Papers

Low-loss stoichiometric silicon nitride on oxidized silicon has emerged as the major wide-band photonics platform for nonlinear and quantum photonics. These films, deposited by low-pressure chemical vapor deposition (LPCVD), are well known to have around 1000 MPa of residual tensile stress, which often leads to cracking when film thickness exceeds 400 nm. Although various stress-mitigation approaches can isolate or prevent cracking, this stress issue remains a major fabrication challenge. Low-residual-stress LPCVD stoichiometric silicon nitride on sapphire has also been reported and recently explored for photonics at telecommunication wavelengths. Here, we advance its development through systematic characterization, including thin-film stress analysis, ellipsometry and prism-coupling measurements of refractive index and loss, and swept-wavelength spectroscopy of fabricated photonic devices across a broad range of wavelengths. We demonstrate the direct growth of crack-free stoichiometric silicon nitride up to 2 {\mu}m thick with only 200 MPa of stress. Our fabricated devices show propagation losses of 0.35 dB/cm around 780 nm, 0.37 dB/cm around 1064 nm, and 0.25 dB/cm around 1630 nm. In addition, prism-coupling measurements indicate film losses below 1 dB/cm at wavelengths down to 500 nm. Our work highlights stoichiometric silicon nitride on sapphire as a promising wide-band photonics platform.

The integer and fractional quantum Hall effects reveal that quantized physical responses can take integer and fractional values, respectively. This naturally raises the question of whether quantization can extend beyond rational numbers to an irrational value. Here we show that such a quantized response can emerge from the limiting behavior of a finite, experimentally realizable geometric network. We develop a method based on infinite geometric structures, in which a sequence of finite networks rapidly converges toward an infinite limit. Following this approach, we identify a planar network whose normalized resistance converges to the irrational value $\sqrt{3}$, and realize a finite approximation with high-precision geometric control. The measured normalized resistance is $1.735 \pm 0.001\, h/e^2$ for a fully occupied Landau level in the designed pattern. Our experimental results extend the concept of quantization beyond rational values and demonstrate that an irrational quantized response can be approached through a finite physical system.

Learned real-time supersampling accumulates jittered low-resolution samples along motion vectors. Many methods use a temporal consistency loss to reduce flicker. We show that with accurate registration, this loss can suppress accumulation: on static content, the network copies its warped history at almost no temporal cost. Camera jitter in motion vectors raises this cost and can hide the effect. We identify four registration errors that training can conceal: jitter-carrying motion vectors, one motion field for both resolution lattices, jittered depth, and an unregistered skip path. A model-free oracle shows that correct registration makes freezing two orders of magnitude cheaper on static pixels. Controlled training shows that registration reverses the temporal weight's effect on PSNR. Under exact registration, increasing the weight reduces pixel variance at the expense of accuracy and perceived quality, with a 0.7 to 0.9 dB loss on static frames at weight 0.25. This reversal recurs across three network families and two scales; the loss also reduces accuracy on unjittered STSS data. With a small temporal weight and a normalised sample accumulator, a plain registered network improves the best published QRISP PSNR by 0.34 to 0.62 dB at $\times$2, $\times$3 and $\times$4 upscaling to 1080p. At $\times$2, its static pixel standard deviation is one quarter that of DLSS 2. Without retraining, it outperforms classical temporal upsampling by at least 4.4 dB on Arm's jittered Bistro capture. Our most accurate $\times$4 model runs in 2.03 ms per 1080p frame on an H200. Code: https://github.com/CV-AC/Frozen_history

We prove orbital stability of minimizing solitary waves at sufficiently small positive speeds for an Airy equation with a quadratic interaction filtered by a Bessel potential of order one. The stability hypotheses are verified: constrained minimization gives one negative direction, the small-speed KdV limit and operator-norm perturbation of the profile resolvent give a one-dimensional translation kernel, and a differentiable rescaled branch with uniformly bounded derivative gives a positive slope of the quadratic invariant. A conserved Lyapunov functional then proves orbital stability in the three-halves-derivative energy space and the perturbation also gives uniqueness of centered minimizers at small speed. The dynamical framework is supplied by a direct Bourgain-space bilinear estimate, local well-posedness in the square-integrable class, unconditional uniqueness from one-quarter derivative onward, and global well-posedness in the half-derivative energy space. For every prescribed positive speed, constrained minimization constructs smooth positive profiles that are symmetric, strictly monotone away from their center, and exponentially localized. At general positive speeds the stability criterion remains conditional on nondegeneracy and the slope condition.

Google Earth AI is a family of geospatial AI models, together with an agent that orchestrates them. We present this agent, Google Earth AI's Geospatial Reasoning agent (GRA), a Gemini-powered research system that combines geospatial data sources, foundation models, and tools to answer questions about our planet. GRA organizes these into modular domain skills and pairs them with upfront geospatial planning to decompose complex queries into multi-stage workflows. Evaluating such systems is hard: real-world problems admit multiple valid workflows, require data discovery across large catalogs, and rely on evolving ground truth. We introduce two benchmarks and release their development splits. GeoExpertQueries poses complex analytical problems grounded in practitioner use cases and refined by geospatial experts; each is paired with a consensus golden workflow from three independent experts and scored by a human-validated, tool-agnostic autorater. GeoKTools consists of templated queries with deterministic reference solvers that systematically vary the number of required tool types. We also use two external benchmarks, CBGB and UnivEARTH; the latter we find to be saturated. To isolate the contribution of GRA's tools and its agentic orchestration, we compare four configurations, each with Gemini 3.1 Pro and 3.8 Flash: a standalone LLM, the LLM with basic tools, the LLM with GRA's skills and tools but without its upfront planning and execution guidance, and GRA itself. With at most basic tools, an advanced model recovers much of the answer on templated queries but satisfies only a small fraction of the expert workflows on open-ended problems; GRA performs well on both. The largest step comes from GRA's skills and tools; on open-ended problems, its planning and execution guidance add a further gain with Gemini 3.8 Flash, letting Flash match or exceed Pro at lower cost.

Levitated dielectric sensors are often treated as perturbative probes of a pre-existing optical field. We show that this picture fails for a novel sensor geometry: a dielectric plate whose transverse extent captures the full optical mode. Even an optically thin sensor with modest refractive index can heavily restructure the intracavity fields, effectively partitioning a cavity into two subcavities with substantial cross-coupling and producing new optical resonances and trapping sites with properties which deviate from the standard intuition. Conventional perturbative expressions for the cavity frequency shift and optomechanical coupling depart from the exact solution at thicknesses as small as $\sim\lambda/(10n)$. Furthermore, cavity locking decreases the coupling and increases the trap stiffness by orders of magnitude, and the coupled subcavity structure permits appreciable optical cooling at zero detuning, a distinct behavior unique to this nonperturbative regime. These effects, sharing a common origin, establish a distinct regime of cavity optomechanics in which the sensor is no longer merely a probe of the cavity field, but an element that helps define it.

Strongly interacting fermionic chains obeying a $Z_n$ symmetry can host topological phases with highly degenerate ground-state manifolds. The different ground states are connected by strongly bound fractionalized edge modes called parafermionic zero modes. While exactly preserved under time evolution at zero temperature, interactions lead to a destabilization of these modes at finite energy densities, and, given a generic Hamiltonian, obtaining the lifetime of these modes is often impossible as it requires knowledge of the full spectrum of the Hamiltonian or long time evolutions. Here, we propose a novel quantum-inspired algorithm that can be used to variationally search for these almost strong zero modes in generic parafermionic chain models. We demonstrate analytically and numerically that our ansatz faithfully reproduces the exact results and finds strongly bound parafermionic operators in the non-integrable Kitaev-Hubbard chain and chiral Potts model. Our results further allow us to verify the existence of a true strong zero mode in the symmetrical limit of the chiral Potts model for $Z_3$-parafermions.

Trailer sway is a persistent contributor to severe crashes among light passenger vehicles towing trailers, yet it remains largely unaddressed by the autonomous vehicle systems now entering this new era of automation. Improper cargo placement rearward of the trailer axle reduces, and can invert, the tongue load transmitted through the hitch. The resulting loss of vertical load at the tow vehicle's rear axle diminishes the lateral tire force available there and degrades yaw stability of the coupled system, admitting a divergent lateral oscillation of the trailer known as trailer sway. Under adverse combinations of speed and loading, this oscillation can progress to jackknifing, in which the articulation angle diverges rapidly and directional control is lost. This work develops a linear state-space model of the yaw-plane dynamics of a light vehicle-trailer combination, parameterized to expose tongue load as an explicit stability-determining quantity. From this model a linear quadratic regulator is synthesized to suppress articulation-angle divergence. Because consumer trailers carry no sensors, the two trailer states the controller needs, trailer yaw rate and articulation angle, are estimated by a Kalman filter from measurements taken only on the tow vehicle. The controller is then implemented and evaluated in a MuJoCo multibody simulation across a range of longitudinal speeds and tongue-load conditions, including the negative-tongue-load regime. Results show a significant decrease in trailer oscillation when controlling a given vehicle's steering with an LQR using tow-vehicle IMU and LiDAR data. These findings indicate that LQR controlled anti-sway systems are a viable control method for future autonomous vehicles.

Spatial interference can break the inversion symmetry of an effective Josephson potential even in a laterally homogeneous junction with inversion-symmetric local energy. We derive a general criterion based on harmonic-dependent complex spatial form factors: a single local harmonic remains a shifted cosine, whereas multiple harmonics can acquire phases incompatible with any common inversion center, permitting cubic interactions and three-wave mixing. Phase reflection reverses all odd-order expansion coefficients about corresponding minima while preserving the even-order ones. For a capacitively shunted junction at fixed charging energy and arbitrary offset charge, it also preserves the complete quantum spectrum, enabling reversal of the odd nonlinear sector without spectral retuning. A centered cross junction with multiple local harmonics is one concrete example. Within the imposed-field model, a three-component magnetic field can generate an inversion-asymmetric effective potential, and reversing an odd number of field components exactly reflects it without changing the quantum spectrum of the capacitively shunted junction.

Multimodal 3D occupancy prediction underpins safe autonomous driving through comprehensive scene understanding. However, existing methods have largely pursued higher accuracy through sophisticated fusion, leaving resilience to missing sensors underexplored. In this paper, we investigate spiking neural networks (SNNs) as a means of building more resilient multimodal perception. Still, the discrete nature of spiking neurons makes optimization challenging. To overcome this difficulty, SpikeOcc is introduced with two complementary components, a Spiking Mixture-of-Experts (SMoE) and Sensor Availability Learning (SAL). Specifically, SMoE organizes RGB, LiDAR, and joint fusion into specialized pathways, enabling full-width unimodal representations while retaining the original multimodal features via a bounded residual. Beyond sensor dropout, SAL employs a staged curriculum with explicit routing supervision, progressively improving robustness under severe sensing conditions. Remarkably, SAL also yields notable gains for conventional models, highlighting its broader potential for resilient multimodal learning. Extensive experiments demonstrate that SpikeOcc maintains competitive full-sensor accuracy while markedly raising the performance floor under sensor outages. SpikeOcc illustrates the promise of neuromorphic computing, pointing toward more reliable and energy-efficient autonomous driving.

We consider symplectic analogues of the fractional Fourier transforms and obtain convenient representations in terms of Bargmann transforms and Wigner distributions. We show that they act as homeomorphisms on Schwartz spaces, the space of square integrable functions, the space of tempered distributions, and on Fourier-invariant Gelfand-Shilov spaces and their corresponding distribution spaces. More generally, we establish their homeomorphism properties on a broad class of spaces defined through (Wigner-)Hermite expansions. In addition, we prove that these transforms preserve norms on Schatten-von Neumann symbol classes associated with the Weyl calculus.

In molecules, electronic coherence typically decays within a few hundred femtoseconds, a timescale usually attributed to the environment. We show that it is set to a large extent by which coherences are excited: the molecule can host coherences that the bath barely sees, and optimal control can find and exploit them. Using gradient-based optimal control (GRAPE) with Redfield dynamics, we maximize the total vibronic coherence between the electronic ground and an optically bright electronic excited state in a three-state model with nonadiabatic features, in two regimes of low and high nonadiabatic mixing and for different system-bath coupling strengths. The Liouvillian spectrum reveals, in both regimes, a well-separated coherence mode with substantial ground-bright character that decays on a picosecond timescale. This mode is protected by mixing with the dark state, which brings the geometry of the optically accessible excited state close to that of the ground state and thus weakens its coupling to the bath. Without prior information, GRAPE finds this mode and keeps the photoexcited wave packet near the Franck-Condon region. At equal fluence, it increases the coherence at 800 fs by more than two orders of magnitude over a Gaussian pulse at the weakest coupling, and by many more at stronger coupling. The Liouvillian predicts the bath dependence of the long-lived coherence to within 1%, and a Gaussian pulse tuned to the mode frequency comes within 10% of the best pulse found. Beyond two-level and spin systems, our study puts forward the vibronic structure of molecules as a route to protected coherence subspaces, identifying a dark-state protection mechanism and a design criterion based on molecular parameters. These results lay the groundwork for our goal of using vibronic molecular systems as a resource for long-lived, controllable coherences in molecular quantum information science.

Scene classification of remote sensing imagery with multimodal large language models (MLLMs) is formulated as a single-forward-pass decision problem: candidate class names are lettered options in the prompt, a registered answer slot is conditioned on the prompt, and classification is read out by scoring letter tokens with the language-modeling head. We show that although this letter binding is highly effective within its training vocabulary (97.71% OA on UCM), it behaves erratically out of vocabulary: zero-adaptation accuracy across four cross-dataset directions fluctuates between 7.35% and 96.40%, and under out-of-vocabulary weights its confidence fails inversely on open-set evaluation, assigning higher confidence to unseen than to seen classes (AUROC 0.38), a failure we term silent hallucination. We retain the slot representation and replace the readout with RSAJEV, a training-free abstaining prototype memory that performs retrieval-based prototype matching: K reference images per class are cached as prototypes via their slot representations under the same prompt, and classification becomes nearest-neighbor retrieval over the prototypes, refined by a shrinkage-Mahalanobis geometric correction. This readout is mechanistically isomorphic to retrieval-augmented generation (RAG) but operates on discriminative decisions; it is the first retrieval-based readout for MLLM slot representations. The retrieval readout reaches 89.43% on the UCM cross-dataset gate with 16 references per class and zero gradient updates, and 96.90% on AID. Its ID-ness scores attain 0.91-0.99 AUROC on open-set evaluation over two datasets, exceeding symbolic confidence (0.38-0.96) in all four open-set settings, restoring a calibratable abstention channel and supporting class addition/removal at zero extra inference cost via offline rescoring of the cached matrix.

Measurements from the NASA Parker Solar Probe (PSP) mission demonstrate that inside $\sim20$ solar radii ($\Rs$) the magnitude of the interplanetary magnetic field does not fall off simply as $1/R^2$, with $\avg{|B|}\,R^2$ 15–20% below its value at $\sim60\,\Rs$. Here we use PSP magnetic field measurements over $10.6$–$56.8\,\Rs$ to show that the outward rise of $\avg{|B|}\,R^2$ is not due to the global structure of the heliospheric magnetic field or proximity to the heliospheric current sheet but is the direct result of the growth of large-amplitude, spherically polarized Alfvén waves (so-called `switchbacks'). Over the same distance range, $\avg{B_R}\,R^2$ stays constant, so that the sector-signed radial flux is conserved. The wave amplitude $\epsilon = B_{\perp,{\rm rms}}/\avg{B_\parallel}$ grows from $0.26$ to $0.57$ as $R^{1/2}\cos\alpha$, and $\avg{|B|}/\avg{B_R}$ rises by $20\%$. A simple model parametrized by the broadening of the distribution of deflection angles reproduces the effect to high accuracy. These results demonstrate that as the relative wave amplitudes grow with increasing solar distance, a dynamo-like process operates to enhance the field magnitude by up to $20\%$. By enhancing magnetic pressure this has implications for the large scale solar wind structure in the low-beta inner heliosphere, as well as partially resolving the long-standing discrepancy between remote measurements of the solar magnetic field and those made in situ by earlier missions farther from the Sun.

Launching a micro unmanned aerial vehicle (UAV) from a hovering carrier and recovering it in flight requires reliable confirmation that the child is securely seated before the carrier is allowed to descend. This is challenging because vision provides relative guidance but does not directly indicate load transfer, and a downward-facing camera can lose its fiducial near touchdown. This letter presents a mothership-child quadrotor system in which mission transitions are gated by complementary nonvisual sensing. An upward-facing single-beam LiDAR on the recovery deck provides a binary occupancy state using hysteresis and a debounce interval derived from the child's occluding footprint and crossing speed; real-time kinematic (RTK) positioning conditions the carrier hover; and carrier thrust provides a proprioceptive indication of load transfer. In ten outdoor missions using an F450-class mothership and a Q250-class child with camera-guided recovery, nine completed successfully while RTK-fixed positioning was maintained through release and capture. A complementary MuJoCo study of fully autonomous fiducial-based recovery achieved touchdown 0.077 m from the deck center, while marker detection was lost within the final approximately 0.2 m above the docked configuration. In simulation, carrier thrust changed by -17.0 N at release and +17.3 N at capture, agreeing within 6% with the child's 18.0 N hover thrust. The results support a sensing architecture in which vision guides the approach, while range and proprioceptive information corroborate the final docking decision.

A deterministic method for the problem of distributed area coverage path planning subject to the inclusion of a headland path is presented. Given a, in general, non-convex area to be covered that may, in general, include obstacle areas, multiple robots, at least one robot depot, and constraints on each machinery such as a working width, work capacity limits, travel capacity limits, nonholonomic dynamical constraints and, possibly, a robot start and end depot assignment, distributed paths for each machinery are calculated that collectively enable full area coverage. Variations are discussed, in particular, a scenario where unconstrained depots assignment can attain a global optimal solution to the problem. The importance of inclusion of a headland path for area coverage, as well as the benefits in comparison to an alternative Boustrophedon-pattern based approach are highlighted.

LLM optimization loops often show the proposer the candidates it has already tried and rejected, but leave it to the proposer to avoid them. We ask whether avoidance measured in a static probe predicts avoidance in the states an optimizer's own trajectory produces. In catalog-based loops where three open LLMs improve a GUI agent by naming one of 40 edits, a re-proposal is identified exactly by the edit's name. Clean base-state probes underpredict re-proposal in the proposers' own loops (macro-averaged loop-minus-probe rate 0.32 [0.27, 0.37]), and proposals of already evaluated harness-edit pairs consume 28-63% of valid proposer iterations. The two settings reject different edits: a probe's rejected edits are mostly ones the proposer did not choose, whereas a loop's are its own earlier choices, which remain preferred when the history is removed. Ordinary history lowers re-proposal for two proposers but raises it for the third, and the effects of failure evidence are model-specific. Stating the rejection status in one sentence lowers re-proposal in every model (-0.16 [-0.22, -0.09] relative to a neutral list of the same IDs). For catalog-based loops with exact edit identities, rejection avoidance should be measured on the rejected sets the loop produces.

A characteristic feature of topological field theories is the absence of local propagating degrees of freedom. In four-dimensional BF theory with cosmological term, we make this feature manifest by exactly reducing the finite non-Abelian shift symmetry. We do so by identifying a higher dressing field, a 1-form compensating the shift, allowing to build from first principles the composite shift-invariant field $\boldsymbol{B}=B-\lambda^{-1}F$. In terms of this dressed field, the dynamics becomes algebraic and sets $\boldsymbol{B}=0$, directly exposing the absence of local propagating degrees of freedom. This provides the first explicit instance of the higher-form Dressing Field Method.

It is known that there can be at most 3 disjoint nontrivial minimal graphs in R^3 with 0 boundary values, and conjectured that there can be at most 2. We explore properties that the graphs would have if there were 3.

This paper presents a geometry-aware predictive control framework for bipedal obstacle crossing across walking and running regimes. The proposed approach integrates a phase-structured trajectory generator with a constrained model predictive control (MPC) tracking layer to generate coordinated pelvis and swing-foot motions during obstacle traversal. The crossing process is represented through five interpretable phases: pre-crossing, leading-leg clearance, straddle, trailing-leg clearance, and recovery. Within this structure, swing-foot elevation is adapted as a function of obstacle height, while lateral stance widening and pelvis repositioning are introduced as functions of obstacle depth to improve transverse clearance and body alignment. A continuous gait-parameter evolution enables smooth walk-to-run transition without discrete controller switching. The generated references are tracked using MPC to regulate motion smoothness, clearance feasibility, touchdown behavior, and lateral admissibility. The framework is evaluated in reduced-order simulation over five progressively difficult obstacles under walking and running conditions. The robot successfully crosses the first four obstacles, achieving minimum body-obstacle clearances of 0.036, 0.036, 0.059, and 0.017 m, respectively. The fifth and most demanding obstacle leads to collision, identifying the current operating boundary of the proposed reduced-order formulation. Overall, the method provides a compact, interpretable, and computationally efficient basis for intelligent obstacle-aware bipedal locomotion.

We present ScenePlacer, a framework for scene-level pose estimation of objects generated by state-of-the-art native shape generation models, placing generated 3D objects into scenes given a partial geometric observation without modifying the underlying generative pipeline. Existing scene generation approaches either generate objects directly in scene coordinates, compromising their canonical-space priors, or regress poses from complete generated shapes in the shape generation process - the latter being a poorly conditioned problem that is sensitive to scale ambiguity and partial observations. ScenePlacer decouples object generation from pose estimation by reformulating alignment as conditional visibility-field learning. Given an observed partial point cloud from reconstruction and a generated mesh, it predicts a visibility field defined in object space, then estimates the rigid transformation by aligning the partial observation to visible parts on the mesh. We propose a pretrained 3D shape VAE as our backbone, preserving its mesh prior while conditioning it on a separately encoded observation. We further introduce visibility-aware, radius-constrained supervision and synthesize training pairs using only object-level meshes and rendered views, eliminating the need for scene-level pose annotations. On 3D-FUTURE, ScenePlacer consistently outperforms existing placement and pose-refinement baselines under both ground-truth and monocular depth estimation. With generated objects, it achieves a Chamfer distance of 0.0077 and an F-score of 92.40 under estimated depth and improves to 0.0023 and 98.27, respectively, under ground-truth geometry.

How charge rearranges within an electrode shapes the behavior of batteries, supercapacitors, and other electrochemical devices. Molecular simulations can capture this response by allowing the charge on each electrode atom to adjust to its surroundings, but repeatedly calculating these charges adds substantial computational cost. We present an open-source implementation of the constant-potential method in GROMACS that integrates these calculations into the program's optimized CPU and GPU routines. Comparisons with established codes validate the electrode charges, while charging, voltammetry, and impedance simulations demonstrate the response to changing electrical conditions. The implementation also supports electrodes that move or deform and models of finite metallic screening and electron spillover. In a pair of charged carbon nanotubes, allowing charge to move away from the facing surfaces reduces the repulsive force at contact by about half at an intermediate charge and keeps the pair bound where a uniform-charge model opens within the simulated time. These capabilities make the electrical response of electrodes accessible within GROMACS for studies of interfacial structure, charging dynamics, and electrode motion.

Pluralistic alignment asks that language models faithfully represent a society's diverse, competing values rather than collapse them into one dominant viewpoint, yet flagship pluralistic-alignment datasets remain heavily skewed toward Western respondents. We introduce NepPlural, a human-anchored benchmark for modelling competing socio-cultural personas in Nepali public discourse on youth migration and brain drain. We curate 2,148 code-mixed (Devanagari, Romanized Nepali, English) YouTube comments, annotate them along a four-axis persona taxonomy (Intent, Primary Driver, Value Orientation, Affect) plus an explicit No-Persona class, and build a 300-comment gold set labelled independently by three native Nepali speakers (mean Fleiss' $\kappa \approx 0.55$). Using this gold set to audit automatic annotation, we find that LLM annotation reliability is driven by protocol. Labels collected in batches through chat interfaces reach only near-chance inter-model agreement ($\kappa \approx 0.17$). Annotating one comment at a time raises agreement above the native-speaker level ($\kappa \approx 0.72$), but consistency is not validity: under a forced-choice schema, models agree on personas for the 39% of comments native speakers judge to express none, keeping macro-F1 against human gold near 0.50. Allowing abstention resolves this: models recover No-Persona comments with F1 of 0.79 to 0.87 and, against leave-one-annotator-out references, match or exceed individual native speakers (macro-F1 0.75 to 0.86 vs. 0.75). Encoders fine-tuned on the batched labels reach only 0.34. Finally, a human-judged steerability probe finds that three frontier models adopt single persona axes with 59 to 77% fidelity but compose full personas only 33% of the time, with no congruence gap detected at this scale between Western-congruent and Nepali-collectivist personas.

Let $F$ be a field of positive characteristic, and let $A$ be a finitely generated commutative associative $F$-algebra. We study Lie subalgebras of $Der \, A$ generated by locally finite derivations. We prove that if all commutators of the generators satisfy a natural local finiteness condition, then the associative subalgebra of $End_{F}\, A$ generated by these derivations is finite-dimensional.

Temporal graph neural networks (TGNNs) achieve superior performance in many dynamic graph learning tasks. These models use node state vectors (i.e., memories) to capture and record graph changes (i.e., events), and then exploit these memories to assist prediction. However, this memory-centric design makes multi-GPU training challenging because predictions for later events often depend on memory updates from earlier events. To maintain model quality, existing multi-GPU TGNN training schemes strictly preserve all event dependencies, which limits parallelism and causes expensive cross-GPU memory synchronization. In this work, we observe that many of these dependencies can be handled more flexibly without harming training quality. Specifically, (a) dependencies in TGNN training have high locality in both time and graph topology, (b) the short-range dependencies missed within a large batch can often be recovered cheaply, and (c) many memory updates are not immediately needed. Based on these observations, we propose DepBloom to accelerate multi-GPU TGNN training with three techniques: (1) temporally bounded spatial partitioning to reduce unnecessary cross-GPU synchronization and improve parallelism; (2) intra-batch memory refinement to recover missing short-range temporal dependencies inside enlarged batches; and (3) critical-path optimization to prioritize urgent memory updates while deferring non-critical work during multi-GPU pipeline execution. Experimental results across representative models and diverse datasets show that DepBloom achieves an average speedup of 2.84$\times$ and up to 5.16$\times$ over state-of-the-art methods, while maintaining comparable or better model accuracy.

When we see a null pointer check, we know that the developer believes that the pointer may not be initialized. If the same pointer is later used without a null pointer check, we may consider this a violation of the developer's belief and ultimately a bug. We develop such a belief-guided program analysis for vulnerability detection (VD). For example, when an authentication function returns the same error for both, a wrong user name and a wrong password, an auditor may infer that the developer believes that both cases are indistinguishable, mitigating a user enumeration attack. This belief is reasonable, as it follows standard security practice. Yet, a function call inducing a difference in response time for both cases, contradicts this belief. This contradiction uncovers a timing side-channel that syntactic analysis tools would miss. Existing static analysis tools struggle with this type of flaw because they require them to be expressible by general, syntactic patterns. Our belief-guided approach detects such domain-specific, high-level semantic security flaws that remain elusive. Given a function, our approach, Creed, infers the developer's security-related beliefs, evaluates how reasonable each one is, and derives the conditions under which it is violated. In contrast to existing ML-based VD techniques, Creed explains the root cause, the symptoms, and the conditions under which each issue is exploitable. To address the context-dependency problem in the evaluation, we introduce two measures for explanation-based VD: recall quantifies how effectively an approach correctly explains a known vulnerability, and overhead quantifies the effort of inspecting its reports. On a recent real-world vulnerability benchmark, Creed detects both semantic and syntactic vulnerabilities more effectively than existing explanation-based VD tools, achieving the best effectiveness-effort trade-off.

Visible watermarks convey document ownership, confidentiality, and status, but can also carry adversarial signals. We introduce Auto-Regional Adversarial Watermarking (AR-AWM), a model-aware white-box attack on end-to-end visual language models (VLMs) for document understanding. AR-AWM constructs a small, spatially distributed watermark-text support and optimizes only its pixels to disrupt visual representations. The attack is untargeted and question-agnostic: generation uses no questions, answers, parsing prompts, language-model gradients, or language-model outputs. Each attacked page is reused across downstream queries. On Qwen2.5-VL-7B, a watermark support covering 7.98% of page pixels under the reporting threshold reduces DocVQA exact match from 0.899 to 0.150 and average normalized Levenshtein similarity (ANLS) from 0.939 to 0.212. Selected qualitative cases retain visually legible source fields. The same support without optimization and with matched random perturbations retains ANLS of 0.932 and 0.933, respectively, identifying adversarial optimization as the primary cause of degradation. Layout-matched optimized rectangles yield ANLS of 0.276, showing an incremental benefit from text-shaped geometry. AR-AWM also degrades all reported text, formula, reading-order, and table metrics on OmniDocBench. Separate white-box optimization on InternVL2.5-8B reduces ANLS from 0.922 to 0.046, while weak cross-model transfer reveals strong encoder specificity. These results identify conventional watermark-shaped supports as an adversarial surface for document VLMs in the evaluated white-box setting.luated white-box setting.

The coexistence of multiple classes of localized states, including different types of topological corner states, is highly desired; yet its realization typically requires sufficiently strong long-range couplings, bringing challenges to experimental realizations across a wide range of systems. Here, we demonstrate the emergence of multiple localized corner modes evolution under purely nearest-neighbor coupling conditions without long-range couplings, not even next-nearest-neighbour couplings, and experimentally confirm their coexistence in a single lattice. Our platform simultaneously hosts type-I and type-II corner states, corner localized bound states in the continuum, flat-band states, and edge states. We experimentally realize such lattice using topolectrical circuits, a system that can strictly exclude long-range coupling by design, enabling direct characterizations of both the energy-band spectrum and the spatial profiles of the localized eigenmode. This work demonstrates that solely modifying nearest-neighbor couplings can replace the role of long-range interactions in realizing multiple types of corner states, providing a deterministic and widely accessible route to multiple corner-state classes and rich localized states, with immediate potential for mode-division multiplexing and multimode topological devices.

The anomalous Hall effect (AHE) in magnetic Weyl semimetals arises from distinct microscopic origins-intrinsic Berry curvature and extrinsic scattering mechanisms-yet reversible switching between these regimes within a single material platform remains elusive. Here, we demonstrate that off-stoichiometric tuning of the Al/Si ratio (from ~1.05 to ~0.75) reversibly drives an intrinsic-extrinsic-intrinsic crossover of the AHE in PrAlSi single crystals. Across four compositions spanning a narrow Si variation (PrAl0.93Si0.89 to PrAl0.79Si1.06), the anomalous Hall conductivity (AHC) exhibits a non monotonic arch (170-1413- 507-164 {\Omega}-1cm-1). The intermediate composition PrAl0.90Si0.94 (Al/Si ~0.96) exhibits high carrier mobility, multi carrier transport, yielding a giant AHC of 1413 {\Omega}-1cm-1 dominated by extrinsic skew-scattering. The Si-poor and Si-rich compositions exhibit lower AHC values and intrinsic Berry-curvature-dominated AHE, establishing a composition-driven intrinsic-extrinsic-intrinsic crossover. Deviating from the intermediate composition suppresses mobility and reverses carrier type from hole-like to electron-like. The Al/Si ratio thus serves as a chemical knob that primarily modulates the Fermi level via compositional band filling, while off-stoichiometry-induced chemical disorder plays a subordinate role in determining the AHE mechanism. These results position off-stoichiometric PrAlSi as a tunable platform where subtle compositional variation reversibly switches the AHE mechanism.

Agent skills encode operational procedures and domain expertise as reusable knowledge assets, yet their distribution in plaintext exposes them to unauthorized copying and leakage. However, practical mechanisms for skill ownership verification and leakage tracing remain underdeveloped. Watermark embedding before distribution offers a potential solution, but structured, compact skill instructions impose stringent requirements for factual and procedural accuracy. To address these challenges, we introduce SkillMarker, a KGW-based watermarking method that jointly targets functionality preservation and strong watermark embedding. SkillMarker unifies quality-constrained rewriting with signal-oriented candidate search. The former separates essential knowledge from flexible expression to preserve instruction fidelity; the latter explores alternative expressions to obtain a stronger watermark signal under preservation constraints. We release an open-source toolkit that integrates watermark embedding, quality auditing, and detection to support the protection of agent skills. SkillMarker targets ownership verification and leakage tracing in distributed skill ecosystems, where content may be copied in whole or in part, or incorporated into other skills. Across four datasets, SkillMarker negligibly affects downstream task performance on average. In our copy-paste tests, window-based detection consistently identifies copied passages containing as few as 1,536 consecutive tokens. Even after merging with one or three background skills, SkillMarker retains watermark detection rates of 72.2–77.8%.