arXiv:2607.17765v1 Announce Type: new
Abstract: We introduce WC2026-Agents, a benchmark and dataset for evaluating large language models (LLMs) as autonomous forecasting agents on real, future events. For every one of the 104 matches of the 2026 FIFA World Cup, four frontier models -- Claude Opus 4.8, ChatGPT (GPT-5.5, high reasoning), Gemini 3.1 Pro, and Grok (Expert Mode) -- ran an identical search-act-reflect loop: gather evidence with a web tool, commit to a 1X2 (team-A win / draw / team-B win) distribution and a virtual 100-USD bet, and, after the match, reflect given only the final score. Because every match kicked off after the models' training cutoffs, the benchmark is contamination-free by construction. Crucially, we pair the four agents with a fifth competitor drawn from the same information environment -- the pre-match betting market -- collected as per-match 1X2 odds, giving an economically grounded baseline and letting us score not just what an agent predicts but what it does with money. The release contains 416 forecasts and 414 reflections with verbatim reasoning, ground truth (including penalty shootouts), odds, and a reproducible evaluation suite. A reference evaluation surfaces findings that raw accuracy hides: the four agents issue an identical top pick in 92% of matches and none beats the market's Brier score; indeed, a naive flat stake on the market favorite out-earns all four agents. Yet the agents diverge sharply as decision-makers: betting return-on-investment ranges from -18% to +10%, fading the market is unprofitable for all four, the share of forecasts that cite the market ranges from 12% to 100%, and self-reported error rates on wrong picks range from 36% to 86%. The benchmark thus measures calibration, decision quality, and self-knowledge -- axes on which frontier models differ even when their predictions do not. Data and code: https://github.com/graphuofm/FIFA2026LLM
Science Journals
arXiv:2607.17839v1 Announce Type: new
Abstract: Collaborative robots are increasingly sharing workspaces with human operators, making tool handover a frequent and safety-critical micro-interaction. However, traditional static handovers often lead to awkward grasps when handling asymmetric industrial tools. This paper presents a receiver-centered voice-driven adaptive handover system for mechanical tools, built on a Franka cobot. Using an LLM for intention recognition and MediaPipe for real-time 3D hand tracking, the framework dynamically adjusts the end-effector's orientation to present tools in an ergonomically optimal, handle-first pose. A within-subjects study compared this adaptive approach with an object-agnostic static baseline. The results demonstrate that the adaptive system reduces the grasp delay for asymmetric tools, improving the fluency of the interaction. Furthermore, the adaptive strategy improved specific trust-related perceptions, particularly motion predictability and perceived task simplicity.
arXiv:2607.18086v1 Announce Type: new
Abstract: Background: LLM judges increasingly score whether clinical language models give overconfident answers under incomplete evidence, yet whether a measured "safety gain" reflects real behavior change or the judge's calibration is unresolved. Using a structured evidence-sufficiency prompt as a test case, we asked whether it reduces unsafe overconfident answers, how far that effect depends on the scoring judge, and what it costs in helpfulness.
Methods: In a retrospective public-data benchmark (Real-POCQi, HealthBench, MedRBench), four models (GPT-5.5, Claude Opus 4.8, Gemini 3.5 Flash, Grok 4.3) answered a fully paired common panel (1,200 cells) with a standard prompt and the wrapper. The pre-specified endpoint was the paired reduction in unsafe overconfidence scored by the primary judge (GPT-5.4-nano); secondary analyses added a different-family judge (Claude Sonnet 5), a correctness judge, matched scaffold controls, and a blinded three-clinician review.
Results: Unsafe overconfidence fell from 49.3% to 24.7%, a paired reduction of 24.7 points (95% CI 21.8-27.7; p<0.001), robust in direction across models and paraphrases. Magnitude was judge-dependent: Sonnet agreed on direction but nearly halved the effect (+13.1 points), with one-directional disagreement. Blinded clinicians characterized the primary judge as a high-sensitivity (1.00), low-specificity (0.55) screen, not a calibrated rate. The gain carried a model-specific helpfulness cost (correct diagnosis 80.3% to 50.3%): near-free for GPT-5.5, near-total for Gemini (-58 points). Matched scaffold controls showed genuine behavior change, not judge circularity.
Conclusions: LLM-judged clinical safety effects should be reported as directional and relative, anchored to human review and evaluated jointly with helpfulness, not as calibrated absolute rates. This does not establish clinical deployment readiness.
arXiv:2607.18018v1 Announce Type: new
Abstract: We propose two new approaches for constructing families of rational functions with shared real poles that nearly uniformly approximate the functions $\exp(-tz)$ for $z\geq 0$ and $t$ in a positive time interval. The first result concerns the case where all real poles coalesce into a single point. With an appropriate choice of a weight function we are able to derive a closed formula for the asymptotically optimal location of such a pole. We then discuss the more general case where all real poles are distinct. Using Zolotarev's construction of certain optimal rational functions, we present a simple algorithm to derive nearly optimal poles efficiently. We analyze the stability of the numerical evaluation of the resulting rational matrix functions in floating-point arithmetic. By controlling the growth of potential ill-conditioning arising from partial fractions, reliable and highly parallelizable exponential propagators are obtained.
arXiv:2607.18102v1 Announce Type: new
Abstract: Financial question answering over U.S. Securities and Exchange Commission (SEC) filings requires retrieving and synthesizing heterogeneous evidence dispersed across long, standardized, and highly redundant disclosures. Existing retrieval-augmented and multi-agent systems typically derive retrieval queries directly from the user's question and rank candidates by semantic similarity. Together, these choices create prior-corpus misalignment: a mismatch between model priors and the target filings' structure, terminology, and evidence standards. As a result, query generation misses corpus-specific evidence, while semantic reranking favors topically similar but evidentially invalid false-positive chunks. We propose FinSAgent, an evidence-grounded multi-agent framework that reframes SEC filing QA as corpus-aligned retrieval planning and corrects both ends with a single principle: inject corpus-side conditioning wherever model priors would otherwise dominate. FinSAgent combines (1) role-specialized agents anchored to the mandated 10-K item structure, (2) database-aware query decomposition that conditions each agent's sub-queries on a lightweight, summary-level view of the local corpus, and (3) multi-path retrieval with a learned feature-gated reranker that separates evidential validity from semantic similarity. Across five offline financial QA benchmarks, FinSAgent improves retrieval coverage and answer correctness over strong single-agent and multi-agent baselines; in a three-arm randomized online experiment with 1,000 anonymous user ratings, it also receives higher scores than baselines.
arXiv:2607.17880v1 Announce Type: new
Abstract: Mixture-of-Experts (MoE) models route each token to only a few expert networks, distributing the serving load across experts whose popularity shifts over time. A serving system must therefore dynamically decide how many GPUs to assign to each expert, trading off service latency against the cost of reconfiguring the assignment. We introduce a formal model of MoE Serving and initiate a principled study of online and offline algorithms for it. Our main result is a polynomial-time $O(\sqrt{\log k})$-competitive online algorithm, where $k$ is the number of GPUs beyond one per expert. We complement it with a matching $\Omega(\sqrt{\log k})$ barrier for the online dual problem underlying our analysis. In the offline setting, we give a constant-factor approximation, show that MoE Serving is NP-hard, and rule out an FPTAS assuming ETH.
arXiv:2607.17789v1 Announce Type: new
Abstract: Early and timely screening of laryngeal cancer is crucial for improving clinical outcomes. In recent years, NBI endoscopy has become a standard diagnostic tool for the detection of laryngeal lesions. However, its effective use requires well-trained clinicians and the procedure is time-consuming and subject to interobserver variability. In this context, the application of artificial intelligence (AI) offers a promising solution to support clinical decision-making. In this work, we proposed applying transformer and attention mechanism for analyzing the narrow band imaging and distinguish benign and malignant lesions. Results show it has good classification performance with F1 (82.72%), accuracy(82.33%). In addition, the result of laryngeal cancer screening is explainable for clinicians. The explainability is utilizing the state of art segmentation method (MedSAM) to provide the useful pathological information area for clinicians. The proposed methodology fusing classification and segmentation provides a translating on laryngeal cancer screening.
arXiv:2607.17896v1 Announce Type: new
Abstract: 2D Gaussian Splatting is an attractive direction for image representation due to its explicit formulation, fast rasterization, and favorable decoding efficiency. The representation quality of this paradigm depends on the proper allocation of Gaussian capacity to the demanding regions. However, existing methods fail to allocate Gaussian capacity efficiently during optimization: under-reconstructed content is often refined in a fragmented pixel-wise manner, while neighboring optimized Gaussians with similar attributes are redundantly retained. This inefficiency motivates the need for a density control framework that jointly addresses insufficient allocation in under-reconstructed regions and redundant allocation in over-reconstructed regions. Our key insight is that this framework should exploit two complementary forms of locality: the local continuity of reconstruction errors in image space for improved Gaussian allocation, and the local similarity of neighboring Gaussians in Gaussian space for redundant elimination. Based on this insight, we propose Locality-Aware Density Control (LocoADC), a plug-and-play framework that improves Gaussian capacity utilization through Region-wise Gaussian Densification (RGD) and Similarity-Driven Gaussian Merging (SDGM) strategies, together with a local color consistency constraint for more reliable merging. Extensive experiments on diverse datasets show that LocoADC consistently improves multiple baselines by enabling more effective local Gaussian allocation, including a 2.93 dB PSNR gain over GI on the CLIC dataset under the same 30k Gaussian budget. Code is available at: \textit{https://github.com/ChenJiaCong-1005/LocoADC}.
arXiv:2607.17930v1 Announce Type: new
Abstract: We consider the combination of the two constrained planarity problems Level- and Clustered Planarity. Traditionally, level-planar drawings with convex clusters have been studied in this setting. Fink et al. (EuroCG 2024) recently introduced a different way of combining level- and clustered planarity by mimicking a classic characterization of clustered planarity in the level-planar setting: The problem (y-)monotone Clustered Level Planarity (mCLP) seeks a level-planar drawing in which it is possible to augment each cluster with edges that do not cross cluster boundaries so that it becomes connected while maintaining level-planarity. This is in line with previous research on clustered planarity that poses certain requirements on the augmentation edges that make each cluster connected, e.g., that they form a path. Fink et al. (EuroCG 2024) showed that mCLP is NP-complete even for biconnected single-source graphs and instances with a constant number of levels and clusters.
We further classify the parameterized complexity of the mCLP problem by, on the one hand, showing hardness even for instances that consist of a forest with trees of bounded size, no isolated vertices, and a small constant number of either clusters or levels. This excludes fixed-parameter tractability for almost all graph-structural parameters, except for vertex cover, even in conjunction with the number of clusters. We complement this by showing fixed-parameter tractability when parameterizing by the vertex cover number and the number of clusters. A major obstacle is the fact that mCLP is non-hereditary, i.e., subinstances of yes-instances may be no-instances and vice versa, which makes it challenging to apply usual reduction techniques.
arXiv:2607.17835v1 Announce Type: new
Abstract: Many state-of-the-art DSP implementations use fixed-point arithmetic due to its reduced hardware complexity and high throughput compared to conventional floating-point arithmetic. In contrast, machine learning accelerators exhibit substantial gains from reduced-precision floating-point formats, enabling improvements in energy efficiency and peak throughput. These advances motivate a re-evaluation of numerical representations for classical DSP workloads. A central question is whether reduced-precision floating-point formats can achieve competitive power, performance, and area compared to fixed-point implementations, while providing advantages in dynamic range and numerical robustness. This paper presents a new cross-layer co-design methodology for DSP kernels that jointly optimizes numerical representations, arithmetic units, and application-level performance. As a case study, we focus on the FFT, a fundamental DSP block across many applications. The FFT is evaluated within optical OFDM transceivers, where it dominates power consumption and silicon area as FFT size and modulation order scale to support data rates beyond 100 Gbit/s. We compare fixed-point and reduced-precision floating-point formats using post-layout power and area results in a 12nm FinFET technology and demonstrate system-level performance in terms of BER versus Eb/N0. For a 256-point FFT engine in a 128 Gbit/s transceiver, we show that 11- and 12-bit custom floating-point formats preserve BER performance close to a 32-bit floating-point reference across multiple modulation orders, while reducing FFT core power by up to 19.8% and area by up to 12.0% compared to representative fixed-point designs. To the best of our knowledge, this is the first investigation of custom reduced-precision floating-point arithmetic for FFT cores in optical OFDM transceivers.
arXiv:2607.17935v1 Announce Type: new
Abstract: Automated fact-checking remains a challenge for Large Language Models (LLMs) due to "query brittleness" in traditional retrieval systems. We propose DeLIVeR (Decomposed Learning for Information-grounded Veracity Recognition), a framework that treats evidence retrieval as a reinforced strategic exploration task. DeLIVeR utilizes a Planner LLM to decompose complex claims into targeted question sets, which are used to traverse structured Knowledge Graphs (KGs) for high-precision evidence. We optimize the Planner's policy using Group Relative Policy Optimization (GRPO) with a reward system prioritizing structural diversity and verdict accuracy. Our evaluation on LIAR, FEVER, and PolitiFact shows that DeLIVeR significantly outperforms state-of-the-art baselines. Using Qwen2.5-7B, our framework achieved peak F1-scores of 83.73, 84.57, and 79.70 respectively, representing a 10-15% improvement over HippoRAG2. By shifting to a reinforced question-planning strategy, DeLIVeR effectively bridges multi-hop reasoning gaps and provides an auditable, transparent path for verifiable misinformation detection.
arXiv:2507.05515v3 Announce Type: replace
Abstract: Vision-language models (VLMs) are facing the challenges of understanding and following multimodal assembly instructions, particularly when fine-grained spatial reasoning and precise object state detection are required. In this work, we explore LEGO Co-builder, a hybrid benchmark combining real-world LEGO assembly logic with programmatically generated multimodal scenes. The dataset captures stepwise visual states and procedural instructions, allowing controlled evaluation of instruction-following, object detection, and state detection. We introduce a unified framework and assess leading VLMs such as GPT-4o, Gemini, and Qwen-VL, under zero-shot and fine-tuned settings. We also evaluated the framework using a reasoning-focused model, GLM-4.1-thinking. Our results show that while object detection achieved high performance (98.16% with fine-tuned InstructBLIP), fine-grained scene understanding and assembly state detection remain challenging: Fine-tuned MiniGPT-v2 reached only 37.52% F1 for identifying theme entities, and even advanced models such as GPT-4o achieved just 40.54% F1 on state detection. This highlights gaps in fine-grained visual understanding among existing models. We release the benchmark, codebase, and generation pipeline to support future research on multimodal assembly assistants grounded in real-world workflows.
arXiv:2607.17873v1 Announce Type: new
Abstract: The increasing penetration of inverter-based resources has led to a significant reduction in system inertia, resulting in faster and more pronounced frequency deviations in modern low-inertia power systems. In such environments, the dynamic behavior of electrical loads becomes increasingly important in shaping overall system frequency response. This paper presents an enhanced load model that incorporates load-side dynamics in addition to conventional static behavior, thereby augmenting the representation of load-frequency control (LFC) models. This improved formulation increases the accuracy of frequency response studies in power systems. A comparison between the proposed augmented model and the conventional LFC representation demonstrates that relying solely on static load modeling can lead to inaccurate results and potentially misleading conclusions. Therefore, accurate modeling of load-side dynamics is essential for reliable frequency stability assessment in modern power systems.
arXiv:2607.18195v1 Announce Type: new
Abstract: Vision models have been found to be susceptible to perturbations such as motion blur induced at runtime by a shaking camera. This impedes their deployment in critical applications since phenomena such as slightly blurred vision might lead to failures, for example an object detector missing objects. While methods such as data augmentation or Adversarial Training can improve empirical robustness, they lack formal safety guarantees, making it difficult to identify and mitigate hidden vulnerabilities. We introduce a novel Certified Training approach that leverages an efficient encoding of convolutional perturbations to train provably robust models. Our method significantly outperforms Adversarial Training, achieving, for example, over 80% robust accuracy against motion blur of reasonable intensity on CIFAR10 while maintaining comparable standard accuracy.
arXiv:2607.17981v1 Announce Type: new
Abstract: Representation learning has enabled classical exploration strategies to be extended to deep Reinforcement Learning (RL), but often makes algorithms more complex and theoretical guarantees harder to establish. We introduce Random Feature Information Gain (RFIG), grounded in Bayesian kernel methods theory, which uses random Fourier features to approximate information gain and compute exploration bonuses in non-countable spaces. We provide error bounds on information gain approximation and avoid the black-box aspects of neural network-based uncertainty estimation, for optimism-based exploration. We present practical details that make RFIG scalable to deep RL scenarios, enabling smooth integration into standard deep RL algorithms. Experimental evaluation across diverse control and navigation tasks demonstrates that RFIG achieves competitive performance with well-established deep exploration methods while offering superior theoretical interpretation.
arXiv:2607.17994v1 Announce Type: new
Abstract: Video understanding has become more and more important with the growth of Artificial Intelligence (AI) for video generation. Recently, Multimodal Large Language Model(M-LLM) has shown its capability in video understanding. Video summarization, a specific domain of video understanding, has proven its importance for efficient navigation and retrieval. Both video understanding and video summarization require a good selection of key frames in a video. Current video summarization methods heavily focus on the selected key frames and correlated segment captions. However, existing approaches overlook the perspective of treating the importance of the frames globally. We argue that using discrete selected frames for summarization will not only reduce the understanding coherence, but also lost important information in the video, as well as wasting the original capacity of the MLLMs. In this paper, we propose HAS, a Highlight-guided Attention Steering method for video summarization. We consider a challenging but practical setting where the video given to MLLMs for summarize should be continuous but with highlight guidance. HAS mainly consists of two parts: The first part is to find a continuous frame-level highlight distribution for the video globally. The second part is to apply the highlight distribution as an attention steering vector for the MLLM, targeting a better understanding of the video, and thus during the model inference time, putting more attention on the highlighted frames, while avoiding lost entire information on less highlighted frames through putting less attention instead of forgetting them. We evaluated HAS on a variety of benchmarks, and it has shown convincing performance in video summarization.
arXiv:2607.16583v1 Announce Type: cross
Abstract: Kirigami turns stiff sheets into compliant, shape-morphing structures, but its reliance on periodic cut patterns comes at a cost: correlated panel rotations couple extension to shear, so stretching one axis drives a parasitic shear that cannot be suppressed, and also confine anisotropic stiffness to a narrow, discrete set of responses that cannot be tuned independently. Biological tissues overcome an analogous constraint through controlled disorder, such as graded fiber orientations in skin and hierarchical anisotropy in myocardium, achieving direction-dependent mechanics unavailable to regular architectures. Here, we show that engineered disorder is a design degree of freedom for kirigami, with stochastic kirigami accessing a continuous and far broader region of mechanical response than periodic patterns. This includes programmable anisotropy with near-complete elimination of extension-shear coupling. Because disordered patterns lack a simple parameterization, we navigate this design space with a geometry-aware graph neural network (GNN) that maps cut topology to the full nonlinear, bidirectional stress-strain response, coupled to a genetic algorithm that inverse-designs patterns reproducing target responses along two perpendicular axes. The GNN trains an order of magnitude faster and more accurately than image-based models. Fabricated elastomer samples reproduce the predicted nonlinear, anisotropic responses, closing the loop from design to physical component. By turning disorder into a variable to control directional stiffness, this work develops architected materials that stretch without parasitic shear, from soft actuators to tissue-interfacing devices matched to the anisotropy of living tissue.
arXiv:2607.18163v1 Announce Type: new
Abstract: PPO and the GRPO baseline studied here use clipped surrogate objectives whose favorable-direction saturation introduces an abrupt change in the scalar objective's derivative. We ask whether Output Reset (OR), a smooth one-sided saturation rule, offers a useful alternative for large language model post-training. PPO-OR and GRPO-OR replace the clipped policy term with an OR squared-margin loss in rollout-relative token log-ratio space; the advantage sign determines the update direction, and a token contributes zero direct OR residual after crossing the favorable margin. We compare PPO-clip with PPO-OR under generalized advantage estimation (GAE), and GRPO with GRPO-OR under group-relative advantages, using \texttt{Llama-3.2-1B-Instruct} on Anthropic \texttt{hh-rlhf} with one shared reward model and three seeds per method. Under GAE, PPO-OR has a mean final training-time reward-model score $0.305$ higher than PPO-clip, with a larger observed across-seed spread. Under group-relative advantages, GRPO-OR does not have a higher mean score, but shows a smaller observed spread, a near-zero terminal OR residual, and a declining overshoot fraction, while the matched GRPO clipped-objective trace remains variable. Both group-relative methods exhibit substantially larger rollout-to-current log-ratio displacement than the GAE methods, and OR does not consistently reduce it. Thus, OR changes optimization behavior in both matched comparisons, but the observed reward effect differs between them. At $G=2$, the GRPO-OR diagnostics do not translate into a reward-score gain. Whether larger groups change this outcome remains open. The reported scores are training-time reward-model measurements, not held-out human-preference performance.
arXiv:2607.18021v1 Announce Type: new
Abstract: Differential privacy (DP) imposes fundamental trade-offs between privacy and statistical fidelity in synthetic data generation. While access to public data has been shown to improve these trade-offs empirically, existing approaches use public data only indirectly, through pre-processing (e.g., using pre-trained generative models) or post-processing steps (e.g., matching target statistics estimated from public datasets), while relying on domain-agnostic DP mechanisms. In this work, we lay the theoretical framework to study the principled incorporation of public data into DP mechanisms themselves. We consider normalized histograms as distribution estimators and characterize the asymptotically optimal domain-aware privacy mechanism within a specific class of DP mechanisms. We introduce PubMix, a public-data-aware DP mechanism that can be used in histogram-based data synthesis pipelines. Our experiments demonstrate that PubMix significantly improves synthetic data generation quality compared to domain-agnostic privacy mechanisms.
arXiv:2607.16504v1 Announce Type: cross
Abstract: In Dynamic Data Driven Applications Systems (DDDAS), non-linear continuous-discrete (CD) tracking algorithms have been proposed to recursively estimate stochastic processes which follow continuous-time stochastic differential equations (SDE) using non-linear discrete-time measurement sequences. In this paper, a new filter in this class which is based on the polynomial chaos expansion (PCE) is proposed as an alternative solution to these tracking problems. Using the orthogonality properties of the PCE basis, PCE coefficient-wise prediction and update steps are derived via the Galerkin projection. This differs from previous PCE based filters where collocation points are independently propagated through the motion model and the PCE is recomputed at each time step. For this reason, we call the proposed filter the CD-PCE coefficient filter (CD-PCE-CF). A case study is provided where the CD-PCE-CF and the CD-Extended Kalman Filter (CD-EKF) are used to track a ballistic object undergoing process noise using radar measurements. It is shown that the proposed method significantly outperforms the CD-EKF in terms of estimation accuracy and stability.
arXiv:2607.17953v1 Announce Type: new
Abstract: Inspired by the dynamical PageRank framework of Gleich and Rossi in 2012, we introduce a continuous-time PageRank model in which the personalization vector evolves as a weighted average of its past values, with the weights determined by a memory function. The resulting dynamics are formulated as an initial value problem for an integro-differential equation, where the initial condition is a probability vector. We investigate how the choice of memory function influences the long-time behavior of the PageRank vector. In particular, for strongly connected networks $\mathcal{G}$, we prove that broad classes of memory functions lead to convergence toward a stationary state that is independent of the initial condition. In contrast, when the memory function is exponential-oscillatory, $\omega(t)=e^{at}\cos(bt)$ for $t\geq0$ with $a,b>0$, we show that the PageRank dynamics exhibit asymptotically periodic behavior, revealing that oscillatory memory can fundamentally alter the qualitative evolution of the ranking process. To establish these results, we first prove the existence and uniqueness of solutions using standard results from the theory of integro-differential equations and show that the solution remains a probability vector for all times, thereby preserving the essential properties of the PageRank model.
arXiv:2607.18063v1 Announce Type: new
Abstract: LLM-based agents process external content, exposing them to prompt injection and multi-turn manipulation. Most safety benchmarks evaluate defenders against fixed attack pools collected before evaluation, single-turn or multi-turn. We present a 21-scenario benchmark for \emph{adaptive multi-round attacks against memoryless LLM defenders}: an autonomous LLM attacker observes prior defender responses and pivots across rounds, while each defender response is evaluated as a fresh interaction. Holding the 21 scenarios, attackers, defenders, and structured-output scoring fixed, restricting scoring to the first attacker turn yields $0$-$1\%$ attack success rate (ASR); allowing 15 rounds of adaptive attack yields $5.4$-$14.0\%$. Pooling three frontier attacker LLMs uncovers $1.4$-$2.2\times$ as many unique successful attacks as the best single attacker, and the generated attacks have low cosine similarity ($0.02$-$0.14$) to attacks in existing benchmarks. Claude Opus 4.6 and GPT-5.4 are tied in aggregate ($5.4\%$ each; overlapping $95\%$ CIs), but their weaknesses differ sharply: on one scenario Opus reaches $60\%$ ASR ($95\%$ CI $36$--$80\%$) while GPT-5.4 and Gemini each stay at $7\%$ (CI $1$-$30\%$; the gap is preserved in a higher-$N$ replication). $13$ of $21$ scenarios distinguish at least one defender pair, yet rankings disagree across scenarios (Kendall's $W = 0.19$). We release the benchmark -- 21 evaluation scenarios, 10 public development scenarios, the orchestrator, baseline harnesses, and a multi-attacker CLI -- plus 945 transcripts from the 3$\times$3 frontier matrix, an attack-replay dataset, and 18{,}422 gpt-oss-20b battles from an open competition's final scoring rounds.
arXiv:2607.18187v1 Announce Type: new
Abstract: Large-scale scientific simulations generate volumetric data at rates that far outpace advances in storage and network bandwidth, making effective lossy compression increasingly critical. However, conventional compressors often struggle to preserve fine structural details at high compression ratios (CRs), and implicit neural representations (INRs) require costly per-volume optimization and produce models with fixed CRs. To respond, we present EVOLVE, an autoencoder (AE)-based volume-compression framework that targets high CRs for offline compression, with three key contributions. First, we construct a large-scale cross-domain database of 6,376 volumes from 21 scientific simulations, curated via perceptual hashing to ensure diversity, enabling the optimized model to extract features that generalize across volumes within the covered scientific simulation domains. Second, we reexamine the design space of AE-based compressors and incorporate several macro- and micro-designs into a vanilla AE to develop EVOLVE, which substantially improves the expressive power and compression capability. Third, we develop a learnable gain mechanism with a three-stage training strategy to enable variable-rate encoding, allowing a single model to support continuous CR adjustment at inference time. Experiments on multiple unseen scientific simulation datasets demonstrate that EVOLVE achieves substantially higher CRs than conventional compressors at comparable reconstruction quality, while delivering compression speeds that are orders of magnitude faster than INR-based methods, highlighting its promise as a strong alternative for compressing scientific data. The code, model weights, and results are available on our project page at https://evolve-vis.github.io.
arXiv:2607.18203v1 Announce Type: new
Abstract: A recent paper [Hazeltine, et al., Phys. Plasmas 33, 072501 (2026)] has questioned whether the standard result, (ultimately) due to Ferraro, that the plasma angular velocity is approximately constant along individual equilibrium magnetic field-lines in a rotating axisymmetric magnetic mirror machine, continues to hold when the rotation becomes sonic or supersonic. In order to resolve this issue, the equilibrium of a rapidly rotating mirror is investigated, starting from first principles, using an ideal two-fluid model with anisotropic pressure. It is found that, as long as the ion gyro-radius is much less than the machine size, and the angular velocity of the plasma is much less than the ion gyro-frequency, the Ferraro result holds good.
arXiv:2607.18230v1 Announce Type: new
Abstract: Modern vision-language models (VLMs) have significantly improved image generation and editing capabilities, making pixel-level image tampering detection increasingly important yet challenging under cross-model and out-of-distribution shifts. This work studies domain generalization for pixel-level image tampering detection in modern VLMs like ChatGPT, Gemini, Qwen-Image, etc., aiming to learn tampering localization models that remain robust across diverse VLM-generated manipulation distributions. We propose a simple yet effective domain-generalized training framework built on two practical strategies. First, we introduce a balanced minibatch sampling scheme that strategically samples tampered and real images in each minibatch, preventing biased optimization toward either manipulated artifacts or clean-image priors and avoiding training collapse, ensuring that each optimization step receives proper sampled gradient signals. Second, we adopt a simple late-injection strategy, where the detector is first trained on large-scale base data until stable convergence, and then exposed to a small amount of newly selected supporting data from emerging VLM distributions, improving adaptability without overfitting to limited new domains. Together, these components provide a simple yet strong recipe for improving pixel-level tampering localization and OOD robustness across modern VLMs. Despite the conceptual simplicity, our framework outperforms the prior state-of-the-art PIXAR by a large margin of 26.1% and 26.8% relative improvement in average gIoU and cIoU, respectively, across OOD VLMs of GPT-Images-2.0, Gemini-3.1, FLUX.2, and Seedream 4.5. Our code is available at https://github.com/VILA-Lab/PIXAR-DG