arXiv:2511.12038v2 Announce Type: replace Abstract: We show that strong thermo-optic feedback in nanophotonic resonators creates forbidden steady state temperatures that are unreachable under any static excitation. This temperature bandgap opens when thermo-optic feedback gain overcomes optical and thermal dissipation, splitting an otherwise continuous thermal landscape into disconnected accessible bands. We experimentally map this temperature band structure using quasi-bound states in the continuum resonances in silicon metasurfaces. Continuous wavelength scans exploit spectrally accumulated thermal energy to access the forbidden interval, reaching up to ~88{\deg}C higher temperature than static excitation, both at the same wavelength and maximum power. A characteristic three-stage temperature rise near the band edge confirms the -1/2 critical exponent of saddle-node bifurcations and enables quantitative extraction of band-edge wavelengths. Combining external thermal bias with optical excitation drives programmable interband transitions, enabling nearly 8.5-fold amplification of temperature rise and wavelength-selective switching between thermal states separated by just 1 nm in excitation wavelength. The bandgap is fully designable through metasurface geometry and hybrid material integration, which tune optical absorption, confinement and resonance linewidth. These results establish the temperature bandgap as a designable degree of freedom in driven nanophotonic systems, opening routes toward all-optical thermal logic and programmable photothermal control of chemical and biological processes.
Science Journals
arXiv:2512.00727v2 Announce Type: replace Abstract: Reinforcement learning has enabled impressive locomotion skills on articulated robots, but common policy representations remain only weakly aligned with robot physics. Generic networks ignore kinematic structure, while graph-based policies encode connectivity without specifying how physical quantities transform across symmetric body parts. We introduce a morphological symmetry graph representation for locomotion policy learning and instantiate it in MS-PPO. Starting from the robot's topological graph, our representation augments each observation and action space with the permutation and sign transformations induced by morphological symmetry. This yields a symmetry-equivariant graph actor and a symmetry-invariant graph critic, enforcing the desired policy and value constraints by construction rather than through reward shaping or data augmentation. We evaluate MS-PPO on a variety of locomotion tasks using both Unitree Go2 quadruped and Unitree G1 humanoid, including command tracking, asymmetric joint failures, out-of-distribution command generalization, and zero-shot sim-to-real deployment. Experiments show improved symmetry generalization, robustness, sample efficiency, and model efficiency over topology- and symmetry-aware baselines.
arXiv:2512.06244v2 Announce Type: replace Abstract: The exploration-exploitation dilemma in reinforcement learning (RL) is a fundamental challenge to efficient RL algorithms. Existing algorithms for finite state and action discounted RL problems address this by assuming sufficient exploration over both state and action spaces. However, this yields non-implementable algorithms and sub-optimal performance. To resolve these limitations, we introduce a new class of methods with auto-exploration, or methods that automatically explore both state and action spaces. Auto-exploration can be applied in both the tabular and linear function approximation setting. Under algorithm-independent assumptions on the existence of an exploring optimal policy, both settings attain $O(\epsilon^{-2})$ sample complexity to solve to $\epsilon$ error. These complexities are novel since they avoid algorithm-dependent parameters seen in prior works, which may be arbitrarily large. The methods are also simple to implement because they are parameter-free. We achieve these results by integrating auto-exploration into policy mirror descent to avoid the (unknown) stationary distribution seen in prior art. In the tabular setting, we introduce a dynamic exploration time with a data-driven stopping time, while for linear function approximation we propose a new sampling distribution based on the discounted visitation distribution that covers a more general class of Markov chains.
arXiv:2512.17796v2 Announce Type: replace Abstract: Recent advances in world models have greatly enhanced interactive environment simulation. Existing methods mainly fall into two categories: (1) static world generation models, which construct 3D environments without active agents, and (2) controllable-entity models, which allow a single entity to perform limited actions in an otherwise uncontrollable environment. In this work, we introduce CustomX, leveraging the realism and structural grounding of static world generation while extending controllable-entity models to support user-specified characters capable of performing open-ended actions. Users can provide a 3DGS scene and a character, then use natural language to direct the character to perform diverse behaviors, ranging from basic locomotion to object-centric interactions, while freely exploring the environment. CustomX synthesizes temporally coherent video clips that preserve visual fidelity with the provided scene and character, formulated as a conditional autoregressive video generation problem. Built upon a pre-trained video generator, our training strategy significantly enhances motion dynamics while maintaining generalization across actions and characters. Our evaluation covers a broad range of aspects, including visual quality, character consistency, action controllability, and long-horizon coherence.
arXiv:2601.17917v3 Announce Type: replace Abstract: Diffusion Large Language Models (dLLMs) offer a compelling paradigm for natural language generation, leveraging parallel decoding and bidirectional attention to achieve superior global coherence compared to autoregressive models. While recent works have accelerated inference via KV cache reuse or heuristic decoding, they overlook the intrinsic inefficiencies within the block-wise diffusion process. Specifically, they suffer from spatial redundancy by modeling informative-sparse suffix regions uniformly and temporal inefficiency by applying fixed denoising schedules across all the decoding process. To address this, we propose Streaming-dLLM, a training-free framework that streamlines inference across both spatial and temporal dimensions. Spatially, we introduce attenuation guided suffix modeling to approximate the full context by pruning redundant mask tokens. Temporally, we employ a dynamic confidence aware strategy with an early exit mechanism, allowing the model to skip unnecessary iterations for converged tokens. Extensive experiments show that Streaming-dLLM achieves up to 68.2X speedup while maintaining generation quality, highlighting its effectiveness in diffusion decoding. The code is available at https://github.com/xiaoshideta/Streaming-dLLM.
arXiv:2603.28011v2 Announce Type: replace Abstract: We present a novel framework that jointly trains a neural network controller and a neural Riemannian metric with rigorous closed-loop contraction guarantees using formal bound propagation. Directly bounding the symmetric Riemannian contraction linear matrix inequality causes unnecessary overconservativeness due to poor dependency management. Instead, we analyze an asymmetric matrix function $G$, where $2^n$ GPU-parallelized corner checks of its interval hull verify that an entire interval subset $X$ is a contraction region in a single shot. This eliminates the sample complexity problems encountered with previous Lipschitz-based guarantees. Additionally, for control-affine systems under a Killing field assumption, our method produces an explicit tracking controller capable of exponentially stabilizing any dynamically feasible trajectory using just two forward inferences of the learned policy. Using JAX and \immrax{} for linear bound propagation, we apply this approach to a full 10-state quadrotor model. In $<$10 minutes of post-JIT training, we simultaneously learn a control policy $\pi$, a neural contraction metric $\Theta$, and a verified 10-dimensional contraction region $X$.
arXiv:2601.23147v2 Announce Type: replace Abstract: The integrity of time in distributed Internet of Things (IoT) devices is crucial for reliable operation in energy cyber-physical systems, such as smart grids and microgrids. However, IoT systems are vulnerable to clock drift, time-synchronization manipulation, and timestamp discontinuities, such as the Year 2038 (Y2K38) Unix overflow, all of which disrupt temporal ordering. Conventional anomaly-detection models, which assume reliable timestamps, fail to capture temporal inconsistencies. This paper introduces STGAT (Spatio-Temporal Graph Attention Network), a framework that models both temporal distortion and inter-device consistency in energy IoT systems. STGAT combines drift-aware temporal embeddings and temporal self-attention to capture corrupted time evolution at individual devices, and uses graph attention to model spatial propagation of timing errors. A curvature-regularized latent representation geometrically separates normal clock evolution from anomalies caused by drift, synchronization offsets, and overflow events. Experimental results on energy IoT telemetry with controlled timing perturbations show that STGAT achieves 95.7% accuracy, outperforming recurrent, transformer, and graph-based baselines with significant improvements (d > 1.8, p < 0.001). Additionally, STGAT reduces detection delay by 26%, achieving a 2.3-time-step delay while maintaining stable performance under overflow, drift, and physical inconsistencies.
arXiv:2403.11425v4 Announce Type: replace Abstract: Cancer treatments are known to introduce cardiotoxicity, negatively impacting outcomes and survivorship. Identifying cancer patients at risk of heart failure (HF) is critical to improving cancer treatment outcomes and safety. This study examined machine learning (ML) models to identify cancer patients at risk of HF using electronic health records (EHRs), including traditional ML, Time-Aware long short-term memory (T-LSTM), and large language models (LLMs) using novel narrative features derived from the structured medical codes. We identified a cancer cohort of 12,806 patients from the University of Florida Health, diagnosed with lung, breast, and colorectal cancers, among which 1,602 individuals developed HF after cancer. The LLM, GatorTron-3.9B, achieved the best F1 scores, outperforming the traditional support vector machines by 39%, the T-LSTM deep learning model by 7%, and a widely used transformer model, BERT, by 5.6%. The analysis shows that the proposed narrative features remarkably increased feature density and improved performance.
arXiv:2512.01085v3 Announce Type: replace Abstract: Medical phrase grounding (MPG) maps textual descriptions of radiological findings to corresponding image regions. These grounded reports are easier to interpret, especially for non-experts. Existing MPG systems mostly follow the referring expression comprehension (REC) paradigm and return exactly one bounding box per phrase. Real reports often violate this assumption. They contain multi-region findings, non-diagnostic text, and non-groundable phrases, such as negations or descriptions of normal anatomy. Motivated by this, we reformulate the task as generalised medical phrase grounding (GMPG), where each sentence is mapped to zero, one, or multiple scored regions. To realise this formulation, we introduce the first GMPG model: MedGrounder. We adopted a two-stage training regime: pre-training on report sentence--anatomy box alignment datasets and fine-tuning on report sentence--human annotated box datasets. Experiments on PadChest-GR and MS-CXR show that MedGrounder achieves strong zero-shot transfer and outperforms REC-style and grounded report generation baselines on multi-region and non-groundable phrases, while using far fewer human box annotations. Finally, we show that MedGrounder can be composed with existing report generators to produce grounded reports without retraining the generator.
arXiv:2602.00585v2 Announce Type: replace Abstract: Information-seeking agents have emerged as a powerful paradigm for knowledge-intensive tasks, yet today's systems remain specialized for the open web, documents, or local knowledge bases, hindering scalable and cross-domain deployment. We present the first systematic empirical study of consolidating these information-seeking agents into a single foundation agentic model. We compare two paradigms -- \emph{data-level mixing}, which trains a unified model on a mixture of datasets, and \emph{parameter-level merging}, which merges independently trained experts in parameter space -- across 3 training scenarios, evaluating \textbf{26} representative parameter-level methods on \textbf{10} benchmarks. To compare across heterogeneous benchmarks, we introduce a geometric Composite Score and an Imbalance Score that describe overall performance and task skew. Our analysis shows that (i) well-designed parameter-level merging attains parity with data mixing at a fraction of its training cost and is order-agnostic; (ii) parameter-level merging structurally preserves out-of-domain capabilities that data mixing universally forgets; and (iii) cross-scenario stability is strongly tied to consolidation quality. We distil our observations into a method-selection guide and design principles for next-generation merging operators.
arXiv:2602.01472v2 Announce Type: replace Abstract: Large reasoning models (LRMs) typically solve reasoning-intensive tasks by generating long chain-of-thought (CoT) traces, leading to substantial inference overhead. We identify a reproducible inference-time phenomenon, termed Self-Compression: when multiple independent and answerable questions are presented within a single prompt, the model spontaneously produces shorter reasoning traces for each question. This phenomenon arises from multi-question contextual pressure during generation and consistently manifests across models and benchmarks. Building on this observation, we propose ConPress (Learning from Contextual Pressure), a lightweight self-supervised fine-tuning approach. ConPress constructs multi-question prompts to induce self-compression, samples the resulting model outputs, and parses and filters per-question traces to obtain concise yet correct reasoning trajectories. These trajectories are directly used for supervised fine-tuning, internalizing compressed reasoning behavior in single-question settings without external teachers, manual pruning, or reinforcement learning. With only 8k fine-tuning examples, ConPress reduces reasoning token usage by 59% on MATH500 and 33% on AIME25, while maintaining competitive accuracy.
arXiv:2512.18471v2 Announce Type: replace Abstract: Continual learning systems face a fundamental geometric obstacle: as experience accumulates on a fixed-capacity manifold, covering numbers grow linearly with time, eventually forcing representational overlap and catastrophic interference. Prevailing approaches attack this problem by \emph{expansion} - projecting into higher-dimensional spaces via kernels, overparameterization, or replay. We argue the solution is the opposite: \emph{contraction}. We formalize abstraction as the \textbf{Urysohn Ladder}, a hierarchy of quotient maps that recursively collapse validated metric neighborhoods into compact tokens, converting unbounded ambient-space search into bounded navigation on a low-dimensional intrinsic scaffold. Geometrically, each collapsed token acts as a shortcut - a region of extreme metric contraction that bridges distant experiences, much like a wormhole in the representational manifold. We establish four results that collectively guarantee \emph{separability} (metric contraction renders nonlinearly entangled structure linearly separable at each quotient level, and this separability propagates faithfully through the entire hierarchy), \emph{bounded capacity} (covering numbers remain $O(1)$ per quotient level, independent of stream length), \emph{stability} (parity-partitioned flow/scaffold subspaces enable unbounded plasticity without catastrophic interference), and \emph{scalability} (inference cost scales with quotient distance, not ambient distance). We validate each claim empirically with pretrained models and real-world datasets. Moreover, we demonstrate the potential of Urysohn Ladder for scalable continual learning via scaffold amortization.
arXiv:2512.21314v2 Announce Type: replace Abstract: We develop a Lyapunov-based small-gain theorem for establishing fixed-time input-to-state stability (FxT-ISS) guarantees in interconnected nonlinear dynamical systems. The proposed framework considers interconnections in which each subsystem admits a FxT-ISS Lyapunov function, providing robustness with respect to external inputs. We show that, under an appropriate nonlinear small-gain condition, the overall interconnected system inherits the FxT-ISS property. In this sense, the proposed result complements existing Lyapunov-based smallgain theorems for asymptotic and finite-time stability, and enables a systematic analysis of interconnection structures exhibiting fixed-time stability. To illustrate the applicability of the theory, we study feedback-based optimization problems with time-varying cost functions, and Nash-equilibrium seeking for noncooperative games with nonlinear dynamical plants in the loop. For both problems, we present a class of non-smooth gradient or pseudogradient-based controllers that achieve fixed-time convergence without requiring time-scale separation and using real-time feedback. Numerical examples are provided to validate the theoretical findings.
arXiv:2606.25278v1 Announce Type: new Abstract: Multi-modal fusion and multi-model ensembling are prevalent in enhancing the performance of 3D semantic segmentation. Despite the impressive performance, these methods either rely on auxiliary input signals or suffer from costly computational expense. To efficaciously enhance the segmentation performance without introducing intolerable costs, we propose to transfer the rich knowledge from the multi-modal model (i.e., point clouds and images) and multiple model experts to the point-cloudbased network through knowledge distillation. Specifically, we present Information-oriented Heterogeneous Distillation (IHD) to help the uni-modal model absorb the complementary knowledge from the multi-modal teacher. We design the Information-Oriented Filtering (IOF) strategy to select informative images from the continuous image sequence for multi-modal fusion. This practice can boost the performance of the multi-modal teacher, thus benefiting the learning of the student. Besides, as opposed to vanilla model ensembling that requires the separate training of each expert, we propose Adept Snapshot Distillation (ASD). ASD treats the freely available model snapshots generated during the training phase as multiple experts, which significantly reduces the training cost for model ensembling. For each expert teacher, it only provides supervision to the student in the class where it is adept. The resulting Heterogeneous and Adept Snapshot Knowledge Distillation, dubbed HAS-KD, attains state-of-the-art results on ScanNetV2 and S3DIS datasets. HAS-KD can be seamlessly integrated into contemporary 3D segmentation algorithms and bring considerable gains without introducing extra inference burdens. The code will be made publicly available upon publication.
arXiv:2512.24755v5 Announce Type: replace Abstract: Multimodal fusion is the default approach for combining heterogeneous sensor streams in industrial monitoring, yet no systematic method exists for determining \textit{when fusion degrades rather than improves} detection performance. We present an \textbf{Asymmetry-Aware Routing Framework} -- a three-step diagnostic procedure (unimodal performance gap, gate weight attribution, modality corruption testing) with formal decision criteria -- that routes multimodal systems toward the appropriate fusion strategy before deployment. We validate the framework on three datasets spanning two routing outcomes: (1)~the OHT/AGV industrial dataset (thermal + sensors, 13{,}121 samples), where the framework correctly identifies severe asymmetry (gap ratio 3.1$\times$) and recommends \textsc{cascade}; (2)~a chain conveyor fault detection scenario (audio + vibration), where moderate asymmetry leads to a \textsc{fuse} recommendation with positive fusion benefit; and (3)~the CWRU bearing dataset, providing controlled validation in both directions. Threshold sensitivity analysis across all three datasets shows that the framework's recommendations are robust to threshold perturbation, with correct routing maintained over a wide parameter plateau. Comparison against simpler diagnostics (gap ratio alone) reveals that Step~1 alone is ambiguous for moderate-asymmetry cases, demonstrating the necessity of the full protocol for reliable routing decisions.
arXiv:2606.25837v1 Announce Type: new Abstract: This preliminary paper presents an interactive Evidence-RAG workspace for editorial assessment of AI-assisted peer review in the Journal of Digital History. The workflow makes model recommendations easier to inspect by linking reviewer comments, paper evidence, retrieval traces, and reproducibility checks. The system does not replace editors or reviewers. It treats large language models as auditable assistants whose outputs must be checked by human scholars. We describe the current pipeline: paper conversion, semantic chunking, vector indexing, retrieval-augmented evidence assessment, and a lightweight editorial interface. This is a preliminary version of a full paper associated with the accepted presentation "Towards an Interactive Evidence-RAG Peer-Review Workspace for the Journal of Digital History" at the conference AI through History, History through AI, C2DH, University of Luxembourg, 15-16 June 2026. The work was submitted to the conference on 26 February 2026. We also report a first editor-annotation analysis of 80 saved decisions for the Claude-Qwen audit configuration. For Claude-Qwen, strict editor-confirmed accuracy is 70.0%, the correct-or-mostly-correct rate is 86.2%. The useful-output rate includes all responses judged correct, mostly correct, or partially correct, since these outputs can still assist with editorial review even if the model's assessment is not fully accurate. This rate was 90.0%. The full version may include additional experiments, extended evaluation, and a more complete release of materials.
arXiv:2606.25873v1 Announce Type: new Abstract: The sensitivity of fitting Zernike coefficients to wavefront data due to an error or uncertainty in the exact position of the edge of the wavefront data is derived. The results indicate that fitting higher and higher order Zernikes rapidly becomes a useless exercise unless the true edge of the wavefront data is known with sufficient precision.
arXiv:2606.25875v1 Announce Type: new Abstract: Let $n$ and $p$ denote the numbers of vertices and labels, respectively, in an undirected edge-labeled graph. Previous work showed that, under the Exponential Time Hypothesis (ETH), there is no deterministic algorithm with running time \[ (np)^{o\left(\frac{\log n}{(\log\log n)^2}\right)}. \] In this paper, we give a deterministic reduction that strengthens this conditional running-time lower bound to \[ (np)^{o\left(\frac{\log n}{\log\log n}\right)}. \]
arXiv:2606.25297v1 Announce Type: new Abstract: Generative models have significantly advanced image generation, resulting in synthesized images that are increasingly indistinguishable from authentic ones. However, the creation of fake images with malicious intent is a growing concern. Low-configured smart devices have become highly popular, making it easier for deceptive images to reach users. Consequently, the demand for effective detection methods is increasingly urgent. In this paper, we introduce a simple yet efficient method that captures pixel fluctuations between neighboring pixels by calculating the gradient, which highlights variations in grayscale intensity. This approach functions as a high-pass filter, emphasizing key features for accurate image distinction while minimizing color influence. Our experiments on multiple datasets demonstrate that our method achieves accuracy levels comparable to state-of-the-art techniques while requiring minimal computational resources. Therefore, it is suitable for deployment on low-end devices such as smartphones. The code is available at https://github.com/vohoaidanh/adof.
arXiv:2606.25876v1 Announce Type: new Abstract: The Internet is transitioning from Web3 toward Web4, where autonomous agents serve as independent economic actors. These agents can now hold crypto wallets, execute on-chain trades, and pay for external API calls. This transition calls for a new infrastructure stack capable of supporting key agent operations, including agent-to-tool interaction, agent-to-agent payments, and verifiable agent identity, represented by emerging protocols such as the Model Context Protocol, x402, and EIP-8004. Despite growing industrial interest in these protocols, the real-world Web4 agent ecosystem remains largely underexplored. To bridge this gap, we conduct the first large-scale empirical study of the Web4 ecosystem. Specifically, our study targets three interconnected questions: how Web4 agents are deployed and used in practice; what engineering challenges developers face when building Web4 agents; how current project communities respond to these challenges. To answer these questions, we analyze 99,448 multi-chain identity registrations, 317,596,323 transaction logs, the source code of 341 MCP projects, and 349 filtered GitHub issues. Our findings reveal that autonomous agents have established a highly active machine-to-machine payment economy, processing millions of daily transactions. However, this growth is built on immature infrastructure, including identity/authorization practice, cross-environment operation, and payment interoperability. Our follow-up analysis shows that community responses are visible but unevenly distributed across repositories, and payment interoperability remains the most persistent unresolved bottleneck. Overall, this study reveals a critical gap between the rapid growth of the Web4 agent economy and its fragile underlying infrastructure, highlighting future directions for building a more secure Web4 agent ecosystem.
arXiv:2606.25307v1 Announce Type: new Abstract: With the rapid growth of the number of academic papers, systematically evaluating the difficulty of research and its relationship to academic impact offers important significance for research topic selection and resource allocation. However, current studies lack quantitative assessments of research difficulty and its correlation with academic impact. This paper proposes a comprehensive evaluation system for research difficulty, incorporating factors such as academic collaboration, content, and references. Taking the field of Natural Language Processing (NLP) as a case study, we extract both internal and external features from academic papers, compute multiple research difficulty indicators. We assign their weights using the entropy weight method and perform a weighted sum to obtain the research difficulty score of academic papers. This paper uses the citation frequency of academic papers to measure academic impact. To validate our approach, NLP experts assessed the difficulty of a sample of papers, and correlation analyses confirmed the reliability of our measurement. Empirical results reveal that in NLP, factors such as the number of pages, reference count, and participation of high-level institutions are significantly associated with academic impact. Moreover, we identify an inverted U-shaped relationship between research difficulty and academic impact. It suggests that moderately difficult research tends to achieve greater academic impact.
arXiv:2606.25319v1 Announce Type: new Abstract: Fine-grained visual reasoning requires multimodal large language models (MLLMs) to identify task-relevant visual evidence and ground their reasoning in local image regions. Existing agentic methods typically rely on reinforcement learning with verifiable rewards or supervised fine-tuning on large-scale annotated reasoning traces, leading to costly exploration, hand-designed verification rules, or heavy dependence on textual supervision. A natural way to avoid such external answer labels is to learn from trajectories sampled by the student itself, which points to On-Policy Distillation (OPD). To understand what OPD can and cannot provide for visual reasoning, we revisit it as negative-free stop-gradient alignment. This perspective shows that, although OPD provides effective token-level correction, its ceiling is constrained by the absence of trajectory-level discrimination. Motivated by these observations, we propose V-Zero, an answer-label-free framework for visual reasoning with contrastive evidence gating. V-Zero uses no annotated textual answer labels; instead, during training it pairs a question-relevant regional crop with a negative visual view to evaluate student-sampled trajectories and gate dense token-level distillation. Experiments on multiple visual reasoning benchmarks show that V-Zero consistently improves fine-grained visual reasoning while preserving strong generalization. Notably, V-Zero is more than 5$\times$ faster than previous supervised fine-tuning methods and more than 10$\times$ faster than reinforcement learning baselines. Code and dataset will be released at https://github.com/eVI-group-SCU/V-Zero
arXiv:2606.25356v1 Announce Type: new Abstract: Large Language Models (LLMs) are increasingly used for automated vulnerability detection, but it remains unclear how program structure and semantics should be represented for LLM-based reasoning. Most prompting-based approaches provide raw source code, implicitly assuming that more source-level context gives the model better evidence. This paper challenges that assumption through RepBench, an empirical benchmark comparing raw source code with static-analysis-based program representations. RepBench converts real-world C/C++ vulnerability testcases into multiple representations: raw source, Abstract Syntax Trees (ASTs), Control-Flow Graphs (CFGs), Program Dependence Graphs (PDGs), their combinations, and an auxiliary track of enriched PDGs (ePDGs). Using a curated PrimeVul-derived corpus of 107 Joern-based testcases across five CWE categories, we evaluate ten representation variants under a fixed Chain-of-Thought and structured-output protocol, plus 19 additional ePDG cases generated through VulChecker/Hector. Representation choice substantially affects LLM vulnerability reasoning. The strongest variant, AST+PDG, achieves 83.2% accuracy, compared with 53.5% for raw source. At the family level, graph-only prompts outperform both source-only and source-plus-graph prompts while requiring far less prompt overhead. These results reveal a context dilution effect: adding raw source code to compact structural graph evidence can degrade reasoning by making vulnerability-relevant evidence less salient. Overall, our findings show that carefully selected structural representations offer a better accuracy-overhead tradeoff than simply giving LLMs more raw input, and suggest that static analysis can serve as an effective prompt-construction layer for security-focused LLM reasoning.
A fast scheme for the homogeneous Boltzmann equation based on lifting and tensor train approximation
arXiv:2606.25914v1 Announce Type: new Abstract: We propose a fast deterministic scheme for the space-homogeneous Boltzmann equation that exploits the low-rank structure of the velocity distribution. This paper consists of two independent contributions. The first is a \emph{lifting-projection (LP) scheme}, inspired by the approach in the recent theoretical breakthroughs \cite{guillen2025landau, imbert2026monotonicity, guillen2025landau2} on the well-posedness of the Landau and Boltzmann equations. In particular, the approach lifts the nonlinear 3D Boltzmann equation to the 6D linear Kac master equation, advanced over a single time step, and projected back to its marginal in 3D. The second contribution is a \emph{low-rank tensor method} for evaluating the collision operator, in which the lifted solution is represented in tensor train (TT) format and computed via a TT cross approximation algorithm with interpolation, complemented by a TT-friendly conservation correction that enforces conservation of mass, momentum, and energy. When the solution is low-rank in velocity, the method scales linearly in $n$ when cubic interpolation is used (and quadratic in $n$ when spectral interpolation is used), where $n$ is the number of grid points in each velocity direction. Therefore, our methods offer significant computational savings over existing deterministic solvers in such cases. Numerical experiments on 2D and 3D benchmarks, including the BKW exact solution and anisotropic initial data, confirm the computational scaling, the expected order of accuracy and verify the effectiveness of the conservation correction.
arXiv:2605.18105v2 Announce Type: replace Abstract: Landslides often hit newsstands due to their destructive and potentially fatal effects. News are a valuable source of information for creating or enriching disaster databases and for expediting media-based studies of the dynamics of media attention. To accomplish that, news datasets must be filtered, geolocated and validated. This paper focuses on how landslides around the world are reported in German newspapers. We analyse almost 55k news articles about 4.5k news events in a 25-year period, compare it with external measures of countries' susceptibility to landslides and provide insights, e.g. the overreporting of Southern and Western Europe, to foster further studies on inequalities in media attention to international disasters.