Fusion

Welcome back to our series on Interstellar Travel, where we look at the many proposals made for sending missions to the stars since the dawn of the Space Age. In our first installment, we examined how Cold War-era developments in nuclear weapons and rockets paralleled advances in space exploration, resulting in proposals for nuclear rockets. We then examined how the creation of thermonuclear weapons led to applications in fusion propulsion.

arXiv:2608.28224v1 Announce Type: cross Abstract: We present a local geometry refinement method and a 2D latent representation that took us to the top of the ConStellaration leaderboard on the geometric task in May 2026.

arXiv:2608.28367v1 Announce Type: new Abstract: Pressure gradient-driven instabilities are investigated in tokamak plasmas using the global gyrokinetic code EUTERPE emphasizing the role of moderate high mode numbers. As the normalized plasma pressure $\beta$ is increased, there is a well-known, sudden transition from ion-temperature-gradient (ITG) instabilities to kinetic ballooning modes (KBM), if the magnetohydrodynamic (MHD) geometry is held fixed. However, if the equilibrium field is recomputed for each value of $\beta$, so that the equilibrium is consistent with the stability calculation, the transition can disappear. In a number of cases, we are only able to find an ITG-KBM transition if inconsistent equilibria are used. In the MHD unstable regime, gyrokinetic simulations and MHD stability calculations show good agreement for moderate ratios of ion temperature gradient to density gradient and small values of the ion gyro-radius. Otherwise, non-MHD contributions are important,

arXiv:2608.28366v1 Announce Type: new Abstract: State estimation in magnetohydrodynamic flows is critical for real-time monitoring of liquid metal blankets in tokamak fusion reactors. Due to the multiphysics nature of these phenomena, high-fidelity simulations are computationally prohibitive for real-time applications. This work investigates a data- driven Reduced Order Model framework: the Shallow Recurrent Decoder (SHRED) coupled with Principal Component Analysis, to map sparse temperature measurements to the full thermo-hydraulic system's state. The major contribution of this work lies in the two-parameter analysis of a fully three-dimensional domain representative of the DEMO breeding blanket configuration. Here, the flow is subjected to an external magnetic field varying in direction and intensity and is hindered by two cylinders acting as a water-cooling system, which impose a temperature boundary condition on their surfaces. This double-parametric magnetic variation induces

arXiv:2608.27578v1 Announce Type: new Abstract: Training effective foundation models requires massive and organized datasets, yet scientific domains such as nuclear fusion present unique challenges due to largely heterogeneous and sparse data. Here we characterize the data used in developing such a model: with over 20 sensor types spanning 5 orders of magnitude in sampling rate, mixed tensor structures (point measurements, spectrograms, images), and nonstationary physics. We analyze our input complexity and discuss trade-offs between temporal context and frequency resolution. Our analysis provides a template for representing multi-modal fluctuation data at scale, with implications for both multi-modal control systems and nuclear fusion.

arXiv:2608.27490v1 Announce Type: new Abstract: Full-wave calculations of ion cyclotron resonance heating (ICRH) under different plasma dielectric conditions require repeated assembly and solution of large-scale discretised systems, limiting parameter sweeps and multi-case response analysis. We therefore propose an anisotropic Maxwell neural operator (AMNO) for rapid parametric modelling of ICRH full-wave responses for the Experimental Advanced Superconducting Tokamak (EAST), which learns, within the one-parameter dielectric-field family generated by varying the hydrogen minority fraction X_H over 0.01-0.05 under otherwise fixed settings, a shared solution operator from the spatially varying complex anisotropic dielectric-tensor field to the three-component complex electric field under frequency-domain Maxwell constraints. It represents global spatial coupling through spectral operator layers and local fine-scale responses, and combines sparse reference-field supervision with the

arXiv:2608.28366v1 Announce Type: cross Abstract: State estimation in magnetohydrodynamic flows is critical for real-time monitoring of liquid metal blankets in tokamak fusion reactors. Due to the multiphysics nature of these phenomena, high-fidelity simulations are computationally prohibitive for real-time applications. This work investigates a data- driven Reduced Order Model framework: the Shallow Recurrent Decoder (SHRED) coupled with Principal Component Analysis, to map sparse temperature measurements to the full thermo-hydraulic system's state. The major contribution of this work lies in the two-parameter analysis of a fully three-dimensional domain representative of the DEMO breeding blanket configuration. Here, the flow is subjected to an external magnetic field varying in direction and intensity and is hindered by two cylinders acting as a water-cooling system, which impose a temperature boundary condition on their surfaces. This double-parametric magnetic variation induces

arXiv:2608.27578v1 Announce Type: cross Abstract: Training effective foundation models requires massive and organized datasets, yet scientific domains such as nuclear fusion present unique challenges due to largely heterogeneous and sparse data. Here we characterize the data used in developing such a model: with over 20 sensor types spanning 5 orders of magnitude in sampling rate, mixed tensor structures (point measurements, spectrograms, images), and nonstationary physics. We analyze our input complexity and discuss trade-offs between temporal context and frequency resolution. Our analysis provides a template for representing multi-modal fluctuation data at scale, with implications for both multi-modal control systems and nuclear fusion.

arXiv:2608.28224v1 Announce Type: new Abstract: We present a local geometry refinement method and a 2D latent representation that took us to the top of the ConStellaration leaderboard on the geometric task in May 2026.

arXiv:2608.27912v1 Announce Type: new Abstract: Deep-research agents answer complex user questions through an iterative sequence of search steps, where the agent autonomously formulates sub-queries to retrieve the evidence needed at each stage. However, existing retriever training typically relies only on the sub-query and its corresponding search results at the current step as training signals, leaving the information accumulated from previous interactions largely underutilized. We introduce iter, an agent interaction-aware dense retriever trained using agent trajectory learning signals. iter represents each query by incorporating not only the current sub-query, but also the main question and preceding sub-queries, and is trained using trajectory-relative learning signals derived from the agent's interactions. Across six agent backbones from three model families, iter consistently outperforms the existing agent-trajectory-trained dense retriever, LRAT, achieving an average

arXiv:2608.27230v1 Announce Type: new Abstract: Over a period of 40 years, the ITER project has provided many examples in large magnet structures, from pre-concept to design to manufacture to delivery and assembly, from which lessons can be learned for the future. This presentation is divided into three parts. The first, 'Making and Assembling Large (steel) Cryogenic Structures for Magnets' is based on ITER Coil Structure experience, particularly the Central Solenoid (CS) and Toroidal Field (TF) structures and particularly on large 316LN forgings weighing about 50t individually, requiring then high accuracy machining and deep (up to 0.3m) welds with controlled distortion. New materials development is often proposed. This is easy on a laboratory scale, but often not transferable to an industrial large scale. Examples are provided. The issues to be considered (and solved) are large scale quality (and repair of defects), joining, tolerances and assembly. The second 'Designing Large Steel

arXiv:2608.26704v1 Announce Type: new Abstract: Direct magnetic field sensors can address integration drift commonly observed in conventional inductive magnetic diagnostics used in fusion systems. In this work, an AlGaN/GaN Hall-effect sensor was fabricated, packaged, and deployed inside the Helically Symmetric eXperiment (HSX)---the first quasi-helically symmetric stellarator, operating with a 1 T on-axis magnetic field and up to 200 kW of launched electron cyclotron resonance heating (ECRH) power---for in-situ magnetic field monitoring near the plasma edge. The sensor leverages the high-mobility two-dimensional electron gas (2DEG) formed in the AlGaN/GaN heterostructure for sensitive magnetic field measurement, while the wide-bandgap GaN material system provides thermal robustness for harsh-environment operation. During 68 consecutive plasma discharge shots, the sensor remained functional and produced clear transient responses associated with plasma ignition and discharge dynamics.

arXiv:2608.26216v1 Announce Type: new Abstract: The deployment of advanced, AI-enabled control algorithms in tokamak experiments requires robust integration with existing plasma control system (PCS) architectures and extensive pre-experimental validation. In this contribution, we describe the integration and testing of neural-network-emulated virtual circuits for plasma shape control within the MAST Upgrade (MAST-U) PCS environment. The neural network models predict the plasma shape using the plasma current, poloidal field coil currents, and plasma profile parameters. In this paper, we explain how they are deployed via a real-time C++ inference server that interfaces with the PCS, returning the shape prediction and its Jacobian, and how, from the latter, virtual circuit matrices and updated coil current requests are computed for real-time actuation. Emphasis is placed on the validation workflow and best practices adopted to ensure confidence in the proposed control framework prior to

arXiv:2608.26216v1 Announce Type: cross Abstract: The deployment of advanced, AI-enabled control algorithms in tokamak experiments requires robust integration with existing plasma control system (PCS) architectures and extensive pre-experimental validation. In this contribution, we describe the integration and testing of neural-network-emulated virtual circuits for plasma shape control within the MAST Upgrade (MAST-U) PCS environment. The neural network models predict the plasma shape using the plasma current, poloidal field coil currents, and plasma profile parameters. In this paper, we explain how they are deployed via a real-time C++ inference server that interfaces with the PCS, returning the shape prediction and its Jacobian, and how, from the latter, virtual circuit matrices and updated coil current requests are computed for real-time actuation. Emphasis is placed on the validation workflow and best practices adopted to ensure confidence in the proposed control framework prior

arXiv:2608.26357v1 Announce Type: new Abstract: Large language models (LLMs) exhibit uneven multilingual performance, especially when dealing with low-resource languages. Inference-time intervention offers a lightweight way to improve cross-lingual transfer by modifying the hidden states produced by the LLMs during the forward pass, without updating model parameters. However, existing cross-lingual intervention methods typically learn separate projections from source to target languages, which limits scalability and prevents knowledge sharing across languages. We propose Centroid Intervention Fusion (CIF), a projection fusion framework that consolidates multiple multilingual intervention projections into a single language-shared operator. Across multilingual commonsense reasoning, natural language inference, factual editing, and machine translation benchmarks, CIF outperforms the strongest prior pairwise intervention baseline by up to +3.378 pp on average across four model backbones,

arXiv:2608.27295v1 Announce Type: new Abstract: This paper develops a unified framework for assessing systemic risk and identifying contagion channels in the global banking system using a Temporal Heterogeneous Multiplex Graph Neural Network. We construct a harmonised quarterly panel combining bank fundamentals, CDS spreads, and macroeconomic indicators, and represent these data as dynamic multiplex networks linking banks through financial similarity and liquidity co-movement, augmented with country-level macroeconomic relationships. The model integrates graph convolutional layers with recurrent GRU dynamics and incorporates a learnable fusion gate to capture time-varying reliance on alternative contagion channels. Empirical results show that the framework outperforms conventional econometric, machine learning, and graph-based benchmarks for short-term changes in CDS spreads. Beyond forecasting, we provide an interpretable framework to quantify bank-level systemic importance via

arXiv:2608.26024v1 Announce Type: new Abstract: A compact Spherical Tokamak(ST) is commissioned at Institute for Plasma Research (IPR) to explore low aspect ratio tokamak physics and technologies that complement to the existing high aspect ratio tokamaks namely ADITYA-U and SST-1 by enabling studies on non-inductive startup, current drive in over dense plasmas, and shaped plasma physics on a low cost platform. The device, India's first spherical tokamak has completed major mechanical, magnetic, and electrical integration, and the coil system has been successfully tested with series of integrated commissioning. First plasma experiments have been carried out with a modest Ohmic system assisted by a 2.45GHz microwave system, supported by a centralized control and data acquisition system. An initial diagnostic set comprising visible imaging, spectroscopy, magnetics, and radiation monitors required for machine operation has been installed. This paper presents the integrated commissioning

arXiv:2608.25945v1 Announce Type: new Abstract: Gyrokinetic (GK) stability strongly influences the performance of high-confinement-mode pedestals in spherical tokamak plasmas. High-fidelity gyrokinetic codes such as GENE can model microinstability-driven transport, but the computational cost limits their routine use in integrated pedestal modeling workflows. Instead, present workflows often rely on reduced transport assumptions, such as the ballooning-critical pedestal model used in EPED. This work investigates machine-learning surrogate models for local linear gyrokinetic simulations in a MAST-U-relevant pedestal parameter space, with the aim of providing faster gyrokinetic-based inputs to reduced pedestal models. A sampling workflow is developed in which pedestal profile parameters are varied within experimentally motivated bounds and used to generate physically self-consistent Grad-Shafranov equilibria. This reduces the dimensionality of the data-generation problem compared with

arXiv:2608.25955v1 Announce Type: new Abstract: Autonomous R\&D agents now write, run, and improve executable artifacts under automated evaluation---but largely as laboratory instruments: shown on curated benchmarks, with gains that are hard to trace to a cause and costs well above what sustained engineering practice absorbs. The limitation is structural. Most systems treat each attempt as nearly self-contained, so logs, memories, and search trees record what happened without establishing which design element produced an improvement, whether its evidence survived validation, or how it recombines with others. Long campaigns therefore keep re-learning the same lessons. We introduce Praxist, a lineage-centered generational system that converts reproducible artifacts and evaluator outcomes into a typed evidence graph of findings, lane-structured frontiers, and agendas. Separating local artifact construction from cohort-level evidence synthesis lets later attempts inherit validated

arXiv:2608.25061v1 Announce Type: new Abstract: GPUs increasingly accelerate database systems, but query-specific peak performance still often relies on hand-written kernels. Existing LLM kernel benchmarks focus on machine learning operators, leaving irregular, heterogeneous, data-movement-heavy database-style operators untested. We introduce DataKernelBench, which translates SQL into validated PyTorch TorchPlan programs and evaluates LLMs that optimize either the core tensor-bounded snippet or the full query in CUDA or Triton through execution-guided repair. Across ten proprietary and open-weight models on TPC-H SF10 with an H100 GPU, the strongest full-query CUDA configuration achieves $2.11\times$ speedup over torch.compile at full pass rate. We find that higher-performing implementations commonly use kernel fusion and execution-strategy changes, stronger models benefit most from full-query specialization, and workload context matters more than hardware context. To handle data

Iran has unveiled and commissioned a new set of indigenous capabilities in its nuclear industry.

Some of the strangest weather in the solar system doesn’t happen on Earth, or even Jupiter’s Great Red Spot - it happens in the interior of the Ice Giants like Neptune and Uranus. Specifically, scientists have long believed that, at certain pressure and temperatures, it literally rains diamonds inside of these planets. And for the first time, scientists have mimicked the process they believe creates that. A new paper by physicists at the Lawrence Livermore National Laboratory (LLNL), published in Nature Physics, resolves a 20 year old scientific mystery, and shows how the same physics that makes it rain diamonds inside Neptune could also help us triple our fusion energy output.

arXiv:2608.23976v1 Announce Type: new Abstract: A new approach to tokamak magnetic control enabling high-precision plasma shaping and novel real-time adaptability is experimentally demonstrated on the Tokamak a Configuration Variable (TCV). The method is motivated by the insight that, under appropriate assumptions, a real-time inverse Grad-Shafranov solver approximates an optimal control policy for plasma boundary regulation. Building on this, a control architecture is developed in which classical controllers enforce operational constraints while a fast surrogate model provides a real-time inverse mapping from the desired plasma boundary to Poloidal Field Coil currents. Experimental results on TCV demonstrate improved plasma shaping with respect to the standard discharge preparation procedure --- albeit without explicit real-time shape feedback --- while enabling flexible response to asynchronous events. It is shown that a single network provides satisfactory performance across a

arXiv:2608.23604v1 Announce Type: new Abstract: Laser-induced-desorption quadrupole-mass-spectrometry (LID-QMS) diagnostics is considered as one of the candidate methods for the remote control of tritium inventory in the ITER first wall. Studies involving LID-QMS generally assume the circular shape of the laser spot on the analyzed surface. At the same time, the diagnostics laser source cannot be always positioned so as to irradiate tokamak tiles under normal angles, which results in the laser spot shape differing from the circular one. In this contribution, we analyze the tritium removal process under sample irradiation by an elliptic Gaussian laser beam, extending the results of our previous analysis [Stepanenko, Gasparyan, Physica Scripta 99 (8), 085604 (2025)]. The thermal desorption model governing the heat transport and tritium removal from the solid is formulated. The new analytical expression describing the sample temperature dynamics is derived. The developed model is used to

arXiv:2608.24795v1 Announce Type: new Abstract: Recently, the rapid advancement of multimodal domains has driven a data-centric paradigm shift in graph ML, transitioning from text-attributed to multimodal-attributed graphs. This advancement significantly enhances data representation and expands the scope of graph downstream tasks, such as modality-oriented tasks, thereby improving the practical utility of graph ML. Despite its promise, limitations exist in the current neural paradigms:(1) Neglect Context in Modality Alignment: Most existing methods adopt topology-constrained or modality-specific operators as tokenizers.These aligners inevitably neglect graph context and inhibit modality interaction, resulting in suboptimal alignment.(2) Lack of Adaptation in Modality Fusion: Most existing methods are simple adaptations for 2-modality graphs and fail to adequately exploit aligned tokens equipped with topology priors during fusion, leading to poor generalizability and performance

arXiv:2608.24073v1 Announce Type: new Abstract: Low-earth-orbit (LEO) satellites enable high-resolution, large-scale Earth observation for applications such as disaster monitoring and environmental surveillance. However, cloud coverage often obscures the Earth's surface, and conventional cloud-removal pipelines that download cloudy images to ground stations for processing suffer from limited contact windows, constrained satellite-to-ground bandwidth, and high latency. In this work, we propose a novel satellite federated learning framework for cloud removal across LEO constellations, named orbital attention leaky integrate-and-fire (OrbitALIF). OrbitALIF performs both onboard training and inference using a compact 2.30,M-parameter spiking neural network (SNN) backbone with an adaptive gated fusion module (AGFM) and a spectral-spatial hybrid attention module (SHAM), combined with a decentralized federated learning strategy that shares model weights via inter-satellite links. Our

arXiv:2608.23726v1 Announce Type: new Abstract: Rapid and accurate fault detection in high-voltage transmission networks is essential for grid reliability and equipment protection. Transmission fault datasets are frequently imbalanced, and certain fault types produce electrical signatures that fall within the normal operating envelope, causing single-model classifiers to fail on safety-critical cases. This paper proposes a hybrid two-stage machine learning pipeline that decouples detection from classification. Stage 1 combines an Isolation Forest anomaly detector with an optional supervised binary detector through an OR-fusion rule; the supervised branch is allocated automatically during training for any fault class the anomaly detector cannot resolve, and is omitted when no such class exists. Stage 2 applies a Random Forest multiclass classifier only to samples flagged by Stage 1. Feature engineering is expressed as a per-measurement-point operator mapping six raw channels to

arXiv:2608.23217v1 Announce Type: new Abstract: Fast and reliable plasma equilibrium prediction is essential for real-time tokamak operation and control, but conventional Grad-Shafranov (GS) solvers are often too costly for real-time deployment. We develop an AI surrogate framework and benchmark five architectures (MLP, CNN, FNO, Transformer, and KAN) on a numerical GS database with 100,000 IID and 10,000 OOD samples. Under a unified protocol, we evaluate accuracy, inference efficiency, model scaling, and robustness. We also establish device-level validation on the EXL-50U tokamak by linking numerical GS solutions, surrogate predictions, and the standard Shape Editor reference to assess simulation-to-device consistency. The surrogates achieve errors of $10^{-3}$-$10^{-2}$ relative to GS solutions, while the GS-to-device discrepancy remains at $10^{-3}$. Transformer gives the best IID accuracy, whereas CNN offers the best balance of accuracy, robustness, and speed, reaching 0.7 ms

arXiv:2608.22677v1 Announce Type: new Abstract: In this work, we propose a model-based feedback controller that regulates the vertical instability growth rate ($\gamma_{gr}$) of a high-elongation, double-null tokamak directly, using only out-vessel poloidal field (PF) coils. High elongation raises the achievable plasma current and fusion performance but makes the plasma vertically unstable, and in a fusion power plant the in-vessel coils that present devices rely on for stabilization may be absent, leaving only distant out-vessel circuits. The controller couples a machine learning surrogate of non-rigid, profile agnostic vertical instability metric to a constrained quadratic program: the surrogate supplies real-time $\gamma_{gr}$ estimates and, via automatic differentiation, the actuator sensitivities, while the program allocates coil voltages to track a target growth rate, maintain double-null divertor balance, and respect electromechanical limits. We tested this method on the

arXiv:2608.22515v1 Announce Type: new Abstract: Reliable early disruption prediction is critical for the safe operation and real-time control of tokamaks. However, machine learning based prediction frameworks have predominantly targeted medium and long pulse devices, with comparatively limited attention given to short pulse tokamaks where available warning time is inherently constrained. In this work, an interpretable machine learning framework is developed for feature engineering and early prediction of disruptions in the ADITYA using the initial plasma evolution information, prior to the activation of the negative converter of the ohmic transformer power supply. Statistical descriptors comprising the mean, variance, skewness, kurtosis and wavelet energy entropy are extracted from routinely available plasma diagnostics over different operation time windows. Decision tree based feature selection is employed to identify physically meaningful disruption precursors and to reduce feature

arXiv:2608.21611v1 Announce Type: new Abstract: We introduce an Absolute eXtended UltraViolet (AXUV) diode-based camera forward-modelling tool to support the validation of mitigated disruption simulations and the interpretation of experimental phenomena, with applications to the ASDEX Upgrade (AUG) tokamak. AXUV diodes measure electromagnetic radiation across a wide spectral range with a significantly higher time resolution (~microseconds) than foil bolometers (~milliseconds), albeit with a non-uniform spectral responsivity. AXUV is suitable for examining fast phenomena, such as shattered pellet injection (SPI), where the radiation localisation and radiated power provide information on the deposition of pellet material. Due to the characteristics and degradation of AXUV diodes, absolute power measurements are subject to large systematic uncertainties, especially when the spectra are time-varying, as in e.g. mixed Ne/D2 SPI experiments. These challenges motivated the development of a

arXiv:2608.23217v1 Announce Type: cross Abstract: Fast and reliable plasma equilibrium prediction is essential for real-time tokamak operation and control, but conventional Grad-Shafranov (GS) solvers are often too costly for real-time deployment. We develop an AI surrogate framework and benchmark five architectures (MLP, CNN, FNO, Transformer, and KAN) on a numerical GS database with 100,000 IID and 10,000 OOD samples. Under a unified protocol, we evaluate accuracy, inference efficiency, model scaling, and robustness. We also establish device-level validation on the EXL-50U tokamak by linking numerical GS solutions, surrogate predictions, and the standard Shape Editor reference to assess simulation-to-device consistency. The surrogates achieve errors of $10^{-3}$-$10^{-2}$ relative to GS solutions, while the GS-to-device discrepancy remains at $10^{-3}$. Transformer gives the best IID accuracy, whereas CNN offers the best balance of accuracy, robustness, and speed, reaching 0.7 ms

arXiv:2608.22515v1 Announce Type: cross Abstract: Reliable early disruption prediction is critical for the safe operation and real-time control of tokamaks. However, machine learning based prediction frameworks have predominantly targeted medium and long pulse devices, with comparatively limited attention given to short pulse tokamaks where available warning time is inherently constrained. In this work, an interpretable machine learning framework is developed for feature engineering and early prediction of disruptions in the ADITYA using the initial plasma evolution information, prior to the activation of the negative converter of the ohmic transformer power supply. Statistical descriptors comprising the mean, variance, skewness, kurtosis and wavelet energy entropy are extracted from routinely available plasma diagnostics over different operation time windows. Decision tree based feature selection is employed to identify physically meaningful disruption precursors and to reduce

arXiv:2608.22500v1 Announce Type: new Abstract: At the heart of composed visual data retrieval is the fusion of a reference visual input and a textual modification into a single query. While current state-of-the-art methods utilize multimodal large language models for this fusion, their complexity introduces prohibitive querytime latency, limiting their scalability. We instead revisit the efficacy of simple linear interpolation within an embedding space, and introduce SRAIN, the first framework that dynamically predicts query-specific interpolation weights. The key challenge lies in the fact that the quality of an interpolation weight should be measured by the interpolated embedding's discriminability from negatives as well as its proximity to true targets; this makes collecting and predicting optimal weights intractable. We overcome this bottleneck through two key innovations: batch-wise rank-aware weight estimation during training, and a compact memory bank that synthesizes hard

arXiv:2608.22370v1 Announce Type: new Abstract: Task-specific LoRA adapters offer a modular way to specialize large language and vision-language models. However, existing adapter composition methods are mostly static and cannot adapt to individual test inputs. To address these issues, we propose \textbf{LiST}, a label-free test-time LoRA fusion framework that converts an existing LoRA bank into a target-conditioned local simplex and searches sample-specific fusion weights at inference time. LiST builds joint task representations from LoRA parameter anchors and prompt-level behavior vectors, retrieves neighboring adapters as a local search space, and performs branch-preserving fusion without updating the backbone or adapters. Candidate weights are selected by a prompt-level energy with prior, geometric, and stochastic-consistency constraints, and are deployed only when they pass a safe acceptance rule. Otherwise, LiST falls back to a target-conditioned prior. Experiments on multimodal

arXiv:2608.22023v1 Announce Type: new Abstract: Context-adaptive Kalman filters calibrate their noise covariance matrices Q and R from innovation residuals via online regression. When the underlying sensor or signal carries periodic structure -- mechanical LiDAR rotation harmonics, engine vibration, ground multipath, weekly and annual demand cycles, dosing-interval rhythms, weekly media-buying cadence -- the regression input is contaminated and the fitted covariance models structural modes rather than genuine state uncertainty. We introduce a four-role FFT pre-filter that solves this problem at $O(N\log N)$ cost and serves three additional roles "for free": (i) it whitens coloured noise before the Kalman update, restoring the optimality assumption; (ii) it cleans innovations before covariance regression, preventing periodic contamination of $\hat{R}$ and $\hat{Q}$; (iii) it generates spectral context features that enrich the downstream bandit's regime-selection state; (iv) it

arXiv:2608.21713v1 Announce Type: new Abstract: Reasoning-augmented text-to-image models such as GoT-R1 emit an explicit textual plan - object names, attributes, and bounding boxes - before generating image tokens. When such a model fails a compositional prompt, is the plan wrong, or is the plan right and the decoder unfaithful? Because the plan is machine-readable it can be edited before decoding, which makes the two separable. We first validate the ruler. Swapping the two bounding boxes inside the model's own chain demonstrably flips the generated layout: detector-based accuracy falls 0.75 -> 0.48 (p

Researchers at Lawrence Livermore National Laboratory (LLNL) have found that implosions designed for inertial fusion energy (IFE) can tolerate significant imperfections before performance abruptly declines, a finding that could inform the design of fuel targets for future fusion power plants.

arXiv:2404.05229v1 Announce Type: cross Abstract: The evaluation of modelled or satellite-derived soil moisture (SM) estimates is usually dependent on comparisons against in-situ SM measurements. However, the inherent mismatch in spatial support (i.e., scale) necessitates a cautious interpretation of point-to-pixel comparisons. The upscaling of the in-situ measurements to a commensurate resolution to that of the modelled or retrieved SM will lead to a fairer comparison and statistically more defensible evaluation. In this study, we presented an upscaling approach that combines spatiotemporal fusion with machine learning to extrapolate point-scale SM measurements from 28 in-situ sites to a 100 m resolution for an agricultural area of 100 km by 100 km. We conducted a four-fold cross-validation, which consistently demonstrated comparable correlation performance across folds, ranging from 0.6 to 0.9. The proposed approach was further validated based on a cross-cluster strategy by using

arXiv:2608.21179v1 Announce Type: new Abstract: It is shown that the ELM energy loss normalized by the plasma stored energy ({\Delta}EELM/Wplasma) for high-density small/QCE ELM regimes scales inversely with the separatrix turbulence parameter a_t. In contrast, the neoclassical electron collisionality at the pedestal top, nu*e,neo, expected to regulate {\Delta}EELM/Wplasma according to the Loarte scaling (Plasma Phys. Control. Fusion 2003 45 1549), does not adequately capture {\Delta}EELM/Wplasma data for peeling-ballooning-limited type-I ELMs and ballooning-limited small/QCE ELMs, limiting its applicability for extrapolation to one scenario window. A multi-machine database including seven tokamaks and with {\Delta}EELM/Wplasma ranging from 0.5% to 14%, has been analyzed. A regression analysis on only type-I ELMs yields (({\Delta}E_ELM)/W_plasma )_(Type-I ) [%]=6.8*T_(e,ped)^0.03 n_(e,ped)^(-0.4) \k{appa}^(-0.4) R_major^0.4, corresponding to {\Delta}EELM/Wplasma =4.5% for nominal

arXiv:2608.20901v1 Announce Type: new Abstract: Vertical position control is essential for sustaining high-performance operation in spherical tokamaks, where increased plasma elongation introduces stringent requirements on fast and robust stabilization. This work presents an experimentally validated reinforcement-learning(RL)-based vertical position control framework for the EXL-50U spherical tokamak. A high-fidelity discharge-reconstructed simulation environment is developed by integrating physics-based plasma-circuit models with experimental equilibrium information, enabling systematic controller synthesis and sim-to-real evaluation. Within this framework, RL is benchmarked in simulation against operational proportional--integral--derivative (PID) and model-based linear quadratic regulator (LQR) controllers under identical plant dynamics, actuator constraints, and measurement imperfections.Simulation results show that RL achieves tracking accuracy comparable to PID with consistently

arXiv:2608.20684v1 Announce Type: new Abstract: Reliable prediction of the energy confinement time is essential for magnetic-confinement fusion. Conventional power-law scalings provide constrained extrapolation trends but cannot represent complex nonlinearities, whereas neural networks interpolate accurately but may behave unpredictably outside the training distribution. We propose a unified power-law-anchored residual-learning framework in which a frozen empirical power-law scaling supplies the global trend and a nonlinear model learns only the systematic residual in logarithmic space. PLR-KAN is developed as the primary implementation, while a parameter-matched PLR-MLP serves as a controlled architecture replacement. Using the ITPA DB5.2.3 H-mode confinement database, we evaluate interpolation and parameter-defined held-out cohorts over ten complete training pipelines. PLR-KAN retains near-best interpolation accuracy, achieving R2=0.9671+/-0.0027, while substantially improving the

arXiv:2608.20652v1 Announce Type: new Abstract: A zero-dimensional (0D) multi-physics-constrained framework for parameter design and optimization of Stable Quasi-Isodynamic Designs (SQuIDs) is presented. Single- and multi-objective optimizations for three staged devices are carried out using an in-house stellarator 0D systems code: YF-1 for discharge demonstration, YF-2 for scientific break even, and YF-3 for a commercial demonstration plant. Pareto searches map the main design trade-offs across the three generations. The equal weight optima for YF-2 and YF-3 both lie in the electron-root favorable regime of the adopted root proxy: YF-2 recovers $Q_{phys} \sim 1$, while YF-3 reaches an ignited point at reactor scale. Future work will couple engineering feasibility and economic assessment modules for integrated plant evaluation.

arXiv:2608.21136v1 Announce Type: new Abstract: Recently, open-vocabulary zero-shot 3D scene understanding using vision foundation models has emerged as a promising alternative to data-intensive supervised methods. However, deploying these models in real-world scenarios is severely hindered by their inability to efficiently handle streaming RGB-D inputs and their inherent vulnerability to noise 2D segmentation masks. To address these critical limitations, we propose Stream3Dv2, a novel training-free framework designed for robust streaming 3D perception. Stream3Dv2 processes sequential data through an original nested local-to-historical architecture, capturing multi-view consistency while circumventing the high computational overhead so as to support timely responses. At its core, we introduce a comprehensive geometric-semantic fusion mechanism that resolves geometric noise and semantic ambiguity by explicitly utilizing semantic guidance and formulating 3D segmentation as solving

arXiv:2608.20260v1 Announce Type: new Abstract: Spatial statistics has grown from kriging for spatial prediction into a broad framework for learning from complex dependent data. This article traces that development from random fields and spectral methods to Bayesian hierarchical models and scalable computation. It then connects these foundations to Spatial AI, where graph learning and neural networks are being adapted to spatially dependent data. The article introduces the main ideas behind kriging and nonstationarity and explains how data fusion and uncertainty quantification extend spatial inference to more complex settings. The central contribution is a unified account of how these developments lead naturally to new forms of Spatial AI. Rather than treating spatial statistics and machine learning as separate traditions, we show how both learn from dependence while preserving interpretable structure. We also examine how spatial geometry and physical knowledge can guide flexible

arXiv:2608.20271v1 Announce Type: new Abstract: The rapid proliferation of memecoins on blockchain platforms has increased the risk of fraudulent activities, particularly rug pulls. While previous studies have focused on Ethereum-based tokens, this paper shifts the spotlight to Solana, the leading blockchain for memecoins by trading volume and token count. Unlike Ethereum, where rug pulls often exploit smart contract backdoors, Solana memecoin rug pulls are predominantly driven by liquidity manipulation and social dynamics. This research pioneers large-scale rug pull early detection in the Solana ecosystem by assembling a dataset of 6.4 million tokens over 7 months. Market analysis reveals that a vast majority of these memecoins exhibit rug pull characteristics within one hour of launch, highlighting the urgency of short-horizon prediction. Despite the absence of code-level features, we demonstrate that classic machine learning models, particularly Gradient Boosting (XGBoost), achieve

Less than a year after earning his doctoral degree in chemical and biomolecular engineering from Rice University, Thiago

Welcome back to our series on interstellar travel! In our first installment, we examined attempts to realize nuclear propulsion and how the technology could be used to reach the nearest star. In our second, we examined how fusion power has also been considered a means of propelling spacecraft to relativistic speeds (a fraction of the speed of light). In both cases, these proposals paralleled Cold War developments in rocketry and nuclear armaments, as they did for most space age advancements.

Extreme experiments have revealed how diamond behaves at crushing pressures beyond those inside Neptune and Uranus, resolving a decades-old conflict between theory and observation. The results could help scientists boost fusion energy output while revealing more about the exotic diamond rain hidden inside ice giant planets.

arXiv:2608.18939v1 Announce Type: new Abstract: Alternative divertor configurations (ADCs) must be evaluated under boundary plasma conditions approaching reactor-level values to be considered a reliable, physics-based solution for tokamak power exhaust. Most ADC experiments performed to date were at relatively low exhaust power. This work presents a high-power scenario on the TCV tokamak enabling the study of a wide variety of divertor magnetic shapes under an expanded SOL and power exhaust parameter space. The scenario is characterized by high power levels of electron cyclotron resonance heating ($2.5\,\text{MW}$ fully absorbed in a $\sim1\,\text{m}^{3}$ plasma) at high plasma current (edge safety factor $q_{95}\approx 2.5$), and low upstream separatrix densities ($n_{e,\text{u}}\approx1\times10^{19}\,\text{m}^{-3}$, Greenwald fraction $f_{\text{G}}\approx 0.1$). Stationary parallel heat fluxes up to $100\,\text{MW m}^{-2}$ are measured at the divertor target, an order of magnitude

arXiv:2608.18325v1 Announce Type: new Abstract: The energy confinement time is a key parameter of a magnetized fusion plasma, helping to determine whether ignition can occur. Experiments in tokamaks and stellarators have shown that the confinement time can be improved via pellet injection. The state of enhanced confinement brought about by a given pellet typically deteriorates over time unless and until a subsequent pellet is injected. In this work, we develop a data-driven model that predicts, at any moment, the remaining time before a plasma in Wendelstein 7-X (W7-X) will lose its enhanced confinement state. This "remaining time" metric effectively sets a deadline for when the next pellet must be injected in order to steadily maintain a high confinement time. We describe the development and training of the model and compare its predictions to observations from previous experiments. At least 90% of the model predictions are accurate to within 51 ms, which is below the typical W7-X

arXiv:2608.19147v1 Announce Type: new Abstract: Modern Intel AI PCs ship capable integrated GPUs and NPUs with 16+ GB of unified memory, and they spend considerable time idle. That is not enough memory to fit a large model such as a 70B-parameter LLM. We show that a handful of AIPCs, working together over an ordinary network, can serve models beyond the capability of any single one. We use pipeline parallelism: a model is split by layer into per-stage shards, each pre-compiled into an OpenVINO graph, so that every machine runs one shard and passes activations to the next. Three techniques make this fast enough to be useful. First, we recover the speed of the unsplit model: a naive per-stage export runs well below monolithic inference because it misses an OpenVINO GPU optimization, and injecting a beam_idx Gather into each shard triggers that optimization (the IndirectKVCache fusion) and brings the shards to parity. Second, we leverage speculative decoding on stateful OpenVINO models.

arXiv:2608.18080v1 Announce Type: new Abstract: We present a review on the applications of large language models (LLMs) in health, e.g., social media analysis, clinical conversational agents, therapy support tools, prompt engineering, multimodal learning, and ethical considerations. We integrate findings from interdisciplinary studies utilizing diverse data sources such as social media posts, electronic medical records, and multimodal inputs to enable early detection of depression, suicide risk assessment, personalized therapy support, and psychoeducational content generation. Our review highlights advancements in LLM models and annotation strategies that enhance interpretability and clinical relevance, while we also emphasize the critical role of prompt engineering for domain adaptation. We also discuss emerging multimodal fusion techniques integrating text, speech, and sensor data for improved mental health diagnosis and monitoring. Finally, we address ongoing ethical,

In the world of fusion energy, scientists and engineers study the fourth state of matter known as plasma in an effort to design and build a new type of power plant. Relying on the heat produced by two small atoms smashing together, a network of such facilities would help create a novel source of stable electricity and help ensure America's energy independence. And while scientists in this endeavor are devoting their attention to complex machinery and temperatures hotter than the surface of the sun, they are also trying to determine the best designs for such a power plant by focusing on geometry.

arXiv:2608.17248v1 Announce Type: cross Abstract: When computational models (either physics-based or data-driven) are used for the sensitivity analysis of engineering systems, the sensitivity estimate is affected by the accuracy and uncertainty of the model. This paper considers global sensitivity analysis (GSA) for situations where both a physics-based model and experimental observations are available, and investigates physics-informed machine learning strategies to effectively combine the two sources of information in order to maximize the accuracy of the sensitivity estimate. Two representative machine learning (ML) techniques are considered, namely, deep neural networks (DNN) and Gaussian process (GP) modeling, and two strategies for incorporating physics knowledge within these techniques are investigated, namely: (i) incorporating loss functions in the ML models to enforce physics constraints, and (ii) pre-training and updating the ML model using simulation and experimental data

arXiv:2608.17839v1 Announce Type: new Abstract: Magnetic reconnection governs the explosive release of magnetic energy in systems from the solar corona to fusion plasmas, yet controlling it in the laboratory has remained out of reach. Here we demonstrate active control of reconnection in high-power laser-driven plasmas using a third, relativistic-intensity laser pulse that injects filaments of electron current into the reconnecting system. Two moderate-intensity lasers drive colliding magnetized plumes that reconnect, forming plasmoids in the current sheet as seen in proton deflectometry. The relativistic laser generates magnetic fields matching the polarity on either side of the layer, and, depending on its arrival time, either accelerates the breakup of the current sheet or suppresses reconnection. Arriving early, before the plumes strongly interact, it builds a pocket of magnetic pressure that repels them via flux pileup; arriving after the current sheet forms, it accelerates

arXiv:2608.17418v1 Announce Type: new Abstract: Fast analysis of microscopic drift-wave instabilities based on linear gyrokinetic simulation is desirable for modeling anomalous transport in fusion device. In this work, we present an orbit-invariant decomposition method for solving collisionless gyrokinetic eigenvalue problems. By discretizing velocity space along orbit invariants using particle energy and magnetic moment, the full eigenvalue matrix is separated into independent orbit blocks that couple with each other through the field equation, greatly reducing both matrix dimension and computational cost without sacrificing physics. Based on this method, we extend the MGK code [Phys.\ Plasmas 24, 072106 (2017)] with both CPU and GPU implementations, supporting collisionless electrostatic linear simulations in $s$--$\alpha$ and Miller equilibrium model with kinetic. For kinetic ion temperature gradient (ITG) and trapped electron mode (TEM) eigenvalue problems, the solver reduces

arXiv:2608.16938v1 Announce Type: new Abstract: Stellarator design explores a vast space of three-dimensional plasma boundaries, only a small fraction of which yields usable equilibria. Data-driven models can narrow this search by learning from existing optimized configurations. Building on the ConStellaration database, we extend conditional boundary generation to four-field-period QI configurations, focusing on the sparsely sampled low-aspect-ratio regime. The approach learns the common geometric structure of known QI equilibria and then adapts it using a small high-fidelity compact dataset, allowing target magnetic properties to guide generation beyond the original data distribution. This adaptation reduces the compact-domain test loss by approximately 87% and yields converged ultra-compact candidates consistent with the prescribed conditions. Several candidates show favorable confinement indicators, and one provides a useful seed for further QI optimization and finite-beta

arXiv:2608.17248v1 Announce Type: new Abstract: When computational models (either physics-based or data-driven) are used for the sensitivity analysis of engineering systems, the sensitivity estimate is affected by the accuracy and uncertainty of the model. This paper considers global sensitivity analysis (GSA) for situations where both a physics-based model and experimental observations are available, and investigates physics-informed machine learning strategies to effectively combine the two sources of information in order to maximize the accuracy of the sensitivity estimate. Two representative machine learning (ML) techniques are considered, namely, deep neural networks (DNN) and Gaussian process (GP) modeling, and two strategies for incorporating physics knowledge within these techniques are investigated, namely: (i) incorporating loss functions in the ML models to enforce physics constraints, and (ii) pre-training and updating the ML model using simulation and experimental data

arXiv:2608.15753v1 Announce Type: new Abstract: Runaway electron (RE) generation is of great concern for high-current tokamak operations. A full-f particle-in-cell (PIC) model for RE dynamics has been developed and coupled with the 3D nonlinear extended magnetohydrodynamic (MHD) model implemented in the NIMROD code. Our model accounts for RE generation using analytical source terms and advances the RE motion along guiding-center (GC) orbits. The model was employed to simulate the formation of RE current plateau during a disruption, in which the plasma current becomes dominated by the RE current. The model agrees well with several RE codes based on fluid models when the RE GC drifts are ignored. For highly relativistic REs, the grad-$B$ and curvature drifts can play a significant role in the RE generation and motion due to the increase in the safety factor profile during the current quench process.

arXiv:2608.15493v1 Announce Type: new Abstract: For an advanced tokamak configuration in the presence of energetic particles (EPs), the dominant instability is found to alternate between infernal modes and Alfv\'en eigenmodes with the variation of the minimum safety factor $q_{\min}$. For relatively high $q_{\min}$, the mode is identified as a reversed-shear Alfv\'en eigenmode (RSAE), characterized by its finite Alfv\'enic frequency and radial localization near the minimum of safety-factor profile. As $q_{\min}$ is further reduced, the dominant branch sequentially transitions through a low-frequency infernal-mode interval, then an energetic-particle-mode (EPM) regime, and finally another low-frequency infernal-mode interval. Increasing the EP beta fraction $\beta_h$ tends to destabilize the RSAE and EPM branches but to stabilize the infernal modes. Phase-space diagnostics further indicate that the destabilizing effects of EPs on the RSAE and EPM branches are mainly associated with

arXiv:2608.16797v1 Announce Type: new Abstract: Industrial recommenders rely on two model families that have evolved largely independently: feature-interaction models over multi-field user/item features, and sequential models over user-behavior histories. Production systems couple them only loosely. To unify the two, we present UniDot, a novel architecture for post-click conversion prediction built from the factorization-machine (FM) point of view: the embedding inner product---which powers collaborative filtering and lets a recommender generalize to unseen user--item pairs---is the same primitive as attention's query dot key scoring, so a single dot-product of tokens can underlie both feature interaction and sequence modeling. UniDot tokenizes non-sequential fields and multi-domain behavioral sequences into one shared token space and stacks a single macro-block in which a token-mixing bus and a sequence-retrieval bus (item tokens cross-attending the histories) run in parallel and

arXiv:2608.16759v1 Announce Type: new Abstract: Design structure matrices (DSMs) are used to comprehensively represent complex systems. They visualize and describe the dependencies between various variables, processes, states, and events. As such they are used in several system engineering approaches, such as requirement and interface management, fault detection, and supervisory control. Currently, a DSM is typically built from knowledge of experts. This may lead to an incomplete or imbalanced DSMs. For instance, elements and links might be missing or superfluous. In this article, we propose a novel method to acquire the DSM using state-of-the-art network identification methods. This demonstrates a proof-of-principle of identifying DSMs from data as an additional tool to the standard heuristic approach. In the future, we plan to embed DSMs in system design and supervisory controllers. We apply this technique to identify the DSM of a fusion reactor modelled by a five-chamber plasma

arXiv:2608.15602v1 Announce Type: new Abstract: While binary quantization theoretically promises extreme compression and acceleration for Large Language Models (LLMs), existing research often overlooks the necessity of specialized hardware kernels, thus failing to unleash the full acceleration potential due to persistent reliance on expensive floating-point arithmetic or runtime dequantization overheads. To bridge this gap, we propose FluxBin (\textbf{F}lexible \textbf{L}UT-based \textbf{U}ltra-low-bit e\textbf{X}ecution with \textbf{Bin}ary bases), an algorithm-kernel co-design that synergizes post-training quantization with a highly optimized CUDA kernel. Algorithmically, we introduce Decoupled Row-Column Binary Decomposition to enhance representational capacity while maintaining hardware efficiency, complemented by a Hessian-guided saliency-aware hybrid bases that preserve critical information. At the kernel level, we implement a Lookup Table Building Approach with Scale Fusion to

arXiv:2608.15447v1 Announce Type: new Abstract: Mobile money has widened financial access across Sub-Saharan Africa and enlarged the surface for money-laundering and terrorism-financing (ML/TF) activity in ecosystems dominated by high-volume, low-value transactions. Rwanda is a case in point: several million active mobile-money users, telecom-led wallets on the MTN and Airtel networks, and a Financial Intelligence Centre (FIC) supervising transaction streams whose scale exceeds static rule-based monitoring. This paper develops and evaluates a transaction-monitoring framework aligned to the Rwandan AML/CFT regime under (i) extreme class imbalance (~0.1% prevalence), (ii) scarce and delayed labels, and (iii) bounded investigator capacity. Using SAML-D, a synthetic dataset of 9,504,852 transactions with 17 laundering typologies, we engineer account-centric behavioural features (rolling velocity, net-flow directionality, counterparty diversity, burstiness) and benchmark supervised

arXiv:2608.14640v1 Announce Type: new Abstract: Stacked bilayer materials exhibit rich stacking-dependent properties driven by the interplay between strong intra-layer bonding and weak inter-layer van der Waals interactions. The computational discovery of such materials is challenging because accurate structure generation typically relies on expensive DFT-based optimization, while existing machine-learning models often fail to explicitly distinguish different interaction types during property prediction. To address these challenges, we propose a machine-learning framework for efficient construction and property prediction of stacked bilayer materials. The framework employs a MatterSim-D3-based structural optimization workflow to generate DFT-quality bilayer structures from monolayer building blocks and stacking configurations at substantially reduced computational cost. For property prediction, we introduce BDIP-Net (Bilayer Dual-Interaction Potential Network), a graph neural network

arXiv:2608.14558v1 Announce Type: new Abstract: Current multimodal models have demonstrated remarkable proficiency in recognizing static visual and auditory content. However, their capacity for abstract perceptual reasoning, inferring unseen information from dynamic, generative processes, remains a critical and underexplored frontier. In this paper, we introduce The Unwritten Benchmark, a new challenge designed to probe this abstract perceptual and cognitive ability. We define the core task as acousto-kinematic word inference: models must decipher words, across 3 different writing styles, being written solely from the audio of pen scratches and the video of hand movements, without any visible ink trace. Our evaluation results reveal a profound gap between human and machine performance: while human participants achieve high ordered letter accuracy (over 80%), leading Multimodal Machine Learning Models, including GPT-4o and Gemini 2.5-Pro, struggle significantly, failing to surpass 10%.

arXiv:2608.15447v1 Announce Type: cross Abstract: Mobile money has widened financial access across Sub-Saharan Africa and enlarged the surface for money-laundering and terrorism-financing (ML/TF) activity in ecosystems dominated by high-volume, low-value transactions. Rwanda is a case in point: several million active mobile-money users, telecom-led wallets on the MTN and Airtel networks, and a Financial Intelligence Centre (FIC) supervising transaction streams whose scale exceeds static rule-based monitoring. This paper develops and evaluates a transaction-monitoring framework aligned to the Rwandan AML/CFT regime under (i) extreme class imbalance (~0.1% prevalence), (ii) scarce and delayed labels, and (iii) bounded investigator capacity. Using SAML-D, a synthetic dataset of 9,504,852 transactions with 17 laundering typologies, we engineer account-centric behavioural features (rolling velocity, net-flow directionality, counterparty diversity, burstiness) and benchmark supervised

arXiv:2608.14186v1 Announce Type: cross Abstract: Tabular anomaly detection is dominated by classical density-proxy methods (Isolation Forest, OCSVM, LOF), reconstruction-based detectors (Autoencoders, VAEs), and modern non-parametric scorers (COPOD, ECOD, Deep SVDD), all of which approximate the inlier distribution only indirectly; explicit energy-based models are largely absent. Motivated by the recent revival of EBMs in deep learning (e.g., Energy-Based Transformers, JEPA), we revisit the classical Deep Boltzmann Machine (DBM) for this task and hypothesize that its mean-field energy combines more effectively with a reconstruction-based score than same-lineage pairs do. We evaluate a two-hidden-layer DBM on two tabular benchmarks spanning distinct domains (UCI Bank Marketing and NSL-KDD) against eight classical and modern baselines across twenty random seeds. The DBM mean-field energy matches the strongest baseline (the Autoencoder) on Bank Marketing and statistically beats it on

arXiv:2608.13829v1 Announce Type: new Abstract: We present an efficient solver framework for the stiff magnetic wave coupling arising in resistive magnetohydrodynamics (MHD) on realistic tokamak geometries. The approach builds on an implicit-implicit (IMIM) time-splitting that separates fast magnetic waves and anisotropic heat transport from slower acoustic dynamics while retaining full coupling (Krzysik et al. 2026). Within this formulation, the magnetic wave subsystem appears as an anisotropic curl-curl operator, enabling the use of scalable auxiliary-space Maxwell (AMS) multigrid solvers. To exploit this structure at the discrete level, we employ curl-conforming finite element spaces for the magnetic field and design the velocity space to preserve the curl-curl structure induced by the Lorentz-force coupling. The resulting compatible discretization preserves the discrete magnetic divergence constraint while producing linear systems directly amenable to efficient AMS-based solvers.

arXiv:2608.14244v1 Announce Type: new Abstract: In large and heavy structures, vibrations arise during motion, posing significant challenges for precise manipulation. To accomplish the desired motion, control algorithms must effectively suppress these structural vibrations. In cutting edge projects, such as remote maintenance of future fusion energy reactors (tokamaks), the manipulation of this type of structure is defined as a crucial task. This paper presents a control strategy to suppress transverse vibrations in flexible payloads during motion using a collaborative payload manipulation approach. Two different industrial robot arms are arranged in a leader follower configuration for the manipulation strategy. The leader robot guides the motion with shaped velocity commands, while the follower robot ensures compliance with the estimated external forces applied by the leader on the payload through an admittance controller. Unlike existing methods, the proposed approach enables

arXiv:2608.14186v1 Announce Type: new Abstract: Tabular anomaly detection is dominated by classical density-proxy methods (Isolation Forest, OCSVM, LOF), reconstruction-based detectors (Autoencoders, VAEs), and modern non-parametric scorers (COPOD, ECOD, Deep SVDD), all of which approximate the inlier distribution only indirectly; explicit energy-based models are largely absent. Motivated by the recent revival of EBMs in deep learning (e.g., Energy-Based Transformers, JEPA), we revisit the classical Deep Boltzmann Machine (DBM) for this task and hypothesize that its mean-field energy combines more effectively with a reconstruction-based score than same-lineage pairs do. We evaluate a two-hidden-layer DBM on two tabular benchmarks spanning distinct domains (UCI Bank Marketing and NSL-KDD) against eight classical and modern baselines across twenty random seeds. The DBM mean-field energy matches the strongest baseline (the Autoencoder) on Bank Marketing and statistically beats it on

arXiv:2608.13829v1 Announce Type: new Abstract: We present an efficient solver framework for the stiff magnetic wave coupling arising in resistive magnetohydrodynamics (MHD) on realistic tokamak geometries. The approach builds on an implicit-implicit (IMIM) time-splitting that separates fast magnetic waves and anisotropic heat transport from slower acoustic dynamics while retaining full coupling (Krzysik et al. 2026). Within this formulation, the magnetic wave subsystem appears as an anisotropic curl-curl operator, enabling the use of scalable auxiliary-space Maxwell (AMS) multigrid solvers. To exploit this structure at the discrete level, we employ curl-conforming finite element spaces for the magnetic field and design the velocity space to preserve the curl-curl structure induced by the Lorentz-force coupling. The resulting compatible discretization preserves the discrete magnetic divergence constraint while producing linear systems directly amenable to efficient AMS-based solvers.

Fusion reactors, devices that generate energy by fusing light atomic nuclei at extremely high temperatures, could contribute to ongoing efforts aimed at producing electricity more sustainably. The extreme environment inside these devices, however, can damage materials that surround the superheated, electrically charged plasma where the nuclear fusion reaction takes place.

arXiv:2608.12963v1 Announce Type: new Abstract: Nuclear batteries powered by alpha decay have been deployed successfully for over 60 years, on a worldwide $^{238}$Pu supply of kilograms per year. We show that the 14 MeV neutrons of a single deuterium-tritium fusion plant can produce alpha emitter battery fuels up to tons per year, in three classes: fuels with completely new production pathways ($^{236}$Pu, $^{227}$Ac, $^{210}$Pb), fuels previously proposed whose scarce feedstock the same pathways now breed at scale ($^{232}$U, $^{228}$Th), and the established $^{238}$Pu. OpenMC simulations of actinide channels in a tokamak blanket give, per GW yr of fusion: 11 to 57 kg of $^{236}$Pu, whose chain releases 18 GJ per gram over a century, ending at stable $^{208}$Pb, plus up to 5.2 t of co-product $^{238}$Pu; up to 1.4 t of $^{231}$Pa from thorium, and, from channel with $^{231}$Pa feedstock, up to $\sim$15 t of $^{232}$U or $\sim$122 kg of $^{210}$Pb, with $^{227}$Ac produced at 21 g/yr

arXiv:2608.12551v1 Announce Type: new Abstract: Compact heavy ion accelerators have numerous applications, ranging from heavy ion fusion to carbon ion radiotherapy, and testing radiation-hardened electronics. The demand could be met by developing high-gradient traveling wave plasma accelerators of high-charge ($\sim\mu\mathrm{C}$) relativistic ion beams. We will discuss a novel ion acceleration regime -- Counter-propagating ionization Front Acceleration (CFA) -- utilizing counter-propagating Ionization Front (IF) and high-current Relativistic Electron Beam (REB). Theoretical modeling and 3D PIC simulations demonstrate the possibility of using typical REBs produced by induction voltage adders propagating through a gas-filled tube undergoing laser ionization to achieve acceleration gradients in excess of $\sim 250 {\rm MeV/m}$ while accelerating micro-Coulombs of ions over meters distance. A unique energy conversion mechanism -- from the REB to electromagnetic fields to the ions is

Future fusion power plants aim to recreate the heart of a star here on Earth to power our future energy needs. While the core fusion plasma will burn at hundreds of millions of degrees, the surrounding structural components must handle sudden, punishing heat loads that rival the extreme temperatures faced by spacecraft upon reentry into Earth's atmosphere. Copper and its alloys are primary candidates for handling these intense heat fluctuations, making it vital to understand exactly how the metal behaves when pushed to its melting point.

arXiv:2608.11335v1 Announce Type: new Abstract: Clinical text can narrow down what to segment, but recent text-guided designs emphasize spatial alignment while overlooking frequency content that governs texture and boundaries. We propose Dual-Domain Cross-Modal Decoding (DD-CMD) for clinical text-guided pulmonary infection segmentation, integrating two complementary forms of language guidance during decoding. In the spatial domain, Text-Guided Spatial Cross-Attention (TGSA) aligns multi-scale visual tokens with text semantics and updates features through gated residual fusion. In the frequency domain, Spectral-Text Adaptive Modulation (STAM) applies a 2D DCT to compute learnable band-energy statistics and predicts text-conditioned FiLM parameters to recalibrate decoder channels for frequency-aware decoding. DD-CMD embeds TGSA and STAM into a coarse-to-fine decoder (7x7 to 56x56) and restores full-resolution masks using a lightweight two-stage refinement module. Experiments on

arXiv:2608.11058v1 Announce Type: new Abstract: Fast and accurate prediction of energetic-particle transport driven by Alfv\'en eigenmode (AE) instabilities is essential for integrated modeling workflows used in the design and optimization of burning plasma fusion reactors. In this work, we develop machine-learning-based surrogate models for rapid prediction of energetic beam and alpha-particle transport fluxes, together with predictive uncertainty estimates, for an ITER steady-state scenario. Two complementary surrogate methodologies, Gaussian process (GP) regression and hierarchical neural networks (NNs), are trained using nonlinear FAR3d gyrofluid simulations of energetic-particle transport. A flux-variability analysis demonstrates that the selected plasma-state representation provides a sufficiently unique parameterization of the nonlinear transport response over most of the sampled feature space, thereby justifying the surrogate formulation. Both surrogate models reproduce the

arXiv:2608.10124v1 Announce Type: new Abstract: Impurity composition, plasma shape, and pedestal density all provide strong levers on fusion power. Here, we explore the ways in which their variation changes fusion power and seek to find the optimum of these parameters. The key impacts of these variables are through changes in the core turbulent transport, the density of the fuel species, the pedestal pressure, and the plasma volume. ITG stabilization due to increased amounts of impurities is observed. The dependence of all of these parameters on the pedestal pressure is especially complicated because of the separate impacts on the peeling and ballooning modes, which can each limit the pedestal. Optimization of this multidimensional operating space is enabled by the use of Bayesian optimization, resulting in an operating point similar to ARC V3A with ~30% more fusion power and a higher fusion power density. Increased shaping parameters, including elongation, triangularity, and

arXiv:2608.10454v1 Announce Type: new Abstract: This perspective examines whether nuclear fusion can provide a scalable, low-carbon power source for rapidly growing AI-driven data center demand. As large language models, cloud computing, and cryptocurrency mining accelerate electricity consumption growth, data centers are projected to account for a substantially larger share of U.S. and global electricity use in the coming decades, creating significant pressure on grid reliability and decarbonization goals. We evaluate the technical and economic alignment between data center load profiles and nuclear power, particularly fusion, through a comparative analysis of capacity factors, levelized cost of electricity, grid interconnection constraints, and deployment pathways. Unlike intermittent renewables, nuclear fission and fusion offer high-capacity-factor, firm baseload generation suited to AI training and inference workloads that require continuous, reliable power. Preliminary

arXiv:2608.08976v1 Announce Type: new Abstract: Parkinson's disease (PD) is the second most common neurodegenerative disorder. Typical machine learning screening methods require PD labels, but the available data is limited by privacy concerns and the need for expert annotation. We propose a label-free face-plus-voice PD screen built entirely on frozen pretrained encoders--a face-expression Vision Transformer and HuBERT--in which no PD label touches any fit; the reference is training controls only. The voice modality uses a synthetic-dysarthria contrastive activation addition (CAA) direction built from time-stretch and breathy degradation of healthy speech; the face modality uses a k-nearest-neighbor anomaly score to the control embedding cluster. We introduce the alignment principle, a post-hoc analysis showing that a synthetic-degradation CAA detector works when the cosine similarity between the synthetic and real disease directions exceeds zero. Measured on the YouTubePD benchmark,

arXiv:2608.07848v1 Announce Type: new Abstract: Robust perception under low-visibility conditions requires fused imagery that jointly preserves infrared thermal saliency and polarization-derived structural details. However, existing infrared-polarization image fusion (IPIF) methods often overemphasize dominant infrared responses, causing weak yet informative polarization textures in dark regions to be suppressed. To address this issue, we propose IRPol-Fuse, an energy-structure coordinated IPIF framework for challenging low-visibility scenarios. The proposed framework contains three key modules: Polarization Attention Fusion for adaptive infrared-polarization allocation, Infrared Highlight Injector for highlight-guided infrared preservation, and Polarization Texture Injector for polarization texture restoration and fine-detail recovery. We further construct LI-PI, a dedicated infrared-polarization evaluation dataset for low-visibility and visually concealed scenes. Experiments on

arXiv:2608.07466v1 Announce Type: new Abstract: The SPARC tokamak will employ $^{238}$U-based fission chambers (FCs) to monitor high-performance deuterium-tritium (DT) plasma operations, spanning neutron yield rates from ${\sim}10^{15}$ to ${>}10^{19}$ n/s. This work validates the $^{238}$U FC design, which utilizes a parallel-plate detector geometry and borated polyethylene collimation to prioritize unscattered DD and DT fusion neutrons. Experimental testing with both DD and DT neutron generators corroborates vendor-specified efficiencies and demonstrates excellent detector linearity, with measured count rates showing good agreement with OpenMC neutronics simulations. Further characterization confirms the $^{238}$U FC's robustness against SPARC-relevant environmental challenges, including stray magnetic fields up to 14 mT and possible signal degradation risks associated with $\sim$30 m long cable runs. These results confirm that the $^{238}$U FC, supported by indirect neutron

arXiv:2608.07165v1 Announce Type: new Abstract: Gyrokinetic simulations of an NSTX spherical-tokamak plasma reveal novel negative-density-gradient (NDG) drift waves at electron-gyroradius-scale that drive turbulent transport comparable to the experimental power flow. Linear simulations indicate that the dominant instability is a trapped-electron electron-scale tearing-parity mode driven mainly by negative electron density gradient, $a/L_{ne} \equiv - a/n_e \, (dn_e/dr)

arXiv:2608.06764v1 Announce Type: new Abstract: A new gyrokinetic code, TEK, was benchmarked in simulating energetic particle (EP) driven toroidal Alfven eigenmodes (TAEs) in the simple tokamak configuration chosen by the ITPA-EP group for code benchmarking purpose. Linear benchmark has been well established by other codes, whereas nonlinear benchmark for this case is lacking. This paper presents, besides the linear benchmark, nonlinear results for both single-n and multiple-n simulations (n is the toroidal mode number). The nonlinear results are in good agreement with an analytical theory on zonal field beat-driven by Alfven eigenmodes, partially verifying correctness of the nonlinear simulations. The saturation level and the resulting EP transport are examined. This provides data for future inter-code nonlinear benchmarking. In TEK, all species (electrons, thermal ions, EPs) are treated on the same footing using the gyrokinetic model (with electrons in the zero Larmor radius limit).

arXiv:2608.06266v1 Announce Type: new Abstract: For more than three centuries, Newton's Second Law (F=ma) has governed mechanics with undisputed success, yet its epistemic authority has rested on empirical adequacy alone. That empirical contingency vanishes under two physical principles - the Principle of Excluded Perpetual Motion (PEPM) and the Weak Equivalence Principle (WEP) - from which the law emerges as a structural necessity of admissible mechanics. Under the PEPM constraint, Suppes' operational measurement protocol grounds gravitational mass and force as independent primitives. The fusion of Stevin's static and Galileo's kinematic inclined-plane analyses establishes F/a as a well-defined, object-intrinsic quantity - the operational definition of inertial mass. The WEP compels the equivalence of inertial and gravitational mass without presupposing Newton's Second Law. Substituting this equivalence into the definition of inertial mass yields the law as a theoretical identity

arXiv:2608.05667v1 Announce Type: new Abstract: Ray tracing codes are useful tools for studying electromagnetic wave propagation and absorption using the geometrical-optics approximation. Existing codes commonly provide either broad radio-frequency coverage in axisymmetric equilibria or three-dimensional capability specialized for electron-cyclotron (EC) applications. BORAY-3D integrates three desirable features in a single version. First, it has a broad frequency range of validity regime from ion-cyclotron, helicon and lower-hybrid waves to EC waves and emission. Second, it provides a unified treatment of arbitrary two- and three-dimensional magnetic-plasma configurations, including both closed and open field-line regions. Third, it incorporates fully relativistic Maxwellian EC absorption. The code extends the axisymmetric BORAY formulation by solving the ray equations in cylindrical coordinates $(r,\phi,z)$ while allowing the toroidal mode number $n_\phi$ to vary. Magnetic-field,

arXiv:2608.05555v1 Announce Type: new Abstract: The Grad-Shafranov (GS) equation governs ideal magnetohydrodynamic equilibrium in tokamak plasmas. Free-boundary GS solvers are central to diverted-equilibrium modeling, but nonlinear Picard iteration introduces computational cost and sample-dependent latency that can become prohibitive in optimization, modeling, and control-oriented loops. Here we train a geometrically conditioned Fourier Neural Operator (FNO) to learn a constrained forward map from spatial coordinates, scalar operating parameters $(P_{\mathrm{axis}}, I_p, f_{\mathrm{vac}})$, and prescribed X-point locations to the poloidal-flux field $\psi(R,Z)$. The model is trained on a controlled family of constrained double-null free-boundary equilibria generated with \textsc{FreeGS} for a single fixed machine geometry and prescribed topology. The best model achieves a mean relative $L^2$ error of $0.05\%$, with test error following an empirical $N^{-0.68}$ power law over

arXiv:2608.05595v1 Announce Type: cross Abstract: Circuit cutting lets a large quantum neural network (QNN) run as independent subcircuits on small devices, but rebuilding its outputs by reconstruction carries a classical sampling overhead exponential in the number of cuts - the dominant runtime cost in prior work. We ask whether, for machine-learning tasks, this step is necessary, and replace it with late fusion: each subcircuit is trained and measured independently, and a small classical head combines their outputs - a linear-cost, decision-level combination borrowed from multimodal learning. To characterize the trade-off we introduce a quantumness dial $Q$, a tunable reconstruction budget interpolating from pure fusion to full reconstruction, and a cut-entanglement diagnostic that indicates how much reconstruction a task needs (Spearman $\rho=0.59$ over $104$ runs). Across synthetic and standard datasets, independently trained late fusion matches full reconstruction accuracy within

arXiv:2608.05843v1 Announce Type: new Abstract: Restoring high-fidelity remote sensing imagery from extreme low-light degradation is indispensable for reliable Earth observation and downstream machine vision. However, under severe noise and illumination corruption, existing methods suffer from attention drift, erroneously aggregating features across distinct physical boundaries and causing severe structural blurring and color distortion. To address this, we propose HALO, a dual-prior-driven enhancement framework that formulates enhancement as a guided feature aggregation problem driven by foundation model priors. Specifically, an illumination-invariant semantic prior provides regional homogeneity as a positive bias for content-consistent aggregation, while a pseudo-3D topological prior provides boundary heterogeneity as a negative penalty to strictly prevent cross-boundary confusion. To cooperatively incorporate these two priors, we propose a Homogeneity-Heterogeneity Cooperative

arXiv:2608.05773v1 Announce Type: new Abstract: A geometry-conditioned, neuro-symbolic closed-loop architecture is proposed for laser powder bed fusion, in which a standards-aligned ontology operates inside the control loop and couples symbolic reasoning with statistical learning to set the targets of a constraint-aware predictive controller. The ontology links the process objectives and constraints to the signals a controller can observe, and a description-logic reasoner converts them into the references and bounds enforced on each scan. The demonstrated case is overhang dross, a quality limit on the melt pool depth, which governs quality yet cannot be measured during the build, is mapped through a geometry- and power-dependent depth-to-width ratio onto a bound on the observable width, with the ratio and its calibrated uncertainty supplied by a Gaussian process. The reasoner classifies each upcoming feature and selects the active constraints-adding a lack-of-fusion floor at

arXiv:2608.03976v1 Announce Type: new Abstract: Protons and neutrons are the smallest forms of measurable matter, and each nucleon has 3 up (+2/3) and down (-1/3) quarks. Quarks treated as point-like particles within polygonal geometries model the structures of stable nuclides through $^{36}$Ar. The model derives from the proton's radius (r=0.8414 fm), the hadron's prolate spheroid shape (from the $\Delta(1232)$ resonance), and the separation distance between bound nucleons ($\approx$0.8 fm, from the Argonne $v_{18}$ $NN$ potential). The prolate nucleon's spatial extent arises from its 3 quarks, which implies a qualitatively linear quark sequence. Spin-spin forces repel the like-flavored quarks to opposite ends of the nucleon, leaving the unlike quark in the middle. Quark-to-quark distance within the nucleon corresponds to the nucleon's radius. Nucleons link by quark-to-quark interactions to form proton-neutron short-range correlated pairs (pn SRC pairs), separated by a distance equal

arXiv:2608.03835v1 Announce Type: new Abstract: Tomographic emissivity reconstruction from plasma diagnostics data relies on a synthetic model mapping the plasma emissivity to the measured signals. The model, referred to as a geometry matrix in the plasma imaging community, is often built using the line-of-sight (LoS) approximation. This approximation neglects the finite width of the detector viewing beams and can therefore introduce systematic errors. Physically correct volume-of-sight (VoS) models remove this inaccuracy by accounting for the full 3D extent of the viewing beams. Their adoption, however, is sometimes hindered by the difficulty of independently validating them. We present an intuitive and easily inspectable voxel-to-detector (V2D) approach for computing physically accurate VoS geometry matrices, based on discretizing the tokamak vessel into voxels and estimating the contribution of each voxel to the measurements of each detector. We apply the V2D approach to the soft

arXiv:2608.03122v1 Announce Type: new Abstract: Advanced stellarator design requires a balance between plasma performance and the manufacturability of three-dimensional modular coils. In conventional two-stage optimization, the coils required to realize an optimized equilibrium can be limited by engineering feasibility. Here, we develop a quasi-single-stage (QSS) framework that incorporates coil feasibility directly into plasma-boundary optimization. QSS uses the maximum normalized normal-field error, evaluated rapidly from surface currents on a uniformly offset winding surface, as a coil-feasibility surrogate. We apply this method to optimize configurations targeting quasi-axisymmetry, quasi-helical symmetry, quasi-isodynamicity, and a combination of omnigenity with piecewise omnigenity. The QSS-optimized configurations exhibit smoother plasma boundaries and winding surfaces, lower normal-field reconstruction errors, and reduced coil complexity, while preserving favourable

arXiv:2608.02763v1 Announce Type: new Abstract: The turbulent transport of toroidal angular momentum helps determine the rotation profiles of tokamak plasmas, and thereby their confinement and stability. The electron contribution therein has an additional consequence: even a modest electron momentum flux can correspond to a substantial turbulent flux of toroidal current, whose divergence could in principle modify the safety-factor profile. Here, using nonlinear gyrokinetic simulations, we show that electromagnetic fluctuations qualitatively alter turbulent momentum transport. In microtearing-mode-driven turbulence, the total momentum transport is inefficient relative to that of heat, yet an electron contribution associated with the turbulent Maxwell stress dominates the momentum flux. We show that this contribution exceeds an estimated scale required for turbulent current redistribution to compete with the collisional processes maintaining the bootstrap current. In the case of