation strategy (rates 1, 3, 5, 9) that mitigates gridding artifacts and improves continuity in narrow road prediction. Unlike conventional models that

arXiv:2508.15505v1 Announce Type: new Abstract: Multimodal Image Fusion (MMIF) aims to integrate complementary information from different imaging modalities to overcome the limitations of individual sensors. It enhances image quality and facilitates downstream applications such as remote sensing, medical diagnostics, and robotics. Despite significant advancements, current MMIF methods still face challenges such as modality misalignment, high-frequency detail destruction, and task-specific limitations. To address these challenges, we propose AdaSFFuse, a novel framework for task-generalized MMIF through adaptive cross-domain co-fusion learning. AdaSFFuse introduces two key innovations: the Adaptive Approximate Wavelet Transform (AdaWAT) for frequency decoupling, and the Spatial-Frequency Mamba Blocks for efficient multimodal fusion. AdaWAT adaptively separates the high- and low-frequency components of multimodal images from different scenes, enabling fine-grained extraction and

One physicist says his design to use nuclear waste as fuel for nuclear fusion could help the U.S. be a leader in the fusion economy.

arXiv:2508.14622v1 Announce Type: new Abstract: A linear gyrokinetic eigenvalue code is developed to study the stability of toroidal Alfv\'en eigenmode (TAE) in general axisymmetric toroidal geometry, with the self-consistent treatment of energetic particle drive and core plasma Landau damping in a non-perturbative way. The general particle responses of both circulating and trapped particles are incorporated in the calculation by means of the action-angle approach, and, particularly, the finite Larmor radius and orbit width effects of energetic particles are fully taken into account. The ballooning-mode representation is adopted to solve the eigenmode equations in order to reduce the computational resource while obtaining a high resolution of the fine radial structure. Furthermore, the code is able to study the physics of wave-particle interaction in great detail, thanks to the development of systematic theory-based numerical diagnostics, including effective mode structure and phase

arXiv:2508.14844v1 Announce Type: new Abstract: Accurately predicting enzyme functionality remains one of the major challenges in computational biology, particularly for enzymes with limited structural annotations or sequence homology. We present a novel multimodal Quantum Machine Learning (QML) framework that enhances Enzyme Commission (EC) classification by integrating four complementary biochemical modalities: protein sequence embeddings, quantum-derived electronic descriptors, molecular graph structures, and 2D molecular image representations. Quantum Vision Transformer (QVT) backbone equipped with modality-specific encoders and a unified cross-attention fusion module. By integrating graph features and spatial patterns, our method captures key stereoelectronic interactions behind enzyme function. Experimental results demonstrate that our multimodal QVT model achieves a top-1 accuracy of 85.1%, outperforming sequence-only baselines by a substantial margin and achieving better

arXiv:2508.14719v1 Announce Type: new Abstract: Photon-Counting Computed Tomography (PCCT) is a novel imaging modality that simultaneously acquires volumetric data at multiple X-ray energy levels, generating separate volumes that capture energy-dependent attenuation properties. Attenuation refers to the reduction in X-ray intensity as it passes through different tissues or materials. This spectral information enhances tissue and material differentiation, enabling more accurate diagnosis and analysis. However, the resulting multivolume datasets are often complex and redundant, making visualization and interpretation challenging. To address these challenges, we propose a method for fusing spectral PCCT data into a single representative volume that enables direct volume rendering and segmentation by leveraging both shared and complementary information across different channels. Our approach starts by computing 2D histograms between pairs of volumes to identify those that exhibit

A 1989 experiment offered the promise of nuclear fusion without the need for high temperatures, but this "cold fusion" was quickly debunked. Now, some of the techniques involved have been resurrected in a new experiment that could actually improve efforts to achieve practical fusion power

Using a small bench-top reactor, researchers at the University of British Columbia (UBC) have demonstrated that electrochemically loading a solid metal target with deuterium fuel can boost nuclear fusion rates.

Practical fusion power that can provide cheap, clean energy could be a step closer thanks to artificial intelligence. Scientists at Lawrence Livermore National Laboratory have developed a deep learning model that accurately predicted the results of a nuclear fusion experiment conducted in 2022. Accurate predictions can help speed up the design of new experiments and accelerate the quest for this virtually limitless energy source.

arXiv:2508.13492v1 Announce Type: new Abstract: Keyhole-induced (KH) porosity, which arises from unstable vapor cavity dynamics under excessive laser energy input, remains a significant challenge in laser powder bed fusion (LPBF). This study presents an integrated experimental and data-driven framework using airborne acoustic emission (AE) to achieve high-resolution quantification of KH porosity. Experiments conducted on an LPBF system involved in situ acquisition of airborne AE and ex situ porosity imaging via X-ray computed tomography (XCT), synchronized spatiotemporally through photodiode signals with submillisecond precision. We introduce KHLineNum, a spatially resolved porosity metric defined as the number of KH pores per unit scan length, which serves as a physically meaningful indicator of the severity of KH porosity in geometries and scanning strategies. Using AE scalogram data and scan speed, we trained a lightweight convolutional neural network to predict KHLineNum with

arXiv:2508.13357v1 Announce Type: new Abstract: Secure Multi-Party Computation (MPC) offers a practical foundation for privacy-preserving machine learning at the edge, with MPC commonly employed to support nonlinear operations. These MPC protocols fundamentally rely on Oblivious Transfer (OT), particularly Correlated OT (COT), to generate correlated randomness essential for secure computation. Although COT generation is efficient in conventional two-party settings with resource-rich participants, it becomes a critical bottleneck in real-world inference on resource-constrained devices (e.g., IoT sensors and wearables), due to both communication latency and limited computational capacity. To enable real-time secure inference, we introduce Silentflow, a highly efficient Trusted Execution Environment (TEE)-assisted protocol that eliminates communication in COT generation. We tackle the core performance bottleneck-low computational intensity-through structured algorithmic decomposition:

arXiv:2508.12783v1 Announce Type: cross Abstract: Muon-catalyzed fusion has recently regained significant attention due to experimental and theoretical developments being performed. The present authors [Phys. Rev. C {\bf 109} 054625 (2024)] proposed the tractable $T$-matrix model based on the Lippmann-Schwinger equation to approximate the elaborate two- and three-body coupled-channel (CC) calculations [Kamimura, Kino, and Yamashita, Phys. Rev. C {\bf 107}, 034607 (2023)] for the nuclear reaction processes in the muonic molecule $dt\mu$, $(dt\mu)_{J=0} \to\!^4{\rm He} + n + \mu + 17.6 \, {\rm MeV}$. % or $(^4{\rm He}\mu)_{nl} + n + 17.6 \,{\rm MeV}$. The $T$-matrix model well reproduced almost all of the results generated by the CC work. In the present paper, we apply this model to the nuclear reaction processes in the $dd\mu$ molecule, $(dd\mu)_{J=1} \to\!^3{\rm He} + n + \mu +3.27 \,$ MeV or $t + p + \mu + 4.03 \,$ MeV, in which the fusion takes place via the $p$-wave $d$-$d$

arXiv:2508.13105v1 Announce Type: new Abstract: Many disruptions are caused by resistive wall tearing modes (RWTM). A database of DIII-D locked mode disruptions provides two main disruption criteria, which are shown to be signatures of RWTMs. The first is that the q = 2 rational surface must be sufficiently close the resistive wall surrounding the plasma to interact with it. If active feedback is used, this implies that RWTMs can be prevented from causing major disruptions. This is demonstrated in simulations. The second criterion is that the current profile is sufficiently peaked. This is caused by edge cooling, such as by impurity radiation and turbulence, which suppress edge current and temperature. This implies the disruptions are not caused by neoclassical tearing modes (NTM), because the bootstrap current is also suppressed. At high $\beta,$ resistive wall modes (RWM) can be stabilized with feedback. Feedback also stabilizes high $\beta$ RWTMs, as shown in NSTX data and in

arXiv:2508.12820v1 Announce Type: new Abstract: The near-axis theory for quasi-isodynamic stellarator equilibria is reformulated in terms of geometric inputs, to allow greater control of the ``direct construction'' of quasi-isodynamic configurations, and to facilitate understanding of the space of such equilibria. This includes a method to construct suitable magnetic axis curves by solving Frenet-Serret equations, and an approach to controlling magnetic surface shaping at first order (plasma elongation), which previously has required careful parameter selection or additional optimization steps. The approach is suitable for studying different classes of quasi-isodynamic stellarators including different axis ``helicities'' and topologies (e.g. knotted solutions), and as the basis for future systematic surveys using higher order near-axis theory. As an example application, we explore a family of configurations with per-field-period axis helicity equal to one half, demonstrating an

arXiv:2508.13048v1 Announce Type: new Abstract: Large Language Models (LLMs) have exhibited remarkable capabilities but remain vulnerable to jailbreaking attacks, which can elicit harmful content from the models by manipulating the input prompts. Existing black-box jailbreaking techniques primarily rely on static prompts crafted with a single, non-adaptive strategy, or employ rigid combinations of several underperforming attack methods, which limits their adaptability and generalization. To address these limitations, we propose MAJIC, a Markovian adaptive jailbreaking framework that attacks black-box LLMs by iteratively combining diverse innovative disguise strategies. MAJIC first establishes a ``Disguise Strategy Pool'' by refining existing strategies and introducing several innovative approaches. To further improve the attack performance and efficiency, MAJIC formulate the sequential selection and fusion of strategies in the pool as a Markov chain. Under this formulation, MAJIC

arXiv:2508.12346v1 Announce Type: new Abstract: The Mamba architecture has emerged as a promising alternative to CNNs and Transformers for image deblurring. However, its flatten-and-scan strategy often results in local pixel forgetting and channel redundancy, limiting its ability to effectively aggregate 2D spatial information. Although existing methods mitigate this by modifying the scan strategy or incorporating local feature modules, it increase computational complexity and hinder real-time performance. In this paper, we propose a structure-aware image deblurring network without changing the original Mamba architecture. Specifically, we design a memory buffer mechanism to preserve historical information for later fusion, enabling reliable modeling of relevance between adjacent features. Additionally, we introduce an Ising-inspired regularization loss that simulates the energy minimization of the physical system's "mutual attraction" between pixels, helping to maintain image

arXiv:2508.12022v1 Announce Type: new Abstract: Major Depressive Disorder is one of the leading causes of disability worldwide, yet its diagnosis still depends largely on subjective clinical assessments. Integrating Artificial Intelligence (AI) holds promise for developing objective, scalable, and timely diagnostic tools. In this paper, we present a comprehensive survey of state-of-the-art AI methods for depression detection and diagnosis, based on a systematic review of 55 key studies. We introduce a novel hierarchical taxonomy that structures the field by primary clinical task (diagnosis vs. prediction), data modality (text, speech, neuroimaging, multimodal), and computational model class (e.g., graph neural networks, large language models, hybrid approaches). Our in-depth analysis reveals three major trends: the predominance of graph neural networks for modeling brain connectivity, the rise of large language models for linguistic and conversational data, and an emerging focus on

arXiv:2508.11695v1 Announce Type: new Abstract: The rapid advancement of Artificial Intelligence Generated Content (AIGC) techniques has unlocked opportunities in generating diverse and compelling advertising images based on referenced product images and textual scene descriptions. This capability substantially reduces human labor and production costs in traditional marketing workflows. However, existing AIGC techniques either demand extensive fine-tuning for each referenced image to achieve high fidelity, or they struggle to maintain fidelity across diverse products, making them impractical for e-commerce and marketing industries. To tackle this limitation, we first construct AdProd-100K, a large-scale advertising image generation dataset. A key innovation in its construction is our dual data augmentation strategy, which fosters robust, 3D-aware representations crucial for realistic and high-fidelity image synthesis. Leveraging this dataset, we propose RefAdGen, a generation

arXiv:2508.11564v1 Announce Type: new Abstract: The simulation of turbulence in the boundary region of a tokamak is crucial for understanding and optimizing the performance of fusion reactors. In this work, the use of low-rank linear algebra techniques is shown to enhance the efficiency of boundary simulations, specifically by accelerating the solution of a kinetic model for the neutral particles. Solving the kinetic model deterministically using the method of characteristics requires the solution of integral equations, which typically result in dense linear systems upon discretization. We employ hierarchical matrix approximations to significantly reduce the computational cost of assembling and solving the linear systems, leading to substantial savings in both time and memory. The hierarchical matrix method is implemented and tested within the GBS simulation code for boundary simulations, achieving over 90\% reduction in computation time and memory, and enabling simulations with

arXiv:2508.11156v1 Announce Type: new Abstract: A novel technique for measuring plasma conditions using monochromatic pump-broadband probe laser interactions has been experimentally demonstrated. Originally proposed in [J. Ludwig et al., Phys. Plasmas \textbf{26}, 113108 (2019)], this method utilizes crossed-beam energy transfer between the broadband probe and the pump, mediated by plasma ion acoustic waves. The complete energy transfer spectrum can be captured in a single shot, enabling the inference of plasma parameters such as density, electron and ion temperatures, and flow velocity. Compared to Thomson scattering, this technique offers signal enhancements typically larger than 9 orders of magnitude, significantly reducing the required probe laser intensity and facilitating interactions that are linear and measurements that are non-perturbative of the plasma. Furthermore, it provides a powerful tool for advancing studies of crossed-beam energy transfer under conditions relevant to

arXiv:2508.11442v1 Announce Type: new Abstract: Learning unified text embeddings that excel across diverse downstream tasks is a central goal in representation learning, yet negative transfer remains a persistent obstacle. This challenge is particularly pronounced when jointly training a single encoder for Information Retrieval (IR) and Semantic Textual Similarity (STS), two essential but fundamentally disparate tasks for which naive co-training typically yields steep performance trade-offs. We argue that resolving this conflict requires systematically decoupling task-specific learning signals throughout the training pipeline. To this end, we introduce CoDiEmb, a unified framework that reconciles the divergent requirements of IR and STS in a collaborative yet distinct manner. CoDiEmb integrates three key innovations for effective joint optimization: (1) Task-specialized objectives paired with a dynamic sampler that forms single-task batches and balances per-task updates, thereby

arXiv:2508.10627v1 Announce Type: new Abstract: Heating and ionization are among the most fundamental processes in ultra-short, relativistic laser-solid interactions. However, capturing their spatiotemporal evolution experimentally is challenging due to the inherently transient and non-local thermodynamic equilibrium (NLTE) nature. Here, time-resolved resonant X-ray emission spectroscopy, in conjunction with simultaneous X-ray absorption imaging, is employed to investigate such complex dynamics in a thin copper wire driven by an optical high-intensity laser pulse, with sub-picosecond temporal resolution. The diagnostic leverages the high brightness and narrow spectral bandwidth of an X-ray free-electron laser, to selectively excite resonant transitions of highly charged ions within the hot dense plasma generated by the optical laser. The measurements reveal a distinct rise-and-fall temporal evolution of the resonant X-ray emission yield-and consequently the selected ion

arXiv:2508.10408v1 Announce Type: new Abstract: Understanding the interaction between turbulence and zonal flows is critical for modeling turbulence transport in fusion plasmas, often described through predator-prey dynamics. However, traditional deterministic models like the Lotka-Volterra equations simplify this interaction and fail to capture the small fluctuations in simulation data. In this study, we develop a neural network model based on stochastic differential equations (SDEs) to represent the predator-prey dynamics using limited data from simulations of the modified Hasegawa-Wakatani system. We extract the drift and diffusion terms via neural networks, incorporating physical constraints and employing the unscented transform to mitigate challenges brought by limited data. The model accurately reproduces key dynamical features, including stagnation phenomena and energy exchange mechanisms, and the state density distribution generated from the model shows a low KL divergence

arXiv:2508.10063v1 Announce Type: new Abstract: Time series forecasting is a critical first step in generating demand plans for supply chains. Experiments on time series models typically focus on demonstrating improvements in forecast accuracy over existing/baseline solutions, quantified according to some accuracy metric. There is no doubt that forecast accuracy is important; however in production systems, demand planners often value consistency and stability over incremental accuracy improvements. Assuming that the inputs have not changed significantly, forecasts that vary drastically from one planning cycle to the next require high amounts of human intervention, which frustrates demand planners and can even cause them to lose trust in ML forecasting models. We study model-induced stochasticity, which quantifies the variance of a set of forecasts produced by a single model when the set of inputs is fixed. Models with lower variance are more stable. Recently the forecasting

arXiv:2508.09169v1 Announce Type: cross Abstract: One key challenge for efficiency and safety in fusion devices is the retention of tritium (T) in plasma-facing components. Tritium retention generates radioactive concerns and decreases the amount of fuel available to generate power. Hence, understanding the behavior of T in tungsten (W), as the main candidate as armor material, is critical to the deployment of fusion as a reliable energy source. In this work, we have studied the effect of a thermal gradient in the transport properties of hydrogen (as a T surrogate) in pure W. Strong thermal gradients develop in the divertor as a result of the intense energy fluxes arriving at the material. We have developed an analytical approach to compute the heat of transport ($Q^*$) that is parameterized from molecular dynamics (MD) simulations. $Q^*$ is a parameter needed in irreversible thermodynamics frameworks to understand mass transport in the presence of thermal gradients. We show that

arXiv:2508.09431v1 Announce Type: new Abstract: We have investigated the method of extracting the temperature from weighted proton-to-neutron yield ratio from fusion reactions as in the previous experiment~[W. Bang, {\it et al.}, Phys. Rev. Lett. \textbf{111}, 055002 (2013).] using the Texas Petawatt laser beam. The Coulomb explosion of deuterium clusters is simulated based on the particle-in-cell model in a box system with periodic boundary conditions, and fusion reactions are incorporated through the stochastic method. As long as the deuteron numbers in deuterium clusters follow a log-normal distribution, the low-energy part of the final deuteron spectrum can be fitted by a Maxwell-Boltzmann distribution, while there are more deuterons in the intermediate- and high-energy region compared to a thermal distribution, and dominate the weighted yield ratio. Therefore, the effective temperature extracted from the weighted yield ratio is generally higher than that from fitting the final

arXiv:2508.09321v1 Announce Type: new Abstract: Stellarator optimization often takes a two-stage approach, where in the first stage the boundary is varied in order to optimize for some physics metrics, while in the second stage the boundary is kept fixed and coils are sought to generate a magnetic field that can recreate the desired stellarator. Past literature dealing with this stage lacks details on the coil cutting procedure and the mathematical and physical properties of the surface current potential which dictates it. In this work, some basic physical quantities of the surface current and how they relate to the parameters in the current potential are presented, and supported for the first time by explicit mathematical derivations. Additionally, the details of how to account for the presence of an external field in the surface current algorithm are explicitly presented. These relations underpin the procedure of discretizing the surface current into coils. Finally, the

arXiv:2508.09533v1 Announce Type: new Abstract: Detecting tiny objects in multimodal Red-Green-Blue-Thermal (RGBT) imagery is a critical challenge in computer vision, particularly in surveillance, search and rescue, and autonomous navigation. Drone-based scenarios exacerbate these challenges due to spatial misalignment, low-light conditions, occlusion, and cluttered backgrounds. Current methods struggle to leverage the complementary information between visible and thermal modalities effectively. We propose COXNet, a novel framework for RGBT tiny object detection, addressing these issues through three core innovations: i) the Cross-Layer Fusion Module, fusing high-level visible and low-level thermal features for enhanced semantic and spatial accuracy; ii) the Dynamic Alignment and Scale Refinement module, correcting cross-modal spatial misalignments and preserving multi-scale features; and iii) an optimized label assignment strategy using the GeoShape Similarity Measure for better

Scientists have developed a lightning-fast AI tool called HEAT-ML that can spot hidden “safe zones” inside a fusion reactor where parts are protected from blistering plasma heat. Finding these areas, known as magnetic shadows, is key to keeping reactors running safely and moving fusion energy closer to reality.

A public‑private partnership between Commonwealth Fusion Systems (CFS), the U.S. Department of Energy's (DOE) Princeton Plasma Physics Laboratory (PPPL) and Oak Ridge National Laboratory has led to a new artificial intelligence (AI) approach that is faster at finding what's known as "magnetic shadows" in a fusion vessel: safe havens protected from the intense heat of the plasma.

Neutrinos—ghostly particles that rarely interact with normal matter—are the sun's secret messengers. These particles are born deep within the sun, a byproduct of the nuclear fusion process which powers all stars.

arXiv:2508.09067v1 Announce Type: new Abstract: We consider a magnetised plasma in contact with an absorbing planar wall, where the angle $\alpha$ between the magnetic field and the wall is small, $\alpha \ll 1$ (in radians) and the system is symmetric tangential to the wall. The finite ratio $\gamma$ of the characteristic electron gyroradius $\rho_{\rm e}$ to the Debye length $\lambda_{\rm D}$, $\gamma = \rho_{\rm e} / \lambda_{\rm D}$, is retained via a grazing-incidence ($\alpha \ll 1$) gyrokinetic treatment [1,2]. Building on a previously developed iterative scheme [2,3] to solve for the steady-state electrostatic potential in the quasineutral magnetic presheath of width $\sim \rho_{\rm S}$, we developed a scheme that simultaneously solves for both the presheath and the non-neutral Debye sheath of width $\sim \lambda_{\rm D}$ in the limit $\lambda_{\rm D} / \rho_{\rm S} \rightarrow 0$. The code, called GYRAZE, thus provides the energy-angle distribution of ions at the wall and the

arXiv:2508.01687v2 Announce Type: replace Abstract: Explaining machine learning (ML) models for time series (TS) classification remains challenging due to the difficulty of interpreting raw time series and the high dimensionality of the input space. We introduce PHAR-Post-hoc Attribution Rules-a unified framework that transforms numeric feature attributions from post-hoc, instance-wise explainers (e.g., LIME, SHAP) into structured, human-readable rules. These rules define interpretable intervals that indicate where and when key decision boundaries occur, enhancing model transparency. PHAR performs comparably to native rule-based methods, such as Anchor, while scaling more efficiently to long TS sequences and achieving broader instance coverage. A dedicated rule fusion step consolidates rule sets using strategies like weighted selection and lasso-based refinement, balancing key quality metrics: coverage, confidence, and simplicity. This fusion ensures each instance receives a concise

arXiv:2508.01057v2 Announce Type: replace Abstract: Autonomous driving (AD) systems relying solely on onboard sensors may fail to detect distant or obstacle hazards, potentially causing preventable collisions; however, existing transformer-based Vehicle-to-Everything (V2X) approaches, which mitigate AD sensing limitations, either lack effective multimodal fusion and reasoning or struggle to meet real-time performance requirements under complex, high-dimensional traffic conditions. This paper proposes the Real-time Edge-based Autonomous Co-pilot Trajectory planner (REACT), a V2X-integrated trajectory optimization framework for AD based on a fine-tuned lightweight Vision-Language Model (VLM). REACT integrates infrastructure-provided hazard alerts with onboard sensor data, capturing intricate surrounding traffic dynamics and vehicle intents through visual embeddings, interpreting precise numerical data from symbolic inputs, and employing contextual reasoning to generate optimized,

arXiv:2508.08917v1 Announce Type: new Abstract: LiDAR-based Place Recognition (LPR) remains a critical task in Embodied Artificial Intelligence (AI) and Autonomous Driving, primarily addressing localization challenges in GPS-denied environments and supporting loop closure detection. Existing approaches reduce place recognition to a Euclidean distance-based metric learning task, neglecting the feature space's intrinsic structures and intra-class variances. Such Euclidean-centric formulation inherently limits the model's capacity to capture nonlinear data distributions, leading to suboptimal performance in complex environments and temporal-varying scenarios. To address these challenges, we propose a novel cross-view network based on an innovative fusion paradigm. Our framework introduces a pseudo-global information guidance mechanism that coordinates multi-modal branches to perform feature learning within a unified semantic space. Concurrently, we propose a Manifold Adaptation and

arXiv:2508.06497v1 Announce Type: new Abstract: Accurate forecasting of commodity price spikes is vital for countries with limited economic buffers, where sudden increases can strain national budgets, disrupt import-reliant sectors, and undermine food and energy security. This paper introduces a hybrid forecasting framework that combines historical commodity price data with semantic signals derived from global economic news, using an agentic generative AI pipeline. The architecture integrates dual-stream Long Short-Term Memory (LSTM) networks with attention mechanisms to fuse structured time-series inputs with semantically embedded, fact-checked news summaries collected from 1960 to 2023. The model is evaluated on a 64-year dataset comprising normalized commodity price series and temporally aligned news embeddings. Results show that the proposed approach achieves a mean AUC of 0.94 and an overall accuracy of 0.91 substantially outperforming traditional baselines such as logistic

arXiv:2508.06761v1 Announce Type: new Abstract: The Polywell fusion concept, originally proposed by Robert W. Bussard in 1985, has been investigated for over four decades as a potential solution for achieving net fusion energy in a compact and economically viable reactor. It combines two distinct approaches: high-beta magnetic cusp confinement of electrons using polyhedral coil configurations and electrostatic ion confinement via a potential well formed by injected electron beams. While the hybrid nature of the Polywell system offers advantages in plasma stability and engineering simplicity, previous efforts have been limited by persistent challenges in achieving sufficient plasma confinement required to generate a net energy gain. In this study, we examine previous work and identify limitations of several Polywell embodiments that have historically impeded progress. We present an updated Polywell physics model incorporating experimental findings and recent first-principles

arXiv:2508.06561v1 Announce Type: new Abstract: Unravelling the internal structure of hadrons and nuclei in terms of the quarks and gluons of Quantum Chromodynamics is a central focus of current nuclear physics research. Directly observing gluonic states in the nucleus would be groundbreaking and is an objective of the future Electron-Ion Collider (EIC). Over thirty years ago, Jaffe and Manohar identified a new double-helicity flip structure function, directly sensitive to exotic gluons. They pointed out that this could be measured in inclusive high-energy electron scattering from a transversely polarized nuclear target with spin $I \ge 1$. Here, we identify the spin-3 nucleus boron-10 as a particularly interesting system to search for exotic gluons. Leveraging technical advances in atomic physics over the past decade, we outline an experimental scheme to directly optically pump a beam of stable boron atoms to polarize the nuclear spin. Technical challenges to realize a spin-polarized

arXiv:2508.06497v1 Announce Type: cross Abstract: Accurate forecasting of commodity price spikes is vital for countries with limited economic buffers, where sudden increases can strain national budgets, disrupt import-reliant sectors, and undermine food and energy security. This paper introduces a hybrid forecasting framework that combines historical commodity price data with semantic signals derived from global economic news, using an agentic generative AI pipeline. The architecture integrates dual-stream Long Short-Term Memory (LSTM) networks with attention mechanisms to fuse structured time-series inputs with semantically embedded, fact-checked news summaries collected from 1960 to 2023. The model is evaluated on a 64-year dataset comprising normalized commodity price series and temporally aligned news embeddings. Results show that the proposed approach achieves a mean AUC of 0.94 and an overall accuracy of 0.91 substantially outperforming traditional baselines such as logistic

arXiv:2508.07796v1 Announce Type: new Abstract: Heterogeneous graph neural networks (HGNNs) excel at processing heterogeneous graph data and are widely applied in critical domains. In HGNN inference, the neighbor aggregation stage is the primary performance determinant, yet it suffers from two major sources of memory inefficiency. First, the commonly adopted per-semantic execution paradigm stores intermediate aggregation results for each semantic prior to semantic fusion, causing substantial memory expansion. Second, the aggregation process incurs extensive redundant memory accesses, including repeated loading of target vertex features across semantics and repeated accesses to shared neighbors due to cross-semantic neighborhood overlap. These inefficiencies severely limit scalability and reduce HGNN inference performance. In this work, we first propose a semantics-complete execution paradigm from a vertex perspective that eliminates per-semantic intermediate storage and redundant

arXiv:2508.07628v1 Announce Type: new Abstract: The future of poultry production depends on a paradigm shift replacing subjective, labor-intensive welfare checks with data-driven, intelligent monitoring ecosystems. Traditional welfare assessments-limited by human observation and single-sensor data-cannot fully capture the complex, multidimensional nature of laying hen welfare in modern farms. Multimodal Artificial Intelligence (AI) offers a breakthrough, integrating visual, acoustic, environmental, and physiological data streams to reveal deeper insights into avian welfare dynamics. This investigation highlights multimodal As transformative potential, showing that intermediate (feature-level) fusion strategies achieve the best balance between robustness and performance under real-world poultry conditions, and offer greater scalability than early or late fusion approaches. Key adoption barriers include sensor fragility in harsh farm environments, high deployment costs, inconsistent

arXiv:2508.07023v1 Announce Type: new Abstract: Complex Visual Question Answering (Complex VQA) tasks, which demand sophisticated multi-modal reasoning and external knowledge integration, present significant challenges for existing large vision-language models (LVLMs) often limited by their reliance on high-level global features. To address this, we propose MV-CoRe (Multimodal Visual-Conceptual Reasoning), a novel model designed to enhance Complex VQA performance through the deep fusion of diverse visual and linguistic information. MV-CoRe meticulously integrates global embeddings from pre-trained Vision Large Models (VLMs) and Language Large Models (LLMs) with fine-grained semantic-aware visual features, including object detection characteristics and scene graph representations. An innovative Multimodal Fusion Transformer then processes and deeply integrates these diverse feature sets, enabling rich cross-modal attention and facilitating complex reasoning. We evaluate MV-CoRe on

arXiv:2508.06566v1 Announce Type: new Abstract: Surface material recognition is a key component in robotic perception and physical interaction, particularly when leveraging both tactile and visual sensory inputs. In this work, we propose Surformer v1, a transformer-based architecture designed for surface classification using structured tactile features and PCA-reduced visual embeddings extracted via ResNet-50. The model integrates modality-specific encoders with cross-modal attention layers, enabling rich interactions between vision and touch. Currently, state-of-the-art deep learning models for vision tasks have achieved remarkable performance. With this in mind, our first set of experiments focused exclusively on tactile-only surface classification. Using feature engineering, we trained and evaluated multiple machine learning models, assessing their accuracy and inference time. We then implemented an encoder-only Transformer model tailored for tactile features. This model not only

Physicists have heated gold to over 19,000 Kelvin, more than 14 times its melting point, without melting it, smashing the long-standing “entropy catastrophe” limit. Using an ultra-fast laser pulse at SLAC’s Linac Coherent Light Source, they kept the gold crystalline at extreme heat, opening new frontiers in high-energy-density physics, fusion research, and planetary science.

In a scientific first, South Korean scientists have provided experimental proof of "multi-scale coupling" in plasma, where interactions between phenomena at the microscopic level and macroscopic level influence each other. The findings could help advance nuclear fusion research and improve our fundamental understanding of the universe.

arXiv:2508.06116v1 Announce Type: new Abstract: We present the first gyrokinetic simulations of multiscale turbulence in a stellarator, using the magnetic geometry of Wendelstein 7-X (W7-X) and experimentally relevant parameters. A broad range of scenarios is explored, including regimes where electron-temperature-gradient (ETG) turbulence coexists with varying levels of ion-temperature-gradient (ITG) turbulence, as well as cases involving microtearing modes (MTMs) relevant to high-$\beta$ and reactor-like conditions. Notably, while ETG turbulence does not form radial streamers as in tokamaks, it can still drive significant transport and interact with ion-scale turbulence. In electrostatic ITG-dominated regimes, electron-scale fluctuations erode zonal flows, enhancing ion-scale transport, while ion-scale turbulence suppresses ETG activity. In contrast, under electromagnetic MTM conditions, the isotropic nature of ETG turbulence limits its suppressive effect, allowing MTMs to persist.

arXiv:2508.06191v1 Announce Type: new Abstract: Pleural effusion semantic segmentation can significantly enhance the accuracy and timeliness of clinical diagnosis and treatment by precisely identifying disease severity and lesion areas. Currently, semantic segmentation of pleural effusion CT images faces multiple challenges. These include similar gray levels between effusion and surrounding tissues, blurred edges, and variable morphology. Existing methods often struggle with diverse image variations and complex edges, primarily because direct feature concatenation causes semantic gaps. To address these challenges, we propose the Dual-Branch Interactive Fusion Attention model (DBIF-AUNet). This model constructs a densely nested skip-connection network and innovatively refines the Dual-Domain Feature Disentanglement module (DDFD). The DDFD module orthogonally decouples the functions of dual-domain modules to achieve multi-scale feature complementarity and enhance characteristics at

arXiv:2508.06157v1 Announce Type: new Abstract: Alzheimer's disease (AD) is a progressive neurodegenerative disorder that severely impairs cognitive function and quality of life. Timely intervention in AD relies heavily on early and precise diagnosis, which remains challenging due to the complex and subtle structural changes in the brain. Most existing deep learning methods focus only on a single plane of structural magnetic resonance imaging (sMRI) and struggle to accurately capture the complex and nonlinear relationships among pathological regions of the brain, thus limiting their ability to precisely identify atrophic features. To overcome these limitations, we propose an innovative framework, MPF-KANSC, which integrates multi-plane fusion (MPF) for combining features from the coronal, sagittal, and axial planes, and a Kolmogorov-Arnold Network-guided spatial-channel attention mechanism (KANSC) to more effectively learn and represent sMRI atrophy features. Specifically, the

arXiv:2508.06147v1 Announce Type: new Abstract: Aiming to obtain a high-resolution image, pansharpening involves the fusion of a multi-spectral image (MS) and a panchromatic image (PAN), the low-level vision task remaining significant and challenging in contemporary research. Most existing approaches rely predominantly on standard convolutions, few making the effort to adaptive convolutions, which are effective owing to the inter-pixel correlations of remote sensing images. In this paper, we propose a novel strategy for dynamically splitting convolution kernels in conjunction with attention, selecting positions of interest, and splitting the original convolution kernel into multiple smaller kernels, named DSConv. The proposed DSConv more effectively extracts features of different positions within the receptive field, enhancing the network's generalization, optimization, and feature representation capabilities. Furthermore, we innovate and enrich concepts of dynamic splitting

arXiv:2508.06146v1 Announce Type: new Abstract: Recent advancements in multimodal vision models have highlighted limitations in late-stage feature fusion and suboptimal query selection for hybrid prompts open-world segmentation, alongside constraints from caption-derived vocabularies. To address these challenges, we propose Prompt-DINO, a text-guided visual Prompt DINO framework featuring three key innovations. First, we introduce an early fusion mechanism that unifies text/visual prompts and backbone features at the initial encoding stage, enabling deeper cross-modal interactions to resolve semantic ambiguities. Second, we design order-aligned query selection for DETR-based architectures, explicitly optimizing the structural alignment between text and visual queries during decoding to enhance semantic-spatial consistency. Third, we develop a generative data engine powered by the Recognize Anything via Prompting (RAP) model, which synthesizes 0.5B diverse training instances through a

arXiv:2508.06113v1 Announce Type: new Abstract: Diffusion-based models are redefining the state-of-the-art in end-to-end autonomous driving, yet their performance is increasingly hampered by a reliance on transformer-based fusion. These architectures face fundamental limitations: quadratic computational complexity restricts the use of high-resolution features, and a lack of spatial priors prevents them from effectively modeling the inherent structure of Bird's Eye View (BEV) representations. This paper introduces GMF-Drive (Gated Mamba Fusion for Driving), an end-to-end framework that overcomes these challenges through two principled innovations. First, we supersede the information-limited histogram-based LiDAR representation with a geometrically-augmented pillar format encoding shape descriptors and statistical features, preserving critical 3D geometric details. Second, we propose a novel hierarchical gated mamba fusion (GM-Fusion) architecture that substitutes an expensive

arXiv:2508.05684v1 Announce Type: new Abstract: The proliferation of multi-modal fake news on social media poses a significant threat to public trust and social stability. Traditional detection methods, primarily text-based, often fall short due to the deceptive interplay between misleading text and images. While Large Vision-Language Models (LVLMs) offer promising avenues for multi-modal understanding, effectively fusing diverse modal information, especially when their importance is imbalanced or contradictory, remains a critical challenge. This paper introduces MM-FusionNet, an innovative framework leveraging LVLMs for robust multi-modal fake news detection. Our core contribution is the Context-Aware Dynamic Fusion Module (CADFM), which employs bi-directional cross-modal attention and a novel dynamic modal gating network. This mechanism adaptively learns and assigns importance weights to textual and visual features based on their contextual relevance, enabling intelligent

arXiv:2508.04881v1 Announce Type: new Abstract: The effect of magnetic shear on ballooning-driven plasma edge turbulence is studied through nonlinear simulations complemented by linear numerical and analytical investigations. Nonlinear, 3D, global, flux-driven simulations using the GBS code show that the scale separation between radial, x, and poloidal, y, size of turbulent eddies, kx

arXiv:2508.04210v1 Announce Type: new Abstract: Sawtooth oscillations, driven by internal kink modes (IKMs), are fundamental phenomena in tokamak plasmas. They can be classified into different types, including normal sawteeth, small sawteeth, and in some cases, evolving into the steady-island state, each having a different impact on energy confinement in fusion reactors. This study investigates the interaction between sawtooth oscillations and energetic particles (EPs) using the initial-value MHD-kinetic hybrid code CLT-K, which can perform long-term self-consistent nonlinear simulations. We analyze the redistribution of EPs caused by sawtooth crashes and the effect of EPs on sawtooth behavior and type transitions. The results show that co-passing EPs tend to re-excite sawtooth oscillations, extending their period, while counter-passing EPs promote the system evolution toward small sawteeth, potentially leading to the steady-island state. Additionally, we provide a physical picture of

arXiv:2508.03734v1 Announce Type: cross Abstract: Visual impairment represents a major global health challenge, with multimodal imaging providing complementary information that is essential for accurate ophthalmic diagnosis. This comprehensive survey systematically reviews the latest advances in multimodal deep learning methods in ophthalmology up to the year 2025. The review focuses on two main categories: task-specific multimodal approaches and large-scale multimodal foundation models. Task-specific approaches are designed for particular clinical applications such as lesion detection, disease diagnosis, and image synthesis. These methods utilize a variety of imaging modalities including color fundus photography, optical coherence tomography, and angiography. On the other hand, foundation models combine sophisticated vision-language architectures and large language models pretrained on diverse ophthalmic datasets. These models enable robust cross-modal understanding, automated

arXiv:2508.04645v1 Announce Type: new Abstract: Link Prediction (LP) is a critical task in graph machine learning. While Graph Neural Networks (GNNs) have significantly advanced LP performance recently, existing methods face key challenges including limited supervision from sparse connectivity, sensitivity to initialization, and poor generalization under distribution shifts. We explore pretraining as a solution to address these challenges. Unlike node classification, LP is inherently a pairwise task, which requires the integration of both node- and edge-level information. In this work, we present the first systematic study on the transferability of these distinct modules and propose a late fusion strategy to effectively combine their outputs for improved performance. To handle the diversity of pretraining data and avoid negative transfer, we introduce a Mixture-of-Experts (MoE) framework that captures distinct patterns in separate experts, facilitating seamless application of the

arXiv:2508.04268v1 Announce Type: new Abstract: This paper addresses the estimation of the State Of Charge (SOC) of lithium-ion cells via the combination of two widely used paradigms: Kalman Filters (KFs) equipped with Equivalent Circuit Models (ECMs) and machine-learning approaches. In particular, a recent Virtual Sensor (VS) synthesis technique is considered, which operates as follows: (i) learn an Affine Parameter-Varying (APV) model of the cell directly from data, (ii) derive a bank of linear observers from the APV model, (iii) train a machine-learning technique from features extracted from the observers together with input and output data to predict the SOC. The SOC predictions returned by the VS are supplied to an Extended KF (EKF) as output measurements along with the cell terminal voltage, combining the two paradigms. A data-driven calibration strategy for the noise covariance matrices of the EKF is proposed. Experimental results show that the designed approach is beneficial

arXiv:2508.04153v1 Announce Type: new Abstract: Enabling multi-task adaptation in pre-trained Low-Rank Adaptation (LoRA) models is crucial for enhancing their generalization capabilities. Most existing pre-trained LoRA fusion methods decompose weight matrices, sharing similar parameters while merging divergent ones. However, this paradigm inevitably induces inter-weight conflicts and leads to catastrophic domain forgetting. While incremental learning enables adaptation to multiple tasks, it struggles to achieve generalization in few-shot scenarios. Consequently, when the weight data follows a long-tailed distribution, it can lead to forgetting in the fused weights. To address this issue, we propose In-Context Meta LoRA Fusion (ICM-Fusion), a novel framework that synergizes meta-learning with in-context adaptation. The key innovation lies in our task vector arithmetic, which dynamically balances conflicting optimization directions across domains through learned manifold projections.

arXiv:2508.04123v1 Announce Type: new Abstract: Underwater image enhancement (UIE) techniques aim to improve visual quality of images captured in aquatic environments by addressing degradation issues caused by light absorption and scattering effects, including color distortion, blurring, and low contrast. Current mainstream solutions predominantly employ multi-scale feature extraction (MSFE) mechanisms to enhance reconstruction quality through multi-resolution feature fusion. However, our extensive experiments demonstrate that high-quality image reconstruction does not necessarily rely on multi-scale feature fusion. Contrary to popular belief, our experiments show that single-scale feature extraction alone can match or surpass the performance of multi-scale methods, significantly reducing complexity. To comprehensively explore single-scale feature potential in underwater enhancement, we propose an innovative Single-Scale Decomposition Network (SSD-Net). This architecture introduces an

arXiv:2508.03776v1 Announce Type: new Abstract: Estimating heat flux in the nuclear fusion device EAST is a critically important task. Traditional scientific computing methods typically model this process using the Finite Element Method (FEM). However, FEM relies on grid-based sampling for computation, which is computationally inefficient and hard to perform real-time simulations during actual experiments. Inspired by artificial intelligence-powered scientific computing, this paper proposes a novel Physics-Informed Neural Network (PINN) to address this challenge, significantly accelerating the heat conduction estimation process while maintaining high accuracy. Specifically, given inputs of different materials, we first feed spatial coordinates and time stamps into the neural network, and compute boundary loss, initial condition loss, and physical loss based on the heat conduction equation. Additionally, we sample a small number of data points in a data-driven manner to better fit the

A new class of advanced steels needs more fine-tuning before use in system components for fusion energy—a more sustainable alternative to fission that combines two light atoms rather than splitting one heavy atom. The alloy, a type of reduced activation ferritic/martensitic or RAFM steel, contains billions of nanoscale particles of titanium carbide meant to absorb radiation and trap helium produced by fusion within a single component.

arXiv:2508.03561v1 Announce Type: new Abstract: A simple model of the ramp-up and ramp-down of the toroidal current in a tokamak plasma is developed. Faraday's law of electric induction is found to limit how rapidly the current can be safety ramped up or down. It is estimated that the minimum safe ramp-up/down times for the JET, SPARC, and ITER tokamaks are 4.2, 2.0, and 14.7 seconds, respectively. The JET ramp rate is in accordance with operational experience. The SPARC and ITER minimum safe ramp rates are less than the ramp rates in the respective designs. Hence, there is no indication that the design ramp rates are infeasible, as was recently suggested in arXiv:2507.05456 (2025). The typical ratios of the inductive electric field to the Connor-Hastie field in SPARC and ITER are found to be less than those in JET. Thus, the fact that the JET tokamak was able to operate successfully without encountering runaway electron problems during current ramps suggests that the future SPARC and

arXiv:2508.02957v1 Announce Type: cross Abstract: Age-related macular degeneration (AMD) is a leading cause of irreversible vision loss, making effective prognosis crucial for timely intervention. In this work, we propose AMD-Mamba, a novel multi-modal framework for AMD prognosis, and further develop a new AMD biomarker. This framework integrates color fundus images with genetic variants and socio-demographic variables. At its core, AMD-Mamba introduces an innovative metric learning strategy that leverages AMD severity scale score as prior knowledge. This strategy allows the model to learn richer feature representations by aligning learned features with clinical phenotypes, thereby improving the capability of conventional prognosis methods in capturing disease progression patterns. In addition, unlike existing models that use traditional CNN backbones and focus primarily on local information, such as the presence of drusen, AMD-Mamba applies Vision Mamba and simultaneously fuses local

arXiv:2508.02712v1 Announce Type: cross Abstract: Modern engineering systems are increasingly equipped with sensors for real-time monitoring and decision-making. However, the data collected by these sensors is often noisy and difficult to interpret, limiting its utility for control and diagnostics. In this work, we propose a physics-informed denoising framework that integrates energy-based model and Fisher score regularization to jointly reduce data noise and enforce physical consistency with a physics-based model. The approach is first validated on benchmark problems, including the simple harmonic oscillator, Burgers' equation, and Laplace's equation, across varying noise levels. We then apply the denoising framework to real thermal emission data from laser powder bed fusion (LPBF) additive manufacturing experiments, using a trained Physics-Informed Neural Network (PINN) surrogate model of the LPBF process to guide denoising. Results show that the proposed method outperforms baseline

arXiv:2508.03598v1 Announce Type: new Abstract: Recent advancements in object detection rely on modular architectures with multi-scale fusion and attention mechanisms. However, static fusion heuristics and class-agnostic attention limit performance in dynamic scenes with occlusions, clutter, and class imbalance. We introduce Dynamic Class-Aware Fusion Network (DyCAF-Net) that addresses these challenges through three innovations: (1) an input-conditioned equilibrium-based neck that iteratively refines multi-scale features via implicit fixed-point modeling, (2) a dual dynamic attention mechanism that adaptively recalibrates channel and spatial responses using input- and class-dependent cues, and (3) class-aware feature adaptation that modulates features to prioritize discriminative regions for rare classes. Through comprehensive ablation studies with YOLOv8 and related architectures, alongside benchmarking against nine state-of-the-art baselines, DyCAF-Net achieves significant

arXiv:2508.02806v1 Announce Type: new Abstract: Recently, a significant improvement in the accuracy of 3D human pose estimation has been achieved by combining convolutional neural networks (CNNs) with pyramid grid alignment feedback loops. Additionally, innovative breakthroughs have been made in the field of computer vision through the adoption of Transformer-based temporal analysis architectures. Given these advancements, this study aims to deeply optimize and improve the existing Pymaf network architecture. The main innovations of this paper include: (1) Introducing a Transformer feature extraction network layer based on self-attention mechanisms to enhance the capture of low-level features; (2) Enhancing the understanding and capture of temporal signals in video sequences through feature temporal fusion techniques; (3) Implementing spatial pyramid structures to achieve multi-scale feature fusion, effectively balancing feature representations differences across different scales. The

arXiv:2508.01325v1 Announce Type: cross Abstract: Effective data partitioning is known to be crucial in machine learning. Traditional cross-validation methods like K-Fold Cross-Validation (KFCV) enhance model robustness but often compromise generalisation assessment due to high computational demands and extensive data shuffling. To address these issues, the integration of the Simple Random Sampling (SRS), which, despite providing representative samples, can result in non-representative sets with imbalanced data. The study introduces a hybrid model, Fusion Sampling Validation (FSV), combining SRS and KFCV to optimise data partitioning. FSV aims to minimise biases and merge the simplicity of SRS with the accuracy of KFCV. The study used three datasets of 10,000, 50,000, and 100,000 samples, generated with a normal distribution (mean 0, variance 1) and initialised with seed 42. KFCV was performed with five folds and ten repetitions, incorporating a scaling factor to ensure robust

arXiv:2508.02423v1 Announce Type: new Abstract: Histopathological analysis has been transformed by serial section-based methods, advancing beyond traditional 2D histology to enable volumetric and microstructural insights in oncology and inflammatory disease diagnostics. This review outlines key developments in specimen preparation and high-throughput imaging that support these innovations. Computational workflows are categorized into multimodal image co-registration, 3D histoarchitecture reconstruction, multiplexed immunohistochemical correlation, and cross-scale data fusion. These approaches exploit serial section-derived spatial concordance to enhance resolution in microenvironmental and molecular profiling. Despite progress, challenges remain in harmonizing heterogeneous datasets, optimizing large-scale registration, and ensuring interpretability. Future directions include spatial transcriptomics, and applications in developmental biology and neuroscience in AI integration,

arXiv:2508.02445v1 Announce Type: new Abstract: Ion Cyclotron Range of Frequencies heating (ICRH) and current drive will be essential for sustaining high-performance plasmas in next-generation fusion devices (e.g. ITER, SPARC). ICRH actuators routinely produce localized hot spots on limiters and nearby components, posing serious risks to antenna reliability, material survivability, and overall plasma performance. Remarkably, these hot spots are often strongly asymmetric, even with nominally symmetric plasma conditions and antenna geometries. We show that such asymmetries exist intrinsically in the wave physics rather than solely being due to misalignment or edge plasma variation. Our results strongly suggest that this asymmetry can be compensated for by using either poloidal phasing control (which e.g. the under construction WEST traveling wave antenna can do) or modified limiter shapes, suppressing peak sputtering by a factor $\sim$3 and reducing total erosion by a factor of $\sim$2

A research team led by Prof. Guo Bin from the Hefei Institutes of Physical Science of the Chinese Academy of Sciences has designed and optimized an organic Rankine cycle (ORC) system specifically for recovering low-grade waste heat from the steady-state Chinese Fusion Engineering Testing Reactor (CFETR) based on organic fluid R245fa, achieving enhanced thermal efficiency and reduced heat loss.

arXiv:2508.00339v1 Announce Type: new Abstract: Enhancement of the scrape-off layer (SOL) heat flux width has been observed in the ADITYA-U Tokamak following the injection of short fuel gas pulses. A notable reduction in parallel heat flux near the last closed flux surface (LCFS) is observed after each pulse. Comparative analysis indicates that pulsed fuelling is more effective in mitigating heat flux with improved core confinement than continuous gas feeding via real-time density control. Analytical and simulation works are also carried out for validation of experimental results. The analytical model shows that SOL width modification cannot be attributed solely to the decrease of temperature due to gas pulse injection; cross-field plasma diffusion also needs to increase. Simulations with the UEDGE code suggest that an increase in both the cross-field diffusion coefficient and inward pinch velocity is necessary to replicate the experimentally observed broadening of the heat flux SOL

arXiv:2505.20884v3 Announce Type: replace Abstract: Fire detection in dynamic environments faces continuous challenges, including the interference of illumination changes, many false detections or missed detections, and it is difficult to achieve both efficiency and accuracy. To address the problem of feature extraction limitation and information loss in the existing YOLO-based models, this study propose You Only Look Once for Fire Detection with Attention-guided Inverted Residual and Dual-pooling Downscale Fusion (YOLO-FireAD) with two core innovations: (1) Attention-guided Inverted Residual Block (AIR) integrates hybrid channel-spatial attention with inverted residuals to adaptively enhance fire features and suppress environmental noise; (2) Dual Pool Downscale Fusion Block (DPDF) preserves multi-scale fire patterns through learnable fusion of max-average pooling outputs, mitigating small-fire detection failures. Extensive evaluation on two public datasets shows the efficient

arXiv:2508.00248v1 Announce Type: new Abstract: Depth map super-resolution technology aims to improve the spatial resolution of low-resolution depth maps and effectively restore high-frequency detail information. Traditional convolutional neural network has limitations in dealing with long-range dependencies and are unable to fully model the global contextual information in depth maps. Although transformer can model global dependencies, its computational complexity and memory consumption are quadratic, which significantly limits its ability to process high-resolution depth maps. In this paper, we propose a multi-scale fusion U-shaped Mamba (MSF-UM) model, a novel guided depth map super-resolution framework. The core innovation of this model is to integrate Mamba's efficient state-space modeling capabilities into a multi-scale U-shaped fusion structure guided by a color image. The structure combining the residual dense channel attention block and the Mamba state space module is

arXiv:2507.22906v1 Announce Type: cross Abstract: As a green MIMO structure, the heterogeneous hybrid analog-digital H2AD MIMO architecture has been shown to own a great potential to replace the massive or extremely large-scale fully-digital MIMO in the future wireless networks to address the three challenging problems faced by the latter: high energy consumption, high circuit cost, and high complexity. However, how to intelligently sense the number and direction of multi-emitters via such a structure is still an open hard problem. To address this, we propose a two-stage sensing framework that jointly estimates the number and direction values of multiple targets. Specifically, three target number sensing methods are designed: an improved eigen-domain clustering (EDC) framework, an enhanced deep neural network (DNN) based on five key statistical features, and an improved one-dimensional convolutional neural network (1D-CNN) utilizing full eigenvalues. Subsequently, a low-complexity and

arXiv:2507.22821v1 Announce Type: new Abstract: Collisions play an important role in turbulence and transport of fusion plasmas. For kinetic simulations, as the collisionality increases in the domain of interest, the size of the time step to resolve the collisional physics can become overly restrictive in an explicit time integration scheme, leading to high computational cost. With the aim of overcoming such restriction, we have implemented an implicit Bhatnagar-Gross-Krook (BGK) collision operator for use in the discontinuous Galerkin (DG) full-f gyrokinetic solver within the Gkeyll framework, which, when combined with Gkeyll's traditional explicit time integrator for collisionless advection, can significantly increase the time step in gyrokinetic simulations of highly collisional regimes. To ensure conservation of density, momentum, and energy, we utilize an iterative scheme to correct the discretized approximation to the equilibrium Maxwellian distribution to which the BGK

arXiv:2507.22516v1 Announce Type: new Abstract: Stellarator coils are known for their complexity and departure from planarity, along with tight manufacturing tolerances in order to achieve the target magnetic field accuracy. These requirements can lead to increased costs and delays in assembly; failure to meet them can compromise the stellarator's performance. Small-scale experiments offer opportunities to develop and benchmark stellarator coil design and evaluation methods more quickly and at lower budget. In this work, we analyze precise 3D scans of the manufacturing deviations of two 3D-printed coil frames (steel, Ti alloy) and one CNC-machined coil frame (Al alloy), as part of assessing these approaches to fabricating high-temperature superconducting (HTS) coils for a tabletop stellarator. The deviations are measured along the coil length, then modeled using Gaussian processes to extract characteristic length scales. Finally a statistical study of field accuracy is performed using

arXiv:2507.22685v1 Announce Type: new Abstract: Leaf wetness detection is a crucial task in agricultural monitoring, as it directly impacts the prediction and protection of plant diseases. However, existing sensing systems suffer from limitations in robustness, accuracy, and environmental resilience when applied to natural leaves under dynamic real-world conditions. To address these challenges, we introduce a new multi-modal dataset specifically designed for evaluating and advancing machine learning algorithms in leaf wetness detection. Our dataset comprises synchronized mmWave raw data, Synthetic Aperture Radar (SAR) images, and RGB images collected over six months from five diverse plant species in both controlled and outdoor field environments. We provide detailed benchmarks using the Hydra model, including comparisons against single modality baselines and multiple fusion strategies, as well as performance under varying scan distances. Additionally, our dataset can serve as a

arXiv:2507.22426v1 Announce Type: new Abstract: Machine learning models are widely used to support stealth assessment in digital learning environments. Existing approaches typically rely on abstracted gameplay log data, which may overlook subtle behavioral cues linked to learners' cognitive strategies. This paper proposes a multimodal late fusion model that integrates screencast-based visual data and structured in-game action sequences to classify students' problem-solving strategies. In a pilot study with secondary school students (N=149) playing a multitouch educational game, the fusion model outperformed unimodal baseline models, increasing classification accuracy by over 15%. Results highlight the potential of multimodal ML for strategy-sensitive assessment and adaptive support in interactive learning contexts.

arXiv:2507.22041v1 Announce Type: new Abstract: Deep learning has witnessed the extensive utilization across a wide spectrum of domains, including fine-grained few-shot learning (FGFSL) which heavily depends on deep backbones. Nonetheless, shallower deep backbones such as ConvNet-4, are not commonly preferred because they're prone to extract a larger quantity of non-abstract visual attributes. In this paper, we initially re-evaluate the relationship between network depth and the ability to fully encode few-shot instances, and delve into whether shallow deep architecture could effectuate comparable or superior performance to mainstream deep backbone. Fueled by the inspiration from vanilla ConvNet-4, we introduce a location-aware constellation network (LCN-4), equipped with a cutting-edge location-aware feature clustering module. This module can proficiently encoder and integrate spatial feature fusion, feature clustering, and recessive feature location, thereby significantly minimizing

arXiv:2507.21489v1 Announce Type: new Abstract: Open-set 3D object retrieval (3DOR) is an emerging task aiming to retrieve 3D objects of unseen categories beyond the training set. Existing methods typically utilize all modalities (i.e., voxels, point clouds, multi-view images) and train specific backbones before fusion. However, they still struggle to produce generalized representations due to insufficient 3D training data. Being contrastively pre-trained on web-scale image-text pairs, CLIP inherently produces generalized representations for a wide range of downstream tasks. Building upon it, we present a simple yet effective framework named Describe, Adapt and Combine (DAC) by taking only multi-view images for open-set 3DOR. DAC innovatively synergizes a CLIP model with a multi-modal large language model (MLLM) to learn generalized 3D representations, where the MLLM is used for dual purposes. First, it describes the seen category information to align with CLIP's training objective

arXiv:2507.20089v1 Announce Type: cross Abstract: Developing effective multimodal data fusion strategies has become increasingly essential for improving the predictive power of statistical machine learning methods across a wide range of applications, from autonomous driving to medical diagnosis. Traditional fusion methods, including early, intermediate, and late fusion, integrate data at different stages, each offering distinct advantages and limitations. In this paper, we introduce Meta Fusion, a flexible and principled framework that unifies these existing strategies as special cases. Motivated by deep mutual learning and ensemble learning, Meta Fusion constructs a cohort of models based on various combinations of latent representations across modalities, and further boosts predictive performance through soft information sharing within the cohort. Our approach is model-agnostic in learning the latent representations, allowing it to flexibly adapt to the unique characteristics of

arXiv:2507.21003v1 Announce Type: new Abstract: Despite significant advances in reducing turbulent heat losses, modern quasi-isodynamic (QI) stellarators -- such as Stellaris -- continue to suffer from poor particle confinement, which fundamentally limits their overall performance. Using gyrokinetic simulations within the GENE--Tango framework, we identify suppressed inward thermodiffusion, caused by unfavorable magnetic geometry, as the primary cause. To overcome this limitation, we design a new configuration with a reduced mirror ratio, which enhances the contribution of passing electrons to the inward particle flux. This facilitates the formation of strongly peaked density profiles, suppresses turbulence, and leads to a substantial improvement in confinement. Our optimized configuration achieves nearly a twofold increase in energy confinement compared to Stellaris, highlighting the crucial role of optimizing particle transport in next-generation stellarator designs.

arXiv:2507.20771v1 Announce Type: new Abstract: We investigate self-consistent, steady-state axisymmetric solutions of incompressible tokamak plasma using a visco-resistive magnetohydrodynamic model. A key contribution of this work is the formulation of Poisson's equation that governs the pressure profile. Our analysis reveals that the current modeling fails to produce realistic pressure levels. To overcome this limitation, we introduce additional non-inductive current drives, akin to those generated by neutral beam injection or radio frequency heating, modeled as modifications to the toroidal current. Numerical simulations validate our enhanced model, showing significant improvements in pressure profile characteristics. In the cases examined, the effect of these current drives on the velocity profiles is moderate, except when the non-inductive current drives induce reversals in the total toroidal current density, leading to non-nested flux surfaces with internal separatrices.

arXiv:2507.20523v1 Announce Type: new Abstract: From a database of 457 experimental and numerical data from 32 machines among solid-walled tokamaks, stellarators and linear plasma devices, we derive physics-informed multi-machine scaling laws predictive of fuel and impurity puffing rates sufficient to access plasma detachment -- leading candidate for a reactor-relevant solution to the open issue of plasma-wall interaction. Validation of our laws in up to 40 plasmas in low- and high-confinement mode also featuring advanced configurations demonstrates accuracy within a factor 1.5 in up to 50% of the instances, and within a factor 2 on average. Divertor volume alone is found to correlate to fuelling. The addition of plasma opaqueness leads to the empirically-calibrated law $\Gamma_{\text{D}}\propto [n_{\text{sep}}\times a\times(S_{\text{div}}/V_{\text{div}})^{-1.5}]^{1.05}$ valid across all toroidal devices. Its simplification to $\Gamma_{\text{D}}^{\text{HDL}}\propto 0.43\times

arXiv:2507.20913v1 Announce Type: new Abstract: The rapid evolution of face manipulation techniques poses a critical challenge for face forgery detection: cross-domain generalization. Conventional methods, which rely on simple classification objectives, often fail to learn domain-invariant representations. We propose HAMLET-FFD, a cognitively inspired Hierarchical Adaptive Multi-modal Learning framework that tackles this challenge via bidirectional cross-modal reasoning. Building on contrastive vision-language models such as CLIP, HAMLET-FFD introduces a knowledge refinement loop that iteratively assesses authenticity by integrating visual evidence with conceptual cues, emulating expert forensic analysis. A key innovation is a bidirectional fusion mechanism in which textual authenticity embeddings guide the aggregation of hierarchical visual features, while modulated visual features refine text embeddings to generate image-adaptive prompts. This closed-loop process progressively

arXiv:2507.20089v1 Announce Type: new Abstract: Developing effective multimodal data fusion strategies has become increasingly essential for improving the predictive power of statistical machine learning methods across a wide range of applications, from autonomous driving to medical diagnosis. Traditional fusion methods, including early, intermediate, and late fusion, integrate data at different stages, each offering distinct advantages and limitations. In this paper, we introduce Meta Fusion, a flexible and principled framework that unifies these existing strategies as special cases. Motivated by deep mutual learning and ensemble learning, Meta Fusion constructs a cohort of models based on various combinations of latent representations across modalities, and further boosts predictive performance through soft information sharing within the cohort. Our approach is model-agnostic in learning the latent representations, allowing it to flexibly adapt to the unique characteristics of each

The alchemist's dream is to make gold from common metals, but can this be done? The physics needed to explain how to change one element into another is well understood and has been used for decades in accelerators and colliders, which smash sub-atomic particles together.

A startup energy company, called Marathon Fusion, may soon be living out the dream of alchemists from the Middle Ages. In a recently released paper posted to the arXiv preprint server, the company outlines a method to turn an isotope of mercury, 198Hg, into 197Au, the most stable form of gold.

arXiv:2507.19425v1 Announce Type: new Abstract: Diagnostics are critical for commercial and research fusion machines, since measuring and understanding plasma features is important to sustaining fusion reactions. The neutron flux (and therefore fusion power) can be indirectly calculated using neutron activation analyses, where potentially large numbers of activation foils are placed in the neutron flux, and delayed gammas from key reactions are measured via gamma spectrometry. In gamma spectrometry, absolute efficiency forms part of the activity calculation, and equals to the ratio of the total number of photons detected to the number emitted by a radioactive sample. Hence, it is imperative that they are calculated efficiently and accurately. This paper presents a novel digital efficiency calculation algorithm, the Machine Learning Based Efficiency Calculator (MaLBEC), that uses state-of-the-art supervised machine learning techniques to calculate efficiency values of a given sample,

arXiv:2507.19319v1 Announce Type: new Abstract: Recent stellarator reactor designs demonstrate mostly outward turbulent particle transport, which, without advanced fueling technology, inhibits the formation of density gradients needed for confinement. We introduce ``SQuID-$\tau$'', a self-fueling quasi-isodynamic stellarator capable of sustaining density peaking through inward particle transport caused by turbulence. Temperature and density profile predictions based on high-fidelity gyrokinetic simulations demonstrate enhanced performance, significantly relaxing constraints on the size and magnetic field strength for reactor designs.

arXiv:2507.18917v1 Announce Type: new Abstract: Turbulence enhances fusion reactivity, enabling ignition at lower temperature. A modified Lawson-like ignition criterion is derived for inertially confined plasmas harboring turbulent kinetic energy. Remarkably, if small-scale turbulence is driven in the hot spot while avoiding mixing at the boundary, less energy is required to ignite a target. The optimal length scale for hot-spot turbulence is quantified, typically lying in the micron range.

arXiv:2507.19229v1 Announce Type: new Abstract: The modeling of genomic sequences presents unique challenges due to their length and structural complexity. Traditional sequence models struggle to capture long-range dependencies and biological features inherent in DNA. In this work, we propose TrinityDNA, a novel DNA foundational model designed to address these challenges. The model integrates biologically informed components, including Groove Fusion for capturing DNA's structural features and Gated Reverse Complement (GRC) to handle the inherent symmetry of DNA sequences. Additionally, we introduce a multi-scale attention mechanism that allows the model to attend to varying levels of sequence dependencies, and an evolutionary training strategy that progressively adapts the model to both prokaryotic and eukaryotic genomes. TrinityDNA provides a more accurate and efficient approach to genomic sequence modeling, offering significant improvements in gene function prediction, regulatory