Cryptography and Security
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Monday 7 September
Large language model (LLM) agents increasingly achieve long-horizon tasks by combining foundation models with explicit skills and implicit procedural knowledge acquired through execution. The resulting task-solving capabilities have become valuable proprietary assets, raising a new security question: can a substantially weaker attacker-controlled agent acquire the capabilities of a stronger proprietary agent through limited black-box interaction? Existing skill-stealing attacks recover explicit skill artifacts, yet we show that artifact leakage does not necessarily transfer capability: a weaker agent may possess the same skills but still fail because it lacks procedural behaviors implicitly realized by the stronger agent. Our key insight is that the skill execution gap itself forms a leakage surface, where missing behaviors are exposed through observable differences between successful victim executions and failed attacker executions. Based on this, we present AgentLeak, a black-box capability-cloning attack that identifies capability-critical behaviors from these execution differences and incorporates them into attacker-side skills, while keeping the attacker's model, harness, and tools unchanged. Across 20 task scenarios comprising 600 instances, diverse agent systems, and multiple backbone models, AgentLeak improves task pass rates by over 40% compared with direct skill reuse and recovers more than 80% of the victim–attacker capability gap. Our findings reveal a confidentiality risk in LLM agents: protecting explicit artifacts alone is insufficient, as observable execution behavior can leak the procedural knowledge required to reconstruct proprietary task-solving capabilities in low-capability and attacker-controlled agents.
High Energy Physics (Phenomenology)
Majoron-driven spontaneous leptogenesis in type-II seesaw model
We investigate a minimal realization of spontaneous leptogenesis in the type-II seesaw model, where the lepton and Higgs asymmetries are generated as decay and inverse-decay processes drive the plasma toward the displaced equilibrium dictated by a rotating Majoron background. We derive the complete set of Boltzmann equations governing the evolution of asymmetric particle densities and analytically and numerically identify the freeze-out and freeze-in regimes. The resulting asymmetries depend primarily on the triplet branching fractions and are nearly independent of the triplet mass. The asymmetry scales as the square root of the smaller branching fraction when either decay mode into leptons or Higgs fields is suppressed, reflecting the symmetry structure; the lepton number is conserved once either coupling is turned off. We also show that the Majoron background does not induce CP-asymmetric decay of the scalar triplet in the massless final state limit, and thus its impact is negligible. Our results establish spontaneous leptogenesis as a simple and robust alternative to thermal type-II leptogenesis that requires neither additional scalar triplets nor explicit CP-violating interactions.
Read the full abstract
Computation and Language
Line-Coupled Language Model
Autoregressive language models generate one token per decoding step, limiting the useful output of each forward pass. Although diffusion models, insertion-based decoding, and multi-token prediction enable parallel generation, they either incur additional training-time token traffic or struggle to predict strongly dependent future tokens. We introduce the Line-Coupled Language Model (LCLM), an autoregressive model that advances multiple text lines together by predicting the next token for every active line while coupling the lines through shared causal context. LCLM interleaves line tokens into a single causal sequence and uses line-staggered rotary positions, retaining the standard next-token objective and causal attention. Controlled experiments show that cross-line targets are substantially less dependent than consecutive same-line targets, supporting lines as parallel generation units. With 881M parameters, LCLM produces an average of 2.94 content tokens per forward pass with a validation cross-entropy loss of 2.44, compared with 1.00 token per forward pass and a loss of 2.39 for the vanilla autoregressive baseline. Most notably, even when LCLM generates 16 tokens per forward pass, its loss is only 0.09 higher than that of the vanilla autoregressive baseline (2.34 vs. 2.25).
Read the full abstract
Artificial Intelligence
EEG-Driven Decoding Framework for Passenger Hazard Perception in Highly Automated Vehicles
Reliable risk assessment remains a central challenge for Autonomous Vehicles (AVs). Despite advances in automation, passenger cognition provides a non-intrusive auxiliary signal that improves both objective and perceived safety without requiring active human intervention. We introduce an Electroencephalogram (EEG)-based Brain-Computer Interface (BCI) that decodes passenger neural responses for both Risk Prediction (RP) and Danger Identification (DI), explicitly modeling humans as passengers to match real-world AV use. To achieve this, we propose the Passenger Cognitive Model (PCM), Risk-aware Sequential Labeling (RSL), and the Passenger EEG Decoding Strategy (PEDS), which integrates a 3D Convolutional Recurrent Neural Network (3D-CRNN) model for joint EEG decoding. Experimental results show that 3D-CRNN achieves a Balanced Accuracy (BA) of $95.3\% \pm 2.7\%$ in RP and improves single-subject DI from $80.9\% \pm 3.9\%$ to $85.0\% \pm 3.2\%$ with RSL. Event-wise analyses further show that 3D-CRNN consistently outperforms other models across different event types in RP and DI. In generalization experiments, 3D-CRNN achieves $77.0\% \pm 5.3\%$ BA in cross-session DI and $77.4\% \pm 1.1\%$ BA on seen subjects in cross-subject evaluation, while maintaining a $64.9\% \pm 8.5\%$ BA on unseen subjects, demonstrating promising generalizability and transferability across both intra-subject and inter-subject variability. These findings establish an Electroencephalogram (EEG) decoding framework for AV passenger hazard perception and suggest that passenger cognitive signals can provide auxiliary supervision for future AV decision-making and Safety of the Intended Functionality (SOTIF) support.
Read the full abstract
Quantum Physics
From Quantum Cryptography to Intuitionism and beyond: relativity, many-worlds and non-locality
After a recollection of some of my encounters with Gilles Brassard, on the occasion of his 70th birthday I present, first, a new idea on creation of new information at a limited rate. Next, I develop the idea that "indeterminacy is relative" [Entropy {23} 1326 (2021)] and how this illuminates the compatibility between relativity and non-locality.
Read the full abstract
Robotics
Beyond Task Success: Stage-Wise Reliability of World Model Planning under Sensing Degradation
In world model planning, sensing inputs pass through an encoder and predictor before affecting planner decisions, so final task success alone cannot reveal where sensing disturbances attenuate or persist in the pipeline. We apply 10 visual and temporal sensing degradations to a world model planner and track their effects across representation, future prediction, planner preference, and physical outcome using paired evaluation on the same 50 tasks. The relative impact of degradations was not preserved across stages: large representation shifts could attenuate downstream, while smaller initial shifts could persist to the outcome, and internal-response ordering did not directly match physical-outcome ordering. Temporal degradations also showed distinct patterns: even with similar overall changes in observation history, responses differed substantially with the location of corrupted information and the planner's actual exposure. This non-uniform stage-wise response was also observed in secondary evaluations with another manipulation task and a different world model. Stage-wise diagnosis can therefore identify where sensing disturbances attenuate or persist and help prioritize subsequent model verification and sensing mitigation.
Read the full abstract
Computer Vision
A Two-Stage Framework for Ego-Centric Key Object Identification via Object State Prediction
This paper presents a novel framework designed to enhance key object identification in autonomous driving. Existing methods primarily focus on either detecting objects independently or leveraging visual relationships, but they do not explicitly consider the ego vehicle's perspective in determining object importance. To address this gap, we propose a structured approach that integrates a virtual ego-vehicle representation and a modular object state predictor, enabling a more accurate estimation of object behaviors relative to the ego-vehicle. Subsequently, our framework employs spatial-temporal reasoning to refine key object identification, prioritizing objects based on their states and relative spatial information rather than relying solely on visual relationships. Experimental results on real-world driving datasets demonstrate the effectiveness of our approach in accurately detecting critical objects in complex traffic environments.
Read the full abstract
Differential Geometry
Finite-Time Singularities of the Kähler–Ricci Flow on a $\mathbb{C}P^m$-Bundle over a Product of Kähler–Einstein Manifolds
Symplectic Geometry
Legendrian Reidemeister moves for the convex surface projection
Computational Engineering
Centering Drives Normalization Gains: Price-Offset Nuisances in Cross-Sectional Return Prediction
Cross-sectional return prediction from raw intraday bars is sensitive to each instrument price level, an additive nuisance under a return-ranking hypothesis. We test whether removing this offset, rather than rescaling amplitudes or changing the encoder, explains gains on a point-in-time CSI~300 five-minute panel. We evaluate eight parameter-matched encoders with and without RevIN normalization; a parameter-free ladder then separates identity, scale-only, centering, last-value referencing, differencing, and standardization across all fields and restricted channels. Centering drives the reliable effect, while scale-only normalization does not help. All eight paired effects are positive and survive Holm correction on raw rank IC, after style residualization, and after further residualizing on short-term reversal. Among six stronger encoders, gains of 0.0376-0.0567 exceed the 0.0109 spread of normalized IC (0.0830-0.0939). Price-only standardization retains 93–101% of the all-field gain. These results place the main effect in transformed price-channel offset removal rather than amplitude scaling or encoder choice.
Read the full abstract
Probability
Uniform asymptotics and entropy structure of the Bingham distribution in arbitrary dimensions
The Bingham distribution is widely used to model directional data with antipodal symmetry, and its normalizing constant $Z$ and moments play a central role in statistical inference and closure models. Their behavior becomes singular when one or more eigenvalue gaps of the parameter matrix become unbounded. In this work, we develop a uniform asymptotic framework for the Bingham distribution in arbitrary dimensions. Starting from an inverse-Laplace integral representation, we derive asymptotic expansions with explicit remainder estimates that are uniform with respect to arbitrary relative scales among multiple diverging eigenvalues. In a particular regime, the asymptotic series becomes absolutely convergent with an exponentially small remainder. These results yield precise asymptotic profiles of the Bingham moments and characterize the degeneration of higher-dimensional distributions to lower-dimensional counterparts. As an application to the Bingham closure, we prove that the entropy admits a decomposition into an explicit logarithmic leading term and a residual that is uniformly Lipschitz continuous on the moment simplex. Moreover, on each boundary face, the residual agrees with its lower-dimensional version up to an explicit additive constant. The results provide a unified description of the singular structure of the Bingham distribution and have implications for numerical closure and $Q$-tensor models of nematic liquid crystals.
Read the full abstract
Computation and Language
PTCG: Persona-guided Tree-based Counterargument Generation
The ability to generate counterarguments is important for critical thinking and balanced discourse, yet existing approaches typically produce only a single counterargument, failing to capture the diversity and persuasiveness required in real-world debates. To address this limitation, we propose Persona-guided Tree-based Counterargument Generation (PTCG), a framework that combines Tree-of-Thoughts-inspired step-wise generation and pruning with speaker persona selection. By estimating the author's persona from the original argument and incorporating speaker personas representing distinct perspectives, PTCG operationalizes perspective-taking and enables the generation of diverse counterarguments. Results from LLM-as-a-Judge, classifier-based assessment, and human evaluations indicate that PTCG shows consistent improvements in both the diversity and persuasiveness of counterarguments compared to baseline methods.
Read the full abstract
High Energy Astrophysical Phenomena
LOFAR follow-up of sources in the First LHAASO Catalogue
The First LHAASO Catalogue lists 90 sources of very-high-energy gamma-ray emission, many without an established counterpart at other frequencies, reopening the question of which Galactic accelerators produce the highest-energy cosmic rays. We search for low-frequency radio counterparts to the 55 1LHAASO sources that fall within the Galactic fields of the LOFAR Two-metre Sky Survey Data Release Three (LoTSS-DR3), using 144~MHz continuum maps at $6''$ and $20''$ resolution, complemented with archival data from other radio surveys where available. We present an overview of the radio emission towards all 55 sources, and identify the structures potentially associated with the gamma-ray emission. We report 24 new supernova remnant (SNR) candidates, nine flat-spectrum radio shells with faint or no infrared counterpart, and six sources with no clear Galactic radio emission present in the LOFAR maps. A bootstrap analysis shows that 23 of the 50 sources in the Galactic plane footprint (46%, for $|b| \leq 5$) overlap a catalogued SNR, whereas random placement yields only $10.0\pm2.8$ matches even if the underlying Galactic SNR population is ten times larger than currently catalogued; in the outer Galaxy, where chance alignment is rarest, 4 of 12 sources coincide with an SNR against a null expectation of $\lesssim2$. We estimate that at least a quarter, and up to half, of the 1LHAASO Galactic-plane sources genuinely reside in SNR environments, consistent with a population dominated by pulsar wind nebulae, SNR–PWN composites, and cosmic-ray-illuminated clouds. Individual results include a new SNR candidate inside the TeV shell HESS~J1912+101, a radio candidate for the SNR hypothesised to explain the GeV emission of 1LHAASO~J1945+2424, and the first proposed counterparts of 1LHAASO~JJ0056+6343u and 1LHAASO~J2200+5643u.
Read the full abstract
Combinatorics
A note on near alternating sign matrices with prescribed row and column weights
We study near alternating sign matrices with prescribed row and column weights. After recalling the basic definitions and the necessary conditions coming from the Gale–Ryser theorem, we give a graph-theoretic characterization of the existence problem. More precisely, we show that a near alternating sign matrix with prescribed row and column weight sequences exists if and only if there exists a binary array with the same weights whose associated adjacency graph is bipartite. This reformulation shows that the Gale–Ryser conditions are not sufficient in general. We then formulate a natural conjecture up to permutations of the row and column weight sequences, and present some positive results, based on convex binary arrays and composition constructions.
Read the full abstract
Computation and Language
The Illusion of Debiasing: Persona Steering Redistributes Rather Than Reduces Bias in LLMs
Prompt-based interventions: system prompts, personas, role instructions, reliably reshape what a language model says, but it is unclear which layer they reach. Do they reconfigure internal structure, or only modulate the output channel? We use persona conditioning as a controlled probe, measuring its effects along a depth axis from self-report, through open-ended generation, to word-level parametric association, across three instruction-tuned models. We find a graded dissociation. Personas are legible but not structural: models follow single-trait instructions yet fail to reproduce human inter-trait covariance. The dissociation deepens with depth: personas hold or amplify closed-form QA bias, shift absolute tone while leaving between-group disparity unchanged, and barely perturb an already saturated associative baseline. Prompt-based steering thus operates in the output channel and has a structural reach limit that surface manipulability can mask.
Read the full abstract
Robotics
MAC-I$^2$: Learned Metrics-Aware Covariance for Robust Visual-Inertial Fusion in Initialization and Calibration
Visual-Inertial (VI) fusion is fundamental to accurate and robust state estimation, where camera and IMU measurements are combined according to their respective uncertainties. Existing methods, however, fuse the two modalities with predefined uncertainties, regardless of how reliable each is in the local context, and thus often struggle under challenging environments involving illumination changes, dynamic objects, and textureless regions. In this paper, we present MAC-I$^2$, which achieves robust VI fusion through learned metric-aware covariance for both modalities, so that vision and IMU compete on their own merits rather than relying on predefined uncertainties. Here, metrics-aware means that each predicted covariance faithfully reflects the actual magnitude of the corresponding measurement noise. On the visual side, we propagate learned feature-matching uncertainties into pose covariances for the fusion. On the inertial side, motivated by the observation that integration error accumulates sharply at the early stage and grows slowly afterward, we design a learned IMU model with a learnable initial covariance, and propose a dedicated fine-tuning strategy on a held-out training subset to enable the metrics-aware covariance on unseen sequences. As a showcase, we build a VI initialization and calibration system, since accurate and robust initialization and calibration are the prerequisite for any reliable VI system. Experiments on EuRoC, and VBR show that MAC-I$^2$ substantially outperforms existing methods: it achieves a 99.9% initialization success rate on EuRoC, reducing gravity and velocity errors by about 60% and 42% over the strongest baseline, and maintains 80% success rate on challenging VBR sequences where baseline methods such as VINS-Mono drop below 10%.
Read the full abstract
Networking
Ollama in the Wild: A Longitudinal Measurement of Exposed Ollama LLM Endpoints at Internet Scale
Self-hosted large language model (LLM) serving is emerging as a distinct category of Internet service, but we still know little about how these deployments appear and change on the public Internet. We present a 365-day longitudinal measurement of exposed Ollama endpoints (port 11434) from February 2025 to February 2026, combining daily active probing with GeoIP/ASN enrichment, PTR and port-443 host observations, and survival analysis. Across 362 observation days and approximately 4.8 million IP$\times$day observations, 26.4% of the 152,137 cumulative IPs appear for a single day; across five selected CVEs, only 0.43-2.90% of below-fix IPs upgraded in place; the top five countries/regions account for over 70% of weighted observations; and cloud and hosting providers dominate the top ASNs. These results characterize exposed Ollama as a structural exposure surface: persistent, growing, and heavily concentrated. At the same time, old versions, common model choices, cloud and hosting ASNs, PTR categories, and TLS certificate patterns remain visible across the year, indicating recurring insecure deployment practices in cloud infrastructure and the potential reach of provider-level mitigation.
Read the full abstract
Systems and Control
Observer-Based Robust Control for an Aerial Manipulator System under Unknown External Disturbances
This paper addresses the mathematical modeling and control of an aerial manipulator system comprising a quadrotor as the uncrewed aerial vehicle and a robotic arm as the manipulator. The dynamic model is established by identifying the overall center-of-mass velocity and the system's orientation as constraints, yielding a simplified, two-decoupled subsystems: a locked (overall translation) subsystem and a shape-space (actuation or overall rotation) subsystem, both subjected to external disturbances. Since the system is mechanically coupled, a critical challenge arises where constant bounded disturbances in the first subsystem manifest as time-varying, state-dependent disturbances in the second subsystem. Given that the quadrotor is inherently unstable, the movement of the robotic manipulator (RM) during flight can further jeopardize the stability of the entire QRM system if these disturbances and coupling effects are not effectively managed. To address this, we present a continuous nonlinear disturbance observer-based feedback control law, which enables the independent control of each subsystem while systematically eliminating cross-coupling effects. The efficacy of the proposed controller is validated through multiple simulations emulating practical operating conditions, thereby substantiating its real-world applicability and highlighting the core contributions of this work.
Read the full abstract
Rings and Algebras
A \(3\times 3\) counterexample to Lin and Wimmer's rank-minimization conjecture associated with Roth's similarity theorem
We give a \(3\times 3\) counterexample, valid over every field, to a rank-minimization conjecture of Lin and Wimmer (Bull. Aust. Math. Soc., 84 (3) (2011), 441–443) related to Roth's similarity theorem for the Sylvester matrix equation. We also prove that, over the complex field, no counterexample can occur when one of the two matrix sizes is less than \(3\). Hence the example is dimensionally minimal over the complex field.
Read the full abstract
Statistical Methodology
COINS: Any-Stage-Valid and Utility-Oriented Sequential Conformal Prediction
In clinical diagnosis, sequential screening, and cascaded prediction systems, information is often acquired in stages, with intermediate predictions guiding whether to act on the available evidence or defer action to gather further information. We study sequential predictive inference in these settings, aiming to construct a sequence of prediction sets that supports timely, resource-efficient decisions while remaining reliable throughout the acquisition process. Our objective is to maximize process-level utility subject to any-stage validity, which controls the probability of miscoverage anywhere in the prediction sequence. We develop COINS, a conformal framework for constructing such sequences, and establish a universality theorem identifying the common exclusion structure underlying all nested, calibration-symmetric, any-stage valid prediction sequences. Building on COINS, we develop Vopt-COINS for data-driven optimization of prediction sequences under task-specific utilities based on prediction-set size, stopping time and acquisition costs. Branchwise and localized extensions further support individualized acquisition and calibration. Numerical studies demonstrate the effectiveness and practical benefits of the proposed framework for reliable, resource-efficient sequential prediction.
Read the full abstract
Robotics
From LLM-Generated Specifications to Learned Quadruped Locomotion
Quadruped robot locomotion policies are often trained using reinforcement learning, which in turn relies heavily on hand-crafted reward functions. Designing reward functions requires substantial manual engineering, and it is often unclear which local rewards will induce the desired global behavior. Shaped rewards from formal specifications in languages like Signal Temporal Logic (STL) can make rewards more interpretable, but writing STL specifications itself still requires domain expertise. We study whether large language models (LLMs) can fill this gap by generating Parametric Signal Temporal Logic (PSTL) specifications that are subsequently used for policy learning. Given a natural language locomotion objective and a constrained specification grammar, GPT-5.5 and Qwen 3.6 independently propose STL templates for command tracking, safety, and gait structure. We instantiate the parameters of the generated PSTL templates using expert trajectories and retain only specifications that are consistent with demonstrated expert behavior. The resulting specifications are then transformed into smooth, finite-history reward functions and used to train a quadruped locomotion policy with Proximal Policy Optimization (PPO) in MuJoCo XLA (MJX). We evaluate both gait-aware and gait-agnostic settings. The former specifies walking-trot, trot, and bound regimes, while the latter allows contact patterns to emerge from the task objective. We compare against hand-engineered rewards, Text2Reward-style LLM-generated reward code, and an expert-switching oracle. Gait-aware Qwen 3.6 specifications achieved 100% survival and command success across all tested speeds (0.3–2.1 m/s) and matched the target gait at high speeds, whereas Text2Reward achieved 0% for both metrics at $\geq 1.9$ m/s. Videos: https://stl-locomotion.github.io/
Read the full abstract
Computer Vision
Discovering Natural Transformation Vulnerabilities in Black-Box Vision Models
Natural adversarial examples (NAEs) reveal that vision models can fail under realistic semantic changes beyond norm-bounded perturbations. However, generating NAEs in a black-box setting remains challenging because existing generative attacks often rely on surrogate models, learned attack priors, or costly query-based optimization, whereas the natural transformations that expose model vulnerabilities are unknown a priori. We propose Adversarial Scenario Attack (ASA), a query-based black-box framework that searches over natural-language editing scenarios using a multimodal language model and a modern text-guided generative editor. ASA jointly explores background, weather, and material/color transformations through winner–loser feedback, and uses a greedy explorer to compose only attack-improving scenarios. Across diverse ImageNet classifiers, ASA achieves substantially higher attack success rates than prior query-based generative attacks while requiring fewer victim-model queries and preserving competitive perceptual quality. Moreover, ASA exhibits both image-level and prompt-level transferability: its adversarial images remain effective across victim-model architectures, while its discovered editing scenarios can be reused across same-class images and, in some cases, across architectures. These findings suggest that vision models possess reusable vulnerabilities to natural transformation patterns, which ASA can efficiently identify in a black-box setting.
Read the full abstract
Robotics
Eventually Optimal and Scalable Multi-Agent Planning for Block Cave Mining
Automation in underground mining has the potential to significantly enhance safety, operational efficiency, and sustainability. However, effectively coordinating fleets of autonomous vehicles in dynamic mine environments introduces substantial challenges in both optimization and motion planning. To address these challenges, we introduce and formalize the Block Cave Mining (BCM) problem, which focuses on computing a transport plan that maximizes ore throughput while satisfying draw ratio constraints. To solve this problem, we propose SAMM, an eventually optimal anytime solver that jointly integrates task assignment, scheduling, and path planning via a mixed-integer linear programming formulation. To improve scalability, we also introduce SAMMS, a variant of SAMM that trades optimality guarantees for efficiency by decomposing the problem into shorter planning subcycles. Experimental evaluations using realistic industrial mine scenarios demonstrate that SAMMS achieves near-optimal throughput and scales effectively to larger fleets and mine layouts.
Read the full abstract
Machine Learning
Online Draft Co-Training for Speculative Decoding in Large-Scale, Long-Context RL Post-Training
Speculative decoding accelerates rollout generation, which dominates the cost of reinforcement learning (RL) post-training. Online co-training keeps the draft aligned with the evolving policy, further increasing acceptance length and speedup. However, training large models on long contexts requires context parallelism (CP) and pipeline parallelism (PP), and co-training drafts under this layout poses two obstacles: (1) the draft's branch attention is unsupported by causal CP, and (2) the target features that the draft consumes are spread across PP stages. In this paper, we present the design and implementation of online draft co-training in NeMo-RL. For CP, we extend packed, load-balanced zigzag ring attention by merging rank-local branch attention with causal main-sequence attention. For PP, TapChannel transports intermediate target features across stages via a separate path, leaving the pipeline schedule unaffected. Experiments show that co-training keeps the policy's reward and accuracy close to the no-draft baseline while delivering $1.48$–$2.25\times$ rollout speedups and $1.19$–$1.88\times$ end-to-end training speedups across model scales up to 122B. Our CP design achieves strong scaling at 256K sequence length with significant memory savings over prior work, and our PP transport incurs modest overhead. Code can be found at https://github.com/NVIDIA-NeMo/RL/issues/3698.
Read the full abstract
Artificial Intelligence
Beyond Sparse Rewards: A New Benchmark and Structure-Aware Graph Alignment for Micro-Drama Understanding
Micro-dramas, characterized by ultra-short durations and hyper-dense storylines, pose unique challenges for video understanding that conventional benchmarks fail to address. To bridge this gap, we introduce M-Drama, the first large-scale bilingual benchmark for micro-drama comprehension, featuring over 35K instances across 9,138 clips. Furthermore, while reinforcement learning can enhance VLMs on complex narratives, existing reward metrics often suffer from sparse and superficial signals, failing to capture intricate character identities and temporal structures. We propose SAGA (Structure-Aware Graph Alignment), a novel graph-matching reward function that models narratives as heterogeneous graphs. SAGA computes dense, rigorous rewards via decoupled semantic triplet and structural temporal matching. Extensive experiments on Qwen3-VL-8B-Instruct demonstrate that SAGA outperforms existing baselines, delivering substantial improvements in open-ended accuracy and summary quality, while maintaining competitive out-of-domain generalization. Code is available at https://github.com/qyx1121/MDrama_SAGA.
Read the full abstract
Materials Science
Coupled structural and electronic evolution under pressure in CuIr2Se4, CuRh2S4, and CuRh2Se4
Spinel chalcogenides provide a platform for investigating the interplay among metallic, superconducting, and pressure-induced insulating states. Here, we combine synchrotron powder X-ray diffraction and electrical-resistivity measurements to investigate the pressure evolution of CuIr2Se4, CuRh2S4, and CuRh2Se4 over pressure ranges extending beyond those previously explored. High-pressure diffraction reveals closely related monoclinic supercells in all three compounds. For CuIr2Se4 and CuRh2S4, constrained profile fits based on structural models relaxed using density functional theory are compatible with Phase-IV-type bond-disproportionated structures, whereas the data for CuRh2Se4 establish a compatible monoclinic unit cell without resolving its atomic-scale ordering pattern. Insulating-like transport develops abruptly over a narrow pressure range in CuIr2Se4 but more gradually over broader pressure ranges in the Rh-based compounds, in close correspondence with their respective structural transformations. We also establish previously unreported bulk superconductivity in CuIr2Se4 at ambient pressure: zero resistance is attained at 0.29 K, and the accompanying ac diamagnetic response is consistent with nearly complete superconducting shielding. These results establish a close relationship between the formation of the high-pressure monoclinic phases and the evolution toward insulating transport, and demonstrate that transition-metal and chalcogen substitutions tune the characteristic pressure scales and the competition with superconductivity within a closely related structural framework.
Read the full abstract
Combinatorics
Monotone Hadwiger Theorems on Spherical and Hyperbolic Convex Sets
For every $n\geq1$, we classify monotone rotation-invariant real-valued valuations on closed spherical convex sets, without assuming continuity or measurability. On proper sets, namely those contained in an open hemisphere, these are precisely the nonnegative linear combinations of the normalized spherical quermassintegrals. On all closed spherical convex sets, they are precisely the linear combinations of the spherical intrinsic volumes with nonnegative, nondecreasing coefficients. The representations are unique, and all such valuations are continuous and invariant under the full orthogonal group. In hyperbolic space, an isometry-invariant real-valued valuation on compact convex sets is continuous if and only if it is a linear combination of the Euler characteristic and the hyperbolic quermassintegrals. This representation is unique. Monotonicity is equivalent to nonnegative coefficients and implies continuity. If monotonicity is required only between nonempty sets, the Euler coefficient is unrestricted in the proper spherical and hyperbolic cases, whereas the classification on all closed spherical convex sets is unchanged. We also obtain corresponding classifications for valuations on closed convex cones that vanish at the zero cone and monotone classifications on compact projectively convex sets contained in an affine chart of real elliptic space.
Read the full abstract
Plasma Physics
Electromagnetic filament coalescence as magnetic island merging with diamagnetic effects
We investigate the nonlinear coalescence of two current-carrying ELM filaments using a three-dimensional electromagnetic fluid model. In the flat-density limit, the coalescence exhibits magnetic island-like reconnection, characterized by X-point formation, current-sheet development, and Sweet-Parker-like resistive scaling. Introducing a blob-like density perturbation modifies the reconnection dynamics: while the peak reconnection rate remains nearly unchanged for weak perturbations, it decreases and is increasingly delayed for larger density amplitudes. Analysis of the induction equation reveals a transition from resistive to increasingly density-dependent advective dynamics. Finite density perturbations also enhance the post-compression rebound, or sloshing, of the filaments. The sloshing amplitude increases with the density-gradient pressure force, establishing density perturbation as an additional control parameter for both reconnection and filament sloshing. These results highlight the coupled electromagnetic and pressure-driven dynamics governing the nonlinear evolution of ELM filaments in the tokamak edge.
Read the full abstract
Computation and Language
Revisiting Complete Reasoning Traces for Post-Training
Large language models (LLMs) are often post-trained on pre-collected reasoning trajectories to improve their reasoning capability. Such trajectories tend to be long due to complex, interwoven paths, which often include detours on the path toward the answer. However, it has been underexplored whether LLMs indeed benefit from learning complete trajectories in post-training, such as supervised fine-tuning (SFT). Starting from our pilot study, we find that full trajectories provide only limited benefit, while partial trajectories are effective even under heavy truncation. We analyze redundancy in reasoning trajectories through attention-based analyses and controlled token-removal studies, both of which show that intermediate tokens contribute minimally to final reasoning quality. This suggests that avoiding redundant information may allow LLMs to internally infer coherent alternatives by inferring missing steps from their internal knowledge, given known trajectory endpoints. Furthermore, we show that training LLMs using endpoints leads to consistent changes in reasoning behavior, and that it also benefits post-training methods based on reinforcement learning or on-policy distillation, highlighting the need to revisit complete reasoning traces. Code is available at https://github.com/naver-ai/revisiting-trace.
Read the full abstract
Instrumentation and Methods for Astrophysics
AstroSpecLM: A Spectrum-Language Model for Evidence-Grounded Astronomical Spectral Analysis
Astronomical spectra encode rich physical information, but drawing scientific conclusions from spectral features typically requires expert interpretation. This paper presents AstroSpecLM, a spectrum-language model that connects one-dimensional DESI spectra with Qwen3-4B to answer questions and provide explanations grounded in spectral evidence. Instead of generating question-answer pairs directly from templates or raw catalog fields, we first distill each spectrum into a compact set of catalog- and spectrum-derived facts, then use these facts as references to generate instruction-following conversations. The resulting model is competitive with specialist supervised baselines on classification and redshift estimation, while additionally producing natural-language explanations that reference specific spectral features. Our results indicate that grounding a language model in one-dimensional scientific spectra is feasible, and that fact-mediated instruction data yields a model capable of both prediction and explanation.