arXiv:2606.21333v2 Announce Type: replace Abstract: Graph Neural Networks (GNNs) have emerged as a powerful paradigm for learning on graph-structured data by iteratively propagating and aggregating information across edges. However, conventional message passing schemes often suffer from over-squashing, whereby exponentially large neighborhoods are compressed into fixed-dimensional embeddings, impeding effective long-range dependency learning. In this work, we introduce Ramanujan Propagation, a graph rewiring strategy that leverages Ramanujan graphs to alleviate topological bottlenecks in GNNs. We first establish that suitably chosen Ramanujan graphs guarantee non-negative resistance curvature, which mitigates over-squashing and facilitates efficient information flow. We then propose an algorithmic framework to construct a Ramanujan rewired graph that preserves the local connectivity of the original graph. Our experiments demonstrate that our method outperforms nine state-of-the-art rewiring techniques. These results establish Ramanujan graphs as a rigorous structural prior for scalable, topology-aware message passing in GNNs.
Science Journals
arXiv:2606.25987v1 Announce Type: new Abstract: Large language models (LLMs) attain remarkable surface fluency on code, yet they neither formally guarantee the syntactic validity of their output nor leverage the hierarchical structure defining the target language. While existing constrained-decoding frameworks address the former, they operate under rigid assumptions that preclude critical lexical mechanisms -- including context-sensitive lexing, maximal-munch tokenization, and keyword extraction -- and only approximate vocabulary masking, sacrificing completeness. For the latter, code LLMs typically inject grammatical structure via predetermined policies rather than learning which structural information to expose. In this work, we introduce Weave of Formal Thought (WoFT), a paradigm uniting rigorous syntactic validation with learned structural representations. First, we present a formal engine and constrained decoder that is sound and complete with respect to the full Tree-sitter specification. By augmenting generalized LR (GLR) parsing with a speculative-lexing construction that maintains concurrent lexer-state hypotheses synchronized with a GLR graph-structured stack, our decoder admits every subword token extending to a valid program prefix and rejects all others. Second, we present a latent-variable fine-tuning method training the language model to interleave non-terminal grammar symbols directly into generation. Utilizing the reweighted wake-sleep (RWS) algorithm to optimize the importance-weighted evidence lower bound (IW-ELBO) of the surface text, the model learns to selectively retain formal derivations as an adaptive structural scratchpad. For Python, fine-tuning StarCoder2-3B with our RWS objective reduces per-token cross-entropy by 14.3% relative to a text-only SFT baseline, demonstrating that discretionary latent syntax recovers critical structural information that flat autoregressive training discards.
arXiv:2606.25711v1 Announce Type: new Abstract: Future multimodal transportation systems require reliable, low-latency communication infrastructures to coordinate autonomous vehicles and moving infrastructure across rail and road networks. Traditional centralized control architectures struggle to meet these requirements in highly dynamic environments due to increased latency, limited scalability, and poor adaptability to changing network conditions. To address this, we propose a distributed communication architecture integrating Software-Defined Networking (SDN) and Multi-Access Edge Computing (MEC), that can create a flexible, programmable and lowlatency network. Results show controller communication and flow setup latency between edge SDN controllers and Pods are lower than reported in literature. The framework uses hierarchical control with regional and edge controllers to support low-latency interface management and edge autonomy. Operational workflows and control logic are defined as representative scenarios. The architecture combines regional policy coordination with edge-level autonomy, enabling local failover and adaptive interface management without central dependency.
arXiv:2606.25732v1 Announce Type: new Abstract: Real-world single image dehazing is highly ill-posed due to spatially and spectrally varying scattering, while practical deployment demands lightweight and low-latency models. Existing approaches either rely on fragile physical inversion under simplified assumptions or adopt heavy blind architectures unsuitable for edge deployment. To overcome these limitations, we propose PGL-Net (Physics-Inspired Global-Local Decoupling Network), a lightweight framework that incorporates physical inductive biases via operator-level emulation, avoiding explicit parameter estimation. It decouples dehazing into global distribution rectification and local structural refinement. A Physics-Inspired Affine Fusion (PAF) module performs globally conditioned alignment across hierarchical skip connections to compensate for haze-induced bias, while a compact Degradation-Aware Modulation (DAM) block adaptively restores spatially and spectrally variant details through dynamic feature modulation. Extensive experiments on multiple real-world benchmarks demonstrate that PGL-Net achieves state-of-the-art restoration quality with significantly reduced complexity. Compared with the recent SOTA SGDN, the Tiny variant (PGL-Net-T) improves PSNR by up to 2.6dB and consistently enhances downstream object detection accuracy, while achieving over a 10x reduction in inference latency. Code is publicly available at: https://github.com/sc-30-bit/PGL-Net.
arXiv:2606.23461v3 Announce Type: replace Abstract: This study investigates heart rate (HR) and heart rate variability (HRV) responses to physical activity in breast cancer survivors using wearable electrocardiogram (ECG) data collected in real-world settings. Reliable HRV analysis in such environments is challenging due to motion artifacts and activity-related signal degradation. To address this, we use an approach that combines accelerometer and gyroscope data for activity intensity segmentation (light, moderate, vigorous) with a robust ECG processing pipeline incorporating R-peak detection and annotation-free signal quality assessment. Because vigorous activity produced unreliable HRV estimates, analyses focused on light and moderate activity levels. Using 30 s, 1 min, and 2 min windows, HR and HRV metrics were computed and compared between breast cancer survivors and healthy controls. Cancer survivors consistently exhibited elevated HR and reduced HRV across activity levels. During light activity, HR increased from 95.7 bpm in controls to 103.4 bpm in cancer survivors. Differences became more pronounced during moderate activity, where RMSSD decreased from 39.7 ms to 22.1 ms and SDNN from 42.6 ms to 25.1 ms. Statistical analyses showed significant group differences with strong and consistent effects across observations. In addition, the proposed ECG quality assessment framework reliably identified high-quality signal segments, achieving near-perfect valid RR ratios (0.99) without manual annotations. Overall, these findings demonstrate impaired and activity-dependent autonomic regulation in cancer survivors and highlight the importance of motion-aware activity segmentation and robust ECG quality control for accurate physiological monitoring in real-world wearable settings.
arXiv:2606.23760v2 Announce Type: replace Abstract: Engineering reliable autonomous systems is an important and growing topic in computer science. As autonomous systems become more prevalent, easy-to-use techniques for building them reliably are increasingly important. This workshop report captures and expands on the discussions at the Lorentz Center Workshop "Engineering Reliable Autonomous Systems" (ERAS), held from 10 to 14 June 2024. The workshop was co-organised by the organisers of the Workshop on Formal Methods for Autonomous Systems (FMAS) and the Workshop on Agents and Robots for reliable Engineered Autonomy (AREA). It brought together members of the FMAS and AREA communities, industry practitioners, and representatives from sectors where autonomous systems pose distinctive engineering challenges. The workshop focused on three main research topics: techniques for verification and validation of autonomous systems; engineering real-world autonomous systems; and software architectures for safe autonomous systems. Its main outcome is a catalogue of challenges in these areas and, most importantly, a pathway to solutions. Some challenges can already be tackled by techniques that are well known in academia but have not yet become regularly used in practice. Other challenges remain unresolved and require further research. This roadmap is intended to support future research and industrial collaboration.
arXiv:2606.23856v2 Announce Type: replace Abstract: Generative molecular models for drug design are a promising direction with much active research. In the next phase of computational drug design, such models will need to understand small molecule structure and protein-ligand interactions, and they will need to possess the machinery to generate molecules de novo. Incorporating each feature poses a critical challenge. Equally important, yet often treated as secondary, is the ability to grow a molecule from a partial starting point -- a scaffold or fragment supplied by a chemist -- which is the central operation of lead optimization. We present Sesame (Spatial Evoformer for a Structure-Aware Molecular Engine), a diffusion-based molecular generation model that leverages a novel spatial pairformer module to condition on partial molecular structure and the surrounding protein pocket, both expressed as continuous spatial density maps. This single conditioning mechanism supports both de novo generation and fragment-conditioned lead optimization, letting a medicinal chemist prune a hit to a scaffold and have Sesame grow it in productive ways. In addition to this module, we also introduce a diffusion framework for joint denoising of atom types, bond types, and positions, along with a trajectory finetuning scheme that trains on the model's own sampling rollouts to improve generation quality. Sesame is trained on a large corpus of ligand-only and protein-ligand datasets.
arXiv:2606.24027v2 Announce Type: replace Abstract: Studies of Internet paths often attach router locations to traceroute hops using commercial geolocation databases, rDNS labels, Geofeeds, and IXP metadata. These sources provide useful hints, but they report point locations without calibrated confidence, leaving researchers unable to tell whether a geographic path is trustworthy. We introduce Path Consistency Scoring (PCS), a passive framework that evaluates router geolocation as a path-level consistency problem. PCS models each traceroute as a sequence of candidate city-level locations and uses a Hidden Markov Model to fuse local evidence with speed-of-light constraints and empirical latency priors. PCS produces a path consistency score summarizing how well metadata and observed RTT increments support a coherent geographic interpretation. Because this score is only meaningful when latency proxies for geography, we also define a Path-Model Alignment metric that compares speed-of-light residual increments of the decoded path against a reference path. We evaluate on 413,354 RIPE Atlas traceroutes and a 6,555-path subset verified by active probing. On validated paths, 94.2% of decoded sequences achieve mean error below 200 km. PCS is largely GeoDB-agnostic; median scores vary by less than 5% across four commercial databases, while the alignment metric reveals that over half of DB-IP and IP2Location paths require substantial correction, compared with 15% for IPinfo. This lets downstream analyses quantify confidence in their geographic conclusions rather than inheriting database accuracy without qualification.
arXiv:2606.25758v1 Announce Type: new Abstract: While test-time adaptation (TTA) empowers vision-language models to adapt without costly retraining, it remains highly vulnerable to out-of-distribution (OOD) outliers prevalent in real-world applications. This discrepancy motivates Noisy TTA (NTTA), an online task to filter noisy OOD samples on the fly while maximizing in-distribution (ID) classification accuracy. Existing zero-shot NTTA approaches typically rely on test-time discriminative training, leading to overconfident misclassifications and significantly degraded inference efficiency. To address these limitations, we propose a novel framework named Dual Distribution Estimation (DDE), shifting the zero-shot NTTA paradigm from instance-level learning to training-free Gaussian distribution modeling. DDE incorporates two novel modules: Positive Feature Distribution Estimation (PFDE) and Negative Label Distribution Estimation (NLDE). PFDE explicitly models class-wise inclusion and exclusion Gaussian distributions to formulate a calibrated contrastive score, robustly enhancing ID accuracy. In parallel, NLDE improves OOD identification by explicitly modeling the negative label distribution to mine highly discriminative labels, effectively mitigating spurious correlations. Extensive experiments show that on the large-scale ImageNet benchmark, DDE achieves an improvement of 3.70\% in harmonic mean accuracy and reduces the FPR95 for OOD detection by 6.20\%, while ensuring highly scalable and efficient online inference. Furthermore, DDE is zero-shot and training-free, demonstrating remarkable robustness in data-scarce scenarios. Codes are available at https://github.com/ZhuWenjie98/DDE.
arXiv:2606.25759v1 Announce Type: new Abstract: Large-scale neural-network training repeatedly aggregates gradients across devices, making communication a central cost in distributed learning. Low-bit gradient aggregation can reduce this cost, but applying it as a static replacement for full-precision communication can destabilize training because safe precision depends on training phase, model structure, runtime bucketization, and the communication substrate. This paper presents NEURON-Fabric, a profile-guided runtime system for controlled low-bit gradient communication. NEURON-Fabric uses calibrated operating profiles, model-aware runtime bindings, online training-health monitoring, and reducer-capacity checks to decide when low-bit aggregation should be admitted, when execution should fall back to FP32, and which model regions are eligible for each route. The runtime preserves model semantics inside mixed DDP buckets and treats reducer admission as an architecture-runtime co-design problem rather than as a standalone compression operator. Across vision, Transformer, and autoregressive language-model workloads, NEURON-Fabric validates the path from calibration to distributed communication-hook execution. Static low-bit communication can collapse training accuracy, while profile-guided control preserves accuracy near full-precision references or calibrated targets and reduces modeled gradient-communication traffic in the evaluated settings. Transformer and billion-parameter language-model checks show that the same routing and fallback mechanisms execute across model families and multi-node deployments. Reducer-side replay and reducer-path measurements identify when compact sign-count aggregation is expected to reduce communication cost and when endpoint capacity should trigger fallback.
arXiv:2606.25760v1 Announce Type: new Abstract: Computer-use agents turn vision-language model (VLM) predictions into executable GUI clicks, so reliable uncertainty estimates are essential for rejection, calibration, miss-severity ranking, and spatial safety regions. Yet evidence on post-hoc uncertainty quantification (UQ) for these agents is fragmented across isolated model and dataset pairs, leaving it unclear whether UQ rankings stay stable when the agent, benchmark, or observable interface changes. We present Argus, a cross-regime benchmark for post-hoc UQ in single-step executable GUI grounding: a 27-method open-weight matrix over 4 VLM agents and 4 datasets, plus an 8-method closed-source matrix across 3 frontier vendors where logits, hidden states, and attention maps are unavailable. Evaluated methods span logit-based scores, sampling and consistency measures, hidden-state and density estimators (Mahalanobis, SAPLMA), attention-based scores, P(True) and verbalised-confidence prompting, and split-conformal prediction. The main finding is selective transfer: UQ rankings are stable across datasets for a fixed model, but degrade across model classes and observable interfaces. Hidden-state and density methods are the most stable open-weight family, while CoCoA-1MCA, Focus, sampling-based scores, and verbalised self-assessment win in specific regimes. Within-model ranking transfer is strong (Spearman rho up to 0.969), but cross-tier transfer to closed-source vendors averages only +0.08, so closed-source UQ should be reranked on the target rather than extrapolated. Conformal click regions show score-level discrimination is not enough for deployment: locally weighted disks shrink radii by 40-60% when the plug-in UQ is calibrated, but coverage degrades under calibration-test or interface mismatch. We release per-item records, calibration/test splits, UQ scores, and analysis scripts for regime-aware UQ selection in GUI agents.
arXiv:2606.25501v1 Announce Type: cross Abstract: We present a novel methodology for studying \emph{large deviations principles} (LDPs) of random matrices. By utilizing a partially lifted variant of \emph{random duality theory} (RDT), we develop a generic LDP framework that completely circumvents traditional random matrix theory (RMT) methods. To demonstrate the framework's simplicity and accuracy, we apply it to the Wishart and Wigner GOE classical statistical ensembles. In both cases, we obtain elegant LDP characterizations of the upper and lower spectral edges that fully match the results achieved through traditional \emph{Coulomb gas} methodologies in [85,95].
arXiv:2606.25790v1 Announce Type: cross Abstract: Many statistical models have natural symmetries described by a group action. We study how such symmetries affect the comparison of two distributions. We work with a transformation model in which a group acts on both the sample space and the parameter space, and the densities transform with a multiplier. Under this assumption, we show that every $f$-divergence is invariant under the group action. As a consequence, an invariant divergence depends only on a maximal invariant of the pair of parameters. When the action on the parameter space is transitive, this maximal invariant is given by a double coset. We apply this result to multidimensional location-scale families, and we show that the same reduction applies to the Fisher--Rao distance.
arXiv:2606.24570v2 Announce Type: replace Abstract: Vision-language contrastive pretraining has become the dominant recipe for 3D medical foundation models, leveraging the large volumes of paired scans and reports produced in clinical practice. However, medical images usually span dozens of organs, and radiological reports are much longer than typical natural image captions and are composed of multiple structured sections. CLIP-style pretraining compresses this structure by encoding each modality into a single global token, at the risk of losing important details. We introduce ConQuer (Concept Queries), an image-text pretraining method that augments CLIP's global alignment with a set of localized alignments, one per concept. ConQuer splits the report into concept-specific sections and learns cross-attention queries that pool the matching image features without using any segmentation mask or spatial supervision. Contrastive learning is then applied independently for each concept. Concepts can be any unit of semantic localization; here, they are anatomical regions, one query per organ or gross body region. As a byproduct, each query learns attention maps focused on its concept, providing built-in spatial interpretability. We use ConQuer to train Jolia, a 3D CT foundation model on chest and abdominal CT. Jolia consistently outperforms a CLIP baseline on findings classification, report generation, and cross-center transfer, and sets a new state of the art across multiple public benchmarks. Jolia's weights are available at https://huggingface.co/raidium/Jolia
arXiv:2606.25776v1 Announce Type: new Abstract: Optical rotatory dispersion (ORD) in chiral media, classically demonstrated as the "sweet monochromator," provides a robust mechanism for liquid-tunable spectral filtering. However, the ultimate physical boundaries governing its spectral bandwidth remain fundamentally unexplored beyond classical electromagnetic theory. Here, we present a comprehensive framework bridging macroscopic optical filtering with the quantum electrodynamic (QED) limits of chiral light-matter interaction. Using time-dependent density functional theory (TD-DFT) combined with Boltzmann conformational averaging, we accurately compute the ORD curves of representative saccharide systems, revealing that the theoretical minimum bandwidth is intrinsically tied to the anomalous dispersion (Cotton effect) near the molecular absorption band. While our macroscopic experiments demonstrate a classical bandwidth limit of ~ 20 nm due to heterogeneous broadening and instrumental constraints, our first-principles calculations predict achievable sub-nanometer bandwidths utilizing visible-absorbing chiral molecules. Furthermore, we extrapolate this framework to the strict quantum limit, proposing a single-photon QED architecture operating at ultra-low temperatures (~ 10 mK). In this regime, the spectral purity is constrained solely by the natural lifetime of the molecular excited state, governed by the Heisenberg uncertainty principle. This work establishes the ultimate theoretical criteria for chiral optical filters and pioneers the concept of the "quantum monochromator" for ultrasensitive chiral spectroscopy and quantum information processing.
arXiv:2606.25778v1 Announce Type: new Abstract: This paper presents a versatile framework for evaluating fuzzy quantification queries over both standard and fuzzy ontologies as well as knowledge graphs. The primary objective is the retrieval of individuals that satisfy queries articulated via Type I or Type II fuzzy quantified expressions. A key advantage of the proposed approach is its inherent adaptability: it remains entirely agnostic to the quantifier type, the underlying evaluation method, and the specific data source of the ontology (i.e., OWL ontologies or RDFS knowledge graphs). Furthermore, we present Q2S2, a publicly accessible implementation of this system developed to support future research.
arXiv:2606.25780v1 Announce Type: new Abstract: The T-shaped mixer is widely used in fundamental studies of chemical engineering. Its transitional regime is well understood, whereas the turbulent dynamics has received scarce attention so far. Here we perform direct numerical simulations of the turbulent regime for Reynolds numbers up to $Re=2000$ at Schmidt number $Sc=1$. Our analysis reveals two distinct stages along the mixing channel prior to relaxation toward duct flow. Near the junction, a jet-like flow forms and exhibits the approximately self-similar behaviour of transitional planar jets. Subsequently, a decay region characterised by power-law decay of turbulent kinetic energy, dissipation and scalar variance emerges. For the velocity field, the observed exponents are consistent with those of decaying turbulence in bounded domains, whereas the scalar-variance exponent is consistent with that of unbounded turbulence. We argue that this apparent discrepancy is a consequence of the mixing process progressing from the center of the channel toward the side walls in the decay region, while turbulence already fills the channel cross-section entirely at the end of the jet region.The time-averaged mixing state presents error-function profiles of the scalar in the transverse direction, similar to the laminar cases, and is quantified here through a stream-wise evolving effective diffusion coefficient.
arXiv:2606.25782v1 Announce Type: new Abstract: With the widespread adoption of large language models (LLMs) in chatbots and everyday applications, companies increasingly need guardrails that are effective while remaining low-cost and low-latency. Safety evaluation of LLM outputs has generally relied on LLM-based judges, which can be effective but are often slow and expensive to deploy at scale. In this paper, we evaluate whether fine-tuned modern encoder classifiers from the ModernBERT family, including ModernBERT and Ettin, can reliably identify harmful LLM outputs in user-model conversations without substantial performance loss relative to LLM-based judges. We benchmark these encoder classifiers against rule-based prefix matching, fine-tuned LLM classifiers, and LLM judges using a range of judge-prompting strategies across open-source adversarial datasets. The LLM judges include evaluation methodologies from StrongReject, ShieldGemma, JailbreakBench, AILuminate, SorryBench, and a Claude-as-a-judge setup, as well as fine-tuned safety classifiers such as LlamaGuard 3 and LlamaGuard 4. The encoder classifiers are fine-tuned on judge-labeled data using a majority-voting label strategy and are then evaluated on a gold-standard holdout dataset to assess their performance relative to LLM judges. We report absolute performance using F1 score, false negative rate, and precision-recall metrics. We also break down results by attack technique, including single-turn prompting, decomposition, escalation, and context manipulation, to identify where encoder classifiers align with or diverge from LLM-based judges. Our findings provide guidance on when encoder classifiers can serve as cost- and latency-efficient alternatives to LLM-based safety evaluation.
arXiv:2605.10913v3 Announce Type: replace Abstract: As LLM agent systems take on more complex tasks, they increasingly rely on meta-agents: higher-order agents that create, operate on and manage other agents. Meta-agent operations such as coordinating agents, halting risky actions before execution, or repairing failed runs, require runtime manipulation of agentic execution. Yet existing agentic substrates make this difficult: they expose only transcripts and environment snapshots, forcing meta-agents to build ad hoc tooling to reconstruct and operate over full execution state. Therefore, we introduce Shepherd, a Python substrate grounded in functional programming principles, where an agent's execution is itself a first-class object that a meta-agent can easily inspect and transform. Every model action, tool call, and environment change becomes a structured event in a reversible, Git-like execution trace, where any past state can be reverted 5x faster than docker commit and fork. Three example use cases show Shepherd's versatility: (1) a supervisor meta-agent prevents conflicts among parallel coding agents, lifting pair-coding pass rate from 28.8% to 54.7% on CooperBench; (2) a counterfactual optimization meta-agent repairs agent workflows by proposing edits and replaying runs from the point of changed behavior, outperforming MetaHarness on Terminal-Bench 2.0 by 12.8% with 58% lower wall-clock; (3) a training meta-agent picks fork points during rollouts to improve credit assignment in long-horizon agentic RL, doubling GRPO's uplift on Terminal-Bench 2.0. We open-source Shepherd to enable principled and efficient operations over agentic execution for both users and meta-agents.
arXiv:2606.25786v1 Announce Type: new Abstract: This document outlines the changes adopted for ACM HotNets 2026, spanning its scope, review process, and program structure. Rather than isolated adjustments, these changes form a coherent effort to clarify and extend HotNets' role as a venue for agenda-setting research, community discussion, and experimentation with how the networking community evaluates, disseminates, and discusses research. In particular, HotNets 2026 broadens its scope to include perspective and community-facing contributions, introduces distinct evaluation criteria for technical and perspective papers, adopts a more collaborative and discussion-oriented review process, rethinks how accepted work is presented and discussed at the workshop, and explores responsible uses of generative AI (GenAI) in reviewing and research dissemination. We believe these changes will help HotNets continue to serve as a home for ambitious, unconventional, and thought-provoking ideas, while also positioning it as a venue for experimenting with new approaches and formats that larger conferences, e.g., SIGCOMM or NSDI, might later adopt. We use this document to solicit feedback from the community, both on these changes and on how HotNets can best serve the networking community in the future. We plan to collect feedback during and after the event and to prepare a follow-up report summarizing the community's reactions and lessons learned.
arXiv:2606.24748v2 Announce Type: replace Abstract: Smoothed-aggregation algebraic multigrid (AMG) is widely used for the linear systems arising from finite-element discretizations of vector PDEs such as elasticity, but its GPU implementations have used scalar sparse matrix formats. These problems carry a natural block structure: matrix nonzeros occur in dense bs x bs blocks sharing one column index, so storing the blocks directly removes most of the index data and raises the arithmetic intensity of the bandwidth-bound kernels that dominate AMG on the GPU. Existing blocked GPU kernels (cuSPARSE, Kokkos Kernels) require equal row and column block sizes, but AMG for elasticity is rectangular-blocked: the near-null space of rigid-body modes makes the coarse block size (6 in 3D) differ from the fine (3), so the prolongator and the Galerkin triple product mix block sizes. We add a portable, Kokkos-backed blocked matrix type to PETSc with rectangular-block kernels, and make every step of the smoothed-aggregation setup operate on the block format directly, with no expansion to scalar form on the coarsening path. The two phases that recur when the hierarchy is reused across solves -- the Galerkin coarse-operator recompute (A_c = P^T A P) and the V-cycle -- are kept resident on the device in blocks, via a native blocked off-process prolongator gather over a PetscSF and a new blocked COO assembly path for dense rectangular blocks. On A100 GPUs for 3D elasticity, the cuSPARSE Galerkin product runs out of GPU memory on a 128^3 grid (6.3M unknowns) packed onto 8 GPUs, where the blocked format fits; the native Kokkos Kernels scalar path also fits, but with a much heavier Galerkin product. Where the formats run, the blocked format is at parity on one GPU and faster at scale: at 27 GPUs it is 1.24x faster on the V-cycle, 1.42x on SpMV, and 1.80x on the coarse-operator recompute, reaching 2.27x on the latter at 64 GPUs.
arXiv:2605.01895v2 Announce Type: replace-cross Abstract: Nonthermal features are ubiquitously observed in electron velocity distribution functions in the solar wind, yet their origin in the collisionless, turbulent, expanding solar-wind plasma remains unclear. We investigate how solar-wind expansion and Alfv\'enic turbulence jointly generate and regulate these features using the first fully kinetic particle-in-cell simulation of an expanding turbulent plasma under heliospheric conditions. In our setup, expansion-driven weakening of the magnetic field adiabatically cools the plasma perpendicularly to the mean field while leaving the parallel temperature largely unchanged, driving the system toward the firehose-instability threshold. Concurrently, strongly anisotropic turbulence leads to perpendicular heating and the development of nonthermal features. Subsequently, we find that suprathermal electron populations preferentially develop in the parallel direction, forming pronounced power-law tails even under weakly compressive, highly Alfv\'enic conditions, and persist despite anisotropy regulation by the firehose instability. The preferentially parallel energization suggests the involvement of parallel electric fields or resonant wave--particle interactions, rather than simple velocity-space redistribution. These results provide the first direct evidence of the emergence of nonthermal-electron features in a unified kinetic framework linking expansion, turbulence, and instabilities in the solar wind.
arXiv:2606.23201v2 Announce Type: replace-cross Abstract: If disk detection is applied to nanobeam electron diffraction datasets, then the results are effectively a list of vectors describing the position of every diffraction peak in real and reciprocal space. This is the natural territory for the application of clustering algorithms, and they are shown to be highly effective at decomposing such datasets and automating imaging and analysis. Examples are shown in both polycrystalline and single crystal (with precipitates) systems. Additionally, automated separation of amorphous or deeply nanocrystalline components is also found to be possible allowing composite images of both amorphous and crystalline components in partially crystallised samples to be easily and automatically generated. These advances promise to increase throughput in atomic structure analysis with nanobeam diffraction, and also make finding minor components much easier. They can also serve as a preliminary step towards more detailed crystallographic or crystal size/shape distribution analysis.
arXiv:2606.25835v1 Announce Type: new Abstract: In this work, we develop a multi-fidelity strategy for kinetic models in epidemiology with uncertain contact dynamics. Assessing and controlling the population-level effects of contact dynamics requires the development of models for understanding observable effects of heterogeneous contact structures, whose formation depends on complex social phenomena. These can be captured taking into account high-dimensional uncertain quantities. The proposed approach combines high-fidelity kinetic solvers with a hierarchy of low-fidelity surrogates, including reduced macroscopic models and coarse kinetic descriptions, remaining applicable even in regimes where a macroscopic closure is unavailable. This hierarchical framework identifies representative parameter samples and reconstructs full solutions via projection-based techniques, enabling efficient uncertainty propagation while drastically reducing computational cost. Numerical experiments in high-dimensional stochastic settings demonstrate that accurate statistical estimates of epidemic observables can be obtained with significantly reduced computational costs compared to standard approaches.
arXiv:2606.25882v1 Announce Type: new Abstract: DGPs are probabilistic models with remarkable prediction performance that concatenate GPs across several layers. Exact inference in DGPs is intractable, and variational inference is often used to approximate the posterior with a parametric distribution tuned by minimizing the Kullback-Leibler divergence. Moreover, finding a good VI approximation is challenging. In particular, a problem of VI is posterior collapse, where VI converges to a variational posterior that matches the prior. In variational DGPs, this implies explaining the data as noise. This work studies posterior collapse in DGPs and identifies its connection to the DSVI algorithm and the widely used linear prior mean function employed in all but the last layer. We show that the benefit of the linear prior mean does not arise from avoiding the non-injective pathology in very deep DGPs, as previously believed, but from improving the conditioning of the optimization problem at initialization. Thus, we propose an alternative initialization of a zero prior mean DGP that mimics a DGP with a linear prior mean at initialization. This enables successful training of DGPs without imposing optimization-driven constraints on the prior, allowing to choose the prior based on modeling assumptions rather than optimization convenience. Our analysis considers three common parameterizations of DGPs and shows that not all of them benefit from a linear prior mean. We also explain why a whitened parameterization of the \DGP provides more stable convergence, something often assumed from experience, but lacking a rigorous analysis. Furthermore, we show that this stability is also beneficial to avoid the posterior collapse problem. Extensive experiments validate our findings: the proposed initialization prevents posterior collapse, improves stability, and achieves performance comparable to (and sometimes better than) DGPs with a linear prior mean.