arXiv subjects

Sinan Wang

Publications and source records attributed to Sinan Wang.

At least 19 recordsLinked to original sources

arXiv2026

WebCQ: Cooperative Multi-Agent Deep Reinforcement Learning for Scalable Web GUI Testing

Multi-agent reinforcement learning (MARL)-based techniques have shown promise for GUI testing. However, as the complexity of modern GUI software increases, existing MARL-based approaches (e.g., MARG and Fastbot) struggle to scale due to the inherent limitations of their underlying tabular reinforcement learning algorithms. This limits their applicability to large-scale commercial GUI software, especially web applications with vast state spaces and many interactive elements. To fill this gap, we propose WebCQ, a novel MARL-based approach for scalable web GUI testing. WebCQ incorporates QTRAN for multi-agent coordination and a lightweight synchronization mechanism, allowing it to work under asynchronous web testing scenarios. It extracts semantic and exploration features for each UI event to form an action vector. This vector is concatenated with the current state vector and fed into the policy network, enabling DQN-based decision making within a dynamic action space. We evaluated WebCQ on eight large-scale commercial websites. Under the same time budget and agent count, WebCQ explored 33.3% more states and executed 42.2% more unique actions than MARG, while triggering more failures on six of the eight websites under test. It also demonstrated strong scalability, maintaining higher action throughput during 20-hour experiments, and achieving greater performance improvements as the number of agents increased. These results show that WebCQ overcomes key limitations of existing MARL-based approaches, providing a scalable and effective solution to enhance modern web GUI testing.

cs.SE↗

arXiv2026

From Cauchy to Gaussian: Sharp Lower Deviations and Ehrhard-Type Inequality

We establish dimension-free, pointwise optimal lower-deviation bounds for convex functions under spherically symmetric Cauchy-type measures. These bounds use either mean-variance normalization or moments about the median, the latter covering the Cauchy case where the mean and variance are unavailable. By approximation with a parametrized family the Cauchy-type measures, our results recover the Gaussian setting of Paouris and Valettas \cite{PaourisValettas18} and strengthen their variance-sensitive estimate to a sharp bound, with a complete equality characterization. This goes beyond the partial improvements obtained in subsequent work. We also prove an Ehrhard-type inequality for arbitrary convex sets under these Cauchy-type measures in every dimension and characterize all equality cases. This resolves a conjecture posed by Byczkowski, Małecki and Żak \cite{ByczkowskiMaleckiZak22}. The proof follows from a variational argument combined with the joint concavity of fiber rearrangement. The geometric inequality reduces distributions of convex observables to one-dimensional convex transforms, yielding the sharp lower-deviation bounds.

math.PR↗

arXiv2026

Rethinking the Roles of Diffusion and Decoding: Predictive-State Flows for Parallel Language Generation

In continuous diffusion language models (DLMs), latents generated by the diffusion prior are mapped almost deterministically to tokens, either by a decoder or through a token embedding matrix. Unlike in image generation, where visual details can vary continuously, the diffusion prior must capture fine-grained discrete token choices in continuous space, which can make the joint latent distribution difficult to model. We argue that generative roles should instead be divided: the diffusion prior should generate high-level predictive information, while the decoder should resolve the remaining token-level choices. Based on this view, we introduce Predictive-State Flows (PSF), which partitions a sequence into blocks and associates each block with a predictive state computed solely from the preceding context. A rectified-flow prior generates the sequence of predictive states, after which a block-causal decoder generates all blocks in parallel, autoregressively only within each block. This design reduces diffusion from L token-level states to K << L predictive states and reduces token-level sequential depth to L/K. On unconditional OpenWebText generation, a 197M-parameter model with K=8 achieves a generative perplexity of 22.8 at entropy 5.16 using 0.31 TFLOPs per 1024-token document, versus 24.0 and 7.2 TFLOPs for ELF-B with 32 sampling steps. The corresponding measured operating points yield MAUVE@512 of 0.93 and 0.73, respectively. Controlled comparisons further show that content states, which additionally encode their own blocks, yield 8-16% higher generative perplexity than predictive states, and that replacing sampled predictive-state trajectories with a shared mean lowers MAUVE@512 by 0.11. These results suggest that latent design for continuous DLMs should consider not only how compactly text is represented, but also how generative roles are divided between the prior and the decoder.

cs.CL↗

arXiv2026

LoCoSplat: Real-Time Feed-Forward 3D Gaussian Splatting with Minimal 3D Reasoning

Feed-forward 3D Gaussian Splatting (3DGS) increasingly aggregates multi-view evidence with heavy learned 3D networks. We propose LoCoSplat (Local-Context Splatting), motivated by the observation that a Gaussian is a local primitive: once depth is predicted, what the 3D stage must add (scale, rotation, opacity) depends on the point cloud around each anchor, and a fixed local average of that neighbourhood is enough to supply it, no heavy network required. LoCoSplat realises exactly this average: it splats a 16-d linear projection of the point features into a fine and a coarse grid and reads both back at each anchor with a 0.14M-parameter pointwise MLP; with no learned 3D network and no dynamic sparse computation, its whole encoder runs as one fp16 CUDA graph. On RealEstate10K, LoCoSplat outperforms every prior feed-forward method on PSNR, SSIM, and LPIPS at 6, 12, and 24 views, with a margin that widens as views densify (+3.3 PSNR over VolSplat, the prior voxel-aligned state of the art, at 24 views) and grows further under zero-shot transfer to ACID and fine-tuning on ScanNet. It reconstructs a 6-view scene in 33 ms on one NVIDIA RTX PRO 6000 GPU, the fastest of seven feed-forward methods and $4.2\times$ faster than the previous state of the art, trains $2.7\times$ faster ($5.9\times$ at 24 views), and uses $6.7\times$ less inference memory.

cs.CV↗

arXiv2026

A Brunn-Minkowski Theory for the Distribution of Random Pairs of Points in Convex Bodies

Let $K\subset\mathbb{R}^n$ be a convex body, let $X$ and $Y$ be independent uniform points of $K$, and set $R_K=(|K|/ω_n)^{1/n}$. We consider the distribution function \[ D_K(ρ)=\mathbb P\{|X-Y|/R_K<ρ\}, \] and its homogeneous counterpart \begin{equation*} \mathcal J_ρ(K)=\iint_{K\times K}\mathbf{1}_{\{|x-y|<ρR_K\}}dxdy. \end{equation*} Thus $D_K(ρ)$ is the proportion of ordered pairs of points of $K$ whose distance is less than $ρR_K$, or equivalently the radial distribution function of the normalized covariogram of $K$. The functional satisfies a Brunn--Minkowski inequality, with equality precisely for homothetic bodies, and its differential gives a first Minkowski inequality. We characterize the Borel measures that occur as Wulff differentials: they are exactly the nonzero finite positive measures with zero centroid which are not concentrated on a great subsphere, and the corresponding body is unique up to translation. We further prove fixed-volume and stationary rigidity of the ball in the subcritical range, together with regularity for the resulting equation. At saturation, the equation is the classical Minkowski equation.

math.AP↗

arXiv2026

A Simulation-Grounded Agentic VLM Framework for Wildfire Monitoring and Reporting

Effective wildfire monitoring requires relating visual evidence to physical fire dynamics, yet real videos with synchronized physical annotations are scarce and high-fidelity 3D simulation is costly. We present a simulation-grounded vision-language model (VLM) framework that automatically converts 2D wildfire simulations into labeled video episodes. A fixed Blender mapping produces low-detail 3D proxies aligned with simulator terrain, fuel layout, fire activity, and wind cues; controllable video generation supplies richer appearance. The proxies are intermediate representations rather than finely rendered final scenes. Generated videos and simulator labels form reusable multimodal memory for a training-free multi-agent VLM system that retrieves reference episodes, reconciles visual and memory-based predictions, and produces structured wildfire reports. On held-out generated episodes, video memory achieves 51.5% exact four-tag accuracy, compared with 22.6% for direct VLM querying and 16-17% for text-only memory; the complete system achieves 77.3% accuracy on six simulator-derived report fields. Component ablations, cross-generator tests, and three real-UAV evaluations assess retrieval, reporting, generator changes, and observable monitoring tasks. The framework connects automatic simulation-to-proxy conversion with memory-based VLM reasoning under scarce real-world physical annotations.

cs.CV↗

arXiv2026

GPUPhysBench: Benchmarking Coding Agents for Correct and Efficient GPU Physics Simulation

Writing fast GPU code for physical simulation is difficult: implementations must preserve numerical accuracy while handling irregular data access, synchronization, and iterative solvers. We introduce GPUPhysBench, a benchmark of 50 tasks testing whether coding agents can meet these demands. Tasks cover fluids, deformable solids, and granular materials, from individual simulation operators to complete simulators. Agents write, compile, test, and optimize GPU code with access to a NVIDIA GPU under fixed time budgets. We report pass rates and runtime performance relative to expert-optimized reference implementations. In a single-attempt evaluation of six frontier model-harness pairs, the two strongest pass all 50 tasks, but even the fastest reaches at least 0.9 the reference speed on only 22% of them, and no submission is more than 5% faster than the reference. The largest gaps arise in collision detection, constraint solving, and iterative solvers. GPUPhysBench brings physical simulation workloads to coding-agent evaluation, testing both the ability to implement numerical methods correctly and the ability to make them run efficiently.

cs.DC↗

arXiv2026

Neural Harmonic Measure Operator

We introduce Neural Harmonic Measure Operator (NHMO), a neural solver for elliptic PDE problems on variable-shape domains. The harmonic measure of a domain is the boundary probability distribution that, integrated against any boundary data, returns the Dirichlet Laplace solution. It depends only on the geometry, not on the boundary data. NHMO parameterizes the density of this measure as a transformer-based boundary kernel supervised by Walk-on-Spheres exit samples, so one trained kernel handles different boundary values on a shape with no retraining. We extend it to Poisson via a classical decomposition, with an auxiliary network amortizing the source-induced correction and avoiding the singular volume quadrature that breaks direct evaluation. At inference, new boundary values and new sources both yield PDE solutions by re-integration against the fitted kernel and lift, with no retraining. NHMO improves over four prior baselines on the MCB-B 3D variable-shape Poisson benchmark across all five categories, and is competitive with major neural-operator baselines on a controlled 2D testbed.

cs.LG↗

arXiv2026

A Variational Optimal Transport Operator on Incompressible Flow

We present the Variational Incompressible Optimal Transport (VIOT) operator, a generative neural operator for amortized incompressible density transport. Given a new source-target density pair, VIOT predicts a divergence-free velocity field and generates the full transport trajectory by feed-forward inference, replacing the hour-scale per-pair optimization used by adjoint fluid solvers and differentiable simulation baselines. The system consists of three components: a stream-function or vector-potential representation that enforces incompressibility by construction, a regularized incompressible transport objective that balances endpoint accuracy and flow smoothness, and a Fourier Neural Operator backbone that amortizes the solve across new pairs and grid resolutions. Together, these components make incompressible transport a reusable neural operator that facilitates various transport processes. Further, the generative capability extends beyond the training distribution, with VIOT producing incompressible transports for user-drawn source-target pairs in a real-time interactive system. We demonstrate VIOT on 2D and 3D density-transport benchmarks. Both 2D and 3D rollouts complete in seconds per pair, while per-instance baselines in our 2D comparisons optimize each new pair from scratch and require on the order of an hour, a roughly $10^4\times$ online speedup.

cs.LG↗

arXiv2026

Hamiltonian Two-Way Coupling of Nonlinear Waves and 3D Flows

Simulating large-scale free-surface water by coupling a localized 3D fluid solver to a cheaper 2D surface model has long faced a mismatch in wave dynamics: efficient 2D wave models used in graphics are typically either linear or non-dispersive. These models are fast, simple, and accurate for calm, small-amplitude seas, but coupling them with strongly nonlinear 3D solvers produces visible reflections and artifacts at the 2D--3D interface. We address this problem by introducing a nonlinear and dispersive 2D wave model based on the canonical Zakharov formulation. Its Hamiltonian structure, in which the surface elevation and surface potential form a canonical pair ($η$, $ψ$) governed by the wave energy, enables a canonically consistent two-way coupling scheme, allowing information to pass smoothly across the 2D--3D interface. Our 2D solver reduces mean wave-height error by 1.7--5$\times$ over SWE, BEM, and Airy baselines while running more than $10^3\times$ faster than BEM; it achieves greater nonlinear accuracy and coupling fidelity than SWE and Airy, with minor losses in speed and stability. Coupling it with a 3D Navier--Stokes solver yields a full system that suppresses visible seam artifacts across a range of experiments, including dispersion-matching and Kelvin-wake tests, and runs over 4$\times$ faster than a pure GPU NB-FLIP simulation on the same domain.

cs.GR↗

arXiv2026

On The Minkowski Problem With Respect To A Mixed Euclidean-Gaussian Density

In this paper, we pose a new class of Minkowski problems corresponding to the mixed Euclidean-Gaussian volume. Using the flow method, we prove the existence of smooth normalized solutions for the corresponding Monge-Ampère-type equation in the symmetric case. Then, by approximation, we obtain the existence of origin-symmetric solutions to the mixed Euclidean-Gaussian Minkowski problem.

math.AP↗

arXiv2026

$L_p$ Brunn-Minkowski inequality for weighted dual quermassintegrals

We investigate the $L_p$ Brunn-Minkowski inequality for dual quermassintegrals in weighted measure spaces, which is a special class of rotationally invariant measures proposed by Cordero-Erausquin and Rotem [Ann. Probab., {\bf 51} (2023)]. Specifically, the weighted dual quermassintegral is defined by integrating the radial density $|x|^{q-n}ϕ(|x|)$ for $q\in(0,n]$, where $ϕ$ is a positive radially non-increasing weight, it recovers the classical dual quermassintegral when $ϕ\equiv1$. For $p\geq1$, we prove the $L_p$ Brunn-Minkowski inequality with concavity exponent $p/q$ under the condition that $t\mapsto\logϕ(e^t)$ is concave, which is exactly the natural convexity condition from Cordero-Erausquin and Rotem's paper in general, improving the exponent $1/n$ when $p=1$. For $p\in(0,1)$, we obtain the result with exponent $p/q$ under more strictly weight assumptions, together with explicit lower bounds for the admissible range of $p$.

math.MG↗

arXiv2026

$L_p$ Minkowski problem and Brunn-Minkowski inequality for dual quermassintegrals

This paper studies the core problems in the $L_p$ dual Brunn-Minkowski theory, encompassing the $L_p$ Minkowski problem and $L_p$ Brunn-Minkowski inequality for dual quermassintegrals. For the case $0<p<q\leq n$, we establish $C^0$ estimates for the $L_p$ dual Minkowski problem without symmetric assumptions, thereby resolving a related problem proposed by Böröczky-Chen-Liu-Saroglou in the smooth sense. We further prove the uniqueness of smooth solutions under appropriate conditions, provided the density function is sufficiently close to a constant in the Hölder norm. Finally, exploiting the fact that the uniqueness of the Minkowski type problem is equivalent to the validity of the Brunn-Minkowski inequality in a certain sense, we study the $L_p$ Brunn-Minkowski inequality for dual quermassintegrals for origin-symmetric convex bodies with $p<q$.

math.AP↗

arXiv2026

Hermite-NGP: Gradient-Augmented Hash Encoding for Learning PDEs

We propose Hermite-NGP, a gradient-augmented multi-resolution hash encoding designed to enable fast and accurate computation of spatial derivatives for neural PDE solvers. Unlike existing NGP-based approaches that rely on automatic differentiation or finite differences and suffer from instability or high cost, Hermite-NGP explicitly stores function values and mixed partial derivatives at hash grid vertices, allowing fully analytic evaluation of gradients, Jacobians, and Hessians via Hermite interpolation. This design preserves the efficiency and spatial adaptivity of NGP while supporting analytic differential operators up to second order. We further introduce a multi-resolution curriculum training strategy analogous to multigrid V-cycles to enable coarse-to-fine optimization. Across a range of 2D and 3D PDE benchmarks, Hermite-NGP achieves up to approximately 20 times lower error than prior neural PDE methods, and reduces wall-clock convergence time by 2 to 10 times compared to other solvers, with per-epoch training times as low as 3.5 ms for models with up to 17M parameters.

cs.LG↗

arXiv2026

Generative Modeling with Orbit-Space Particle Flow Matching

We present Orbit-Space Geometric Probability Paths (OGPP), a particle-native flow-matching framework for generative modeling of particle systems. OGPP is motivated by two insights: (i) particles are defined up to permutation symmetries, so anonymous indexing inflates per-index target variance and yields curved, hard-to-learn flows; and (ii) particles live in physical space, so the flow terminal velocity has physical meaning and can encode geometric attributes, e.g., surface normals. OGPP instantiates three key components: (1) orbit-space canonicalization of the probability-path terminal endpoint, (2) particle index embeddings for role specialization, and (3) geometric probability paths with arc-length-aware terminal velocities that generate normals as a byproduct of the flow. We evaluate OGPP on minimal-surface benchmarks, where it reduces metric error by up to two orders of magnitude in a single inference step; on ShapeNet, where it matches the state of the art with 5x fewer steps and reaches airplane EMD comparable to DiT-3D with 26x fewer parameters and 5x fewer steps; and on single-shape encoding, where it produces normals and reconstructions competitive with 6D generators while operating entirely in 3D.

cs.GR↗

arXiv2026

Towards Understanding Android APIs: Official Lists, Vendor Customizations, and Real-World Usage

Android apps are built on APIs that abstract core Android system functionalities. These APIs are officially documented in multiple files distributed with the Android source code or SDK, which we collectively refer to as Android API Lists (AALs). Prior Android research has relied on specific AALs, often treating them as interchangeable ground truth. However, recent studies suggest that different AALs can lead to substantially different research outcomes, raising concerns about the validity and reproducibility of Android API-based analyses. To address this issue, we present the first in-depth empirical study of four official AALs that are widely used in prior work. We systematically characterize their contents and analyze their evolution across Android releases. We then perform a fine-grained comparison of the APIs recorded in each AAL to uncover their underlying API inclusion policies and inconsistencies. To assess the practical impact of these differences, we further examine API availability on nine Android devices, including both stock Android and vendor-customized systems. Finally, we analyze API usage in 17,759 real-world Android apps (including open-source apps, commercial apps, and malware) to quantify how the choice of AAL affects empirical Android research. Our results reveal that official AALs are neither stable nor mutually consistent, and that discrepancies among them can substantially influence research conclusions. We also observe that vendor-customized APIs are actively used by normal apps, yet remain largely overlooked by existing studies. Based on these findings, we discuss their implications for Android API-based research and provide actionable suggestions to help researchers select and interpret AALs more reliably.

cs.SE↗

arXiv2026

An Impulse-formed Navier-Stokes Solver based on Long-range Particle Flow Maps

We present a particle-grid characteristic-mapping framework that extends long-range characteristic mapping from inviscid flows to general Navier-Stokes dynamics with viscosity, body forces, and complex boundaries. Unlike traditional grid-based and vorticity-centered characteristic methods, our method is built on the observation that particle trajectories naturally provide the long-range flow map, enabling geometric quantities and their gradients to be transported in a direct and effective manner. We identify the impulse, the gauge variable of the velocity field, as the primary quantity mapped along characteristics while remaining compatible with standard velocity-based incompressible solvers. Using the 1-form representation of the impulse equation, we derive an integral formulation that decomposes the impulse evolution into a component transported geometrically along the particle flow map and a complementary component generated by viscosity and body forces evaluated through path integrals accumulated along particle trajectories. These components together yield a unified characteristic-mapping solver capable of handling incompressible Navier-Stokes flows with viscosity and body forces while maintaining the accuracy and geometric fidelity of characteristic transport.

cs.CE↗

arXiv2025

Are We on the Right Way to Assessing LLM-as-a-Judge?

LLM-as-a-Judge has been widely adopted as an evaluation method and served as supervised rewards in model training. However, existing benchmarks for LLM-as-a-Judge are mainly relying on human-annotated ground truth, which introduces human bias that undermines the assessment of reliability and imposes scalability constraints. To overcome these limitations, we introduce Sage, a novel evaluation suite that assesses the quality of LLM judges without necessitating any human annotation. Inspired by axioms of rational choice theory, Sage introduces two new lenses for measuring LLM-as-a-Judge: local self-consistency (pair-wise preference stability) and global logical consistency (transitivity across a full set of preferences). We curate a dataset of 650 questions by combining structured benchmark problems with real-world user queries. Our experiments demonstrate both the stability of our metrics and their high correlation with supervised benchmarks like LLMBar and RewardBench2, confirming Sage's reliability as an evaluation suite for the robustness and accuracy of LLM-as-a-Judge. Based on Sage, we reveal that current state-of-the-art LLMs exhibit significant reliability problems when acting as judges in both scoring and pairwise settings; even the top-performing models, Gemini-2.5-Pro and GPT-5, fail to maintain consistent preferences in nearly a quarter of difficult cases. We attribute this to a new phenomenon called situational preference, which explains why explicit rubrics or criteria can help the model judge consistently across answer pairs. Our further analysis shows that finetuned LLM-as-a-Judge is a feasible method to boost performance, and the panel-based judge as well as deep reasoning can enhance the judging consistency. We also find substantial inconsistency in human judgments, which indicates that human annotation may not be a reliable gold standard.

cs.CL↗