Searcharxiv⌕ Search

arXiv subjects

Dong Wang

Publications and source records attributed to Dong Wang.

At least 91 records · Page 5Linked to original sources
arXiv2026

MammoExpert: Benchmarking Chain-of-Thought Reasoning in Mammography Diagnosis

Mammography is an essential tool for breast cancer detection, with millions of examinations conducted annually. However, publicly available high-quality mammography datasets for AI development remain limited in both scale and annotation richness, particularly regarding pathological subtype coverage and structured diagnostic reasoning annotations. In this paper, we present MammoExpert, the first mammography dataset with Chain-of-Thought reasoning annotations across three diagnostic phases: (i) primal observation, (ii) factual assessment, and (iii) diagnostic synthesis. Comprising 2,379 mammography images covering 67 WHO-classified histopathology subtypes, each exam provides 42 radiographic features annotated by nine senior radiologists. We evaluate its performance on the breast lesion classification task, demonstrating superior accuracy and reasonability compared to existing classification models. Combining public dataset CBIS-DDSM with MammoExpert yields 7.1\% classification accuracy improvement, while the training model to learn CoT reasoning achieves another 4\% gain on the MammoExpert test set. Similar improvements are observed on INBreast and Vindr datasets, where the full approach yields accuracy gains of 6.9\% and 6.7\%, respectively. MammoExpert can serve as a benchmark for interpretable breast lesion diagnosis through explicit CoT reasoning.

cs.CV↗
arXiv2026

Universal photon blockade via two-photon light-matter interaction at chiral exceptional points

The photon blockade (PB) effect is a hallmark non-classical phenomenon in quantum optics and finds important applications for building quantum sources, while the control of PB by the non-Hermitian exceptional points remains largely unexplored. In this work, we theoretically investigate universal photon blockade in a microcavity harboring chiral exceptional points (CEPs) for building multiplexing quantum sources with nonreciprocal photon statistics. The results reveal that the presence of the CEPs leads to a stark contrast in the photon statistics of two whispering-gallery modes with opposite propagating directions. That is, one mode exhibits a strong PB effect while the other displays either sub-Poissonian or super-Poissonian distribution. Our findings thus may pave the way for advanced applications of photon blockade, and provide a theoretical foundation for the selective generation of single-photon and two-photon emission

quant-ph↗
arXiv2026

Harsher on Male? Evaluating LLMs on Gender-Asymmetric Moral Framing Across Diverse Conflict Scenarios

Existing studies on gender bias in LLMs have largely focused on stereotypes, occupational associations, or explicit harmful outputs. In this work, we ask whether LLMs apply consistent response standards to the same negative behavior under matched male-actor and female-actor conditions. We introduce GAMA-Bench, a gender-mirrored benchmark of 1,298 scenarios covering intimate relationship and public social conflicts. It constructs gender-neutral misconduct templates through controlled grids and cross-model review, then compiles them into paired first-person prompts with matched actor-gender and role-reference variations. We further design a structured response-framing protocol to measure how models allocate punishment, empathy, escalation, instruction, and blame. Experiments on 10 representative LLMs reveal a consistent male-disadvantaging asymmetry: male actors receive more punitive, escalatory, and blame-centered framing, whereas female actors receive more therapeutic and empathy-oriented framing for the same misconduct. Further analyses show that this pattern persists across model families, scenario tracks, model scale, and explicit thinking-style reasoning. The official code is available at https://github.com/xufeiqiong/GAMA-Bench.

cs.CL↗
arXiv2026

Unlocking the Potential of Movable Antennas: General and Practical Antenna Position Optimization

Recently, movable antenna (MA) has attracted wide attention in wireless communications due to its potential in enhancing wireless communication performance via local movement within a confined region. However, antenna position optimization (APO) has emerged as a major challenge for MAs, due to the lack of a tractable, analytical, and accurate channel model in terms of antenna positions. Although existing works have developed various algorithms for APO, most of them are based on simplified theoretical channel models, which limit their generality. To address this challenge, in this article, we present more general and effective APO algorithms for different purposes, categorized as continuous APO and discrete APO, respectively. Continuous APO is mainly applied for flexible array signal processing to boost large-scale communication performance, while discrete APO is applied for small-scale multi-path channel reshaping. Specifically, the discrete APO discretizes the antenna movement region into multiple sampling points and employs discrete algorithms to determine the optimal MA positions based on the point-wise channel state information (CSI), without the need for an analytical channel model. To reduce the overhead for CSI acquisition, we also present more efficient learning-based APO algorithms that operate without requiring full point-wise CSI. Finally, we compare the application scenarios of the proposed algorithms and validate their effectiveness with numerical results.

eess.SP↗
arXiv2026

Decoherence across phase-space scales: From compass states to general quantum states

Environmental decoherence occurs when a quantum system interacts with its surroundings, progressively reducing quantum interference and coherence, complicating the preservation of critical quantum features over time, especially during experimental implementation. The quantum features of a state can be represented in phase space via the Wigner function, which manifests across multiple scales, with decoherence potentially influencing each scale differently, as examined in this work. We consider the compass state and its photon-added and photon-subtracted variants (optimized compass states) as our representative examples, each of which exhibits phase-space features with dimensions beyond the Planck scale, making them suitable for quantum sensing applications. We investigate the interaction of these states with a heat reservoir by employing a range of well-established theoretical tools. We observe that compass states with finer-scale phase-space features are more fragile to decoherence, with parameters favoring greater sub-Planckness in phase space concomitantly increasing the fragility of these compass states to decoherence. Our findings are then validated for generic quantum states interacting with the heat reservoir, for which we provide analytical and numerical investigations, exploring the relationship between quantum state robustness to decoherence and the sizes of their phase-space features; that is, phase-space features at smaller scales decay faster under decoherence, and vice versa.

quant-ph↗
arXiv2026

Most Subradiant Bound Photon Pairs from Chirality-Mediated Dispersion Softening

We study the subradiant bound states (BSs) in a two-level atom array chirally coupled to a one-dimensional waveguide. We demonstrate that the chiral interaction can drive BSs to become the most subradiant two-excitation states across a wide spacing range. This phenomenon is rooted in a mechanism of chirality-mediated dispersion softening, where the BS band distortion suppresses the band curvature $|α_2|$ at an extremum point. We rigorously prove that the BS decay rate follows the scaling $Γ\sim |α_2|/N^3$, revealing that the reduction of $|α_2|$ is key to suppressing emission and enhancing subradiance. We also show the existence of chiral BSs in a realistic nanofiber interface.

quant-ph↗
arXiv2026

Genuine tripartite entanglement in Bhabha scattering with an entangled spectator particle

From the perspective of quantum information science, we investigate tree-level Bhabha scattering between an incident electron $A$ and a positron $B$, where $B$ is initially entangled with a spectator electron $C$, which does not participate in the scattering interaction. We find that the quantum electrodynamics (QED) scattering between $A$ and $B$ can drive the global $ABC$ system into a genuine tripartite entangled (GTE) state. Using four canonical tripartite entanglement metrics, we systematically characterize and quantify the GTE of the composite system, and demonstrate that the scattering momentum of the $A$-$B$ pair and the initial $B$-$C$ entanglement are the key resources governing GTE generation. We further analyze the monogamy of quantum correlations, which imposes fundamental constraints on the shareability of quantum resources in multipartite systems. Specifically, we systematically study the monogamy relations for the squared entanglement of formation and squared quantum discord in our scattering model, and find that monogamy constraints are markedly relaxed in the non-relativistic regime, enabling enhanced shareability of quantum correlations across the three particles. This work uncovers novel quantum correlation properties of fundamental QED scattering processes, and provides direct theoretical guidance for the development of QED-based quantum information processing protocols.

quant-ph↗
arXiv2026

Cross-Project Flakiness: A Case Study of the OpenStack Ecosystem

Automated regression testing is a cornerstone of modern software development, often contributing directly to code review and Continuous Integration (CI). Yet some tests suffer from flakiness, where their outcomes vary non-deterministically. Flakiness erodes developer trust in test results, wastes computational resources, and undermines CI reliability. While prior research has examined test flakiness within individual projects, its broader ecosystem-wide impact remains largely unexplored. In this paper, we present an empirical study of test flakiness in the OpenStack ecosystem, which focuses on (1) cross-project flakiness, where flaky tests impact multiple projects, and (2) inconsistent flakiness, where a test exhibits flakiness in some projects but remains stable in others. By analyzing 649 OpenStack projects, we identify 1,535 cross-project flaky tests and 1,105 inconsistently flaky tests. We find that cross-project flakiness affects 55% of OpenStack projects and significantly increases both review time and computational costs. Surprisingly, 70% of unit tests exhibit cross-project flakiness, challenging the assumption that unit tests are inherently insulated from issues that span modules like integration and system-level tests. Through qualitative analysis, we observe that race conditions in CI, inconsistent build configurations, and dependency mismatches are the primary causes of inconsistent flakiness. These findings underline the need for better coordination across complex ecosystems, standardized CI configurations, and improved test isolation strategies.

cs.SE↗
arXiv2026

An efficient and stable diffusion generated method for quadrilateral mesh generation in general domains

This paper introduces a novel, robust, and computationally efficient framework for high-quality quadrilateral mesh generation on general two-dimensional domains. The core of the proposed approach is a novel method for computing cross fields by minimizing a modified and relaxed Ginzburg--Landau-type energy functional. A key innovation is the extension of the problem from the original, potentially complex domain to a larger regular computational domain. This extension transforms the central computational procedure into an iterative scheme that requires only two straightforward and efficient operations: linear diffusion solved globally via the Fast Fourier Transform (FFT) and point-wise normalization. Notably, our method eliminates the conventional need for generating an intermediate triangular mesh or solving complex nonlinear optimization problems on the irregular domain. We provide a rigorous theoretical analysis, proving that the proposed iterative algorithm guarantees unconditional monotonic decay of the objective functional. Comprehensive numerical experiments demonstrate the method's robustness across a wide range of complex geometries, its significant computational efficiency afforded by the FFT-based diffusion, and its consistent generation of high-quality quadrilateral meshes. This work presents a reliable and theoretically sound alternative to existing mesh generation techniques, with strong potential for practical applications in scientific computing.

math.NA↗
arXiv2026

EgoProx: Evaluating MLLMs on Egocentric 3D Proximity Reasoning Across a Cognitive Hierarchy

Humans constantly reason about 3D proximity, the relations between their body and surrounding objects, to guide perception and action in daily life. Whether multimodal large language models (MLLMs) can perform such embodied 3D reasoning remains unclear. To this end, we introduce EgoProx, a benchmark for egocentric 3D proximity reasoning. We organize our tasks along a cognitive chain, covering intention, exploration, exploitation, and chain-of-actions reasoning. We also design an agent based data engine that produces diverse and consistent QA pairs at scale. We benchmark prevailing MLLMs on EgoProx and conduct additional analyses with dataset specific and task specific instruction tuning. We observe large cross-domain gains, indicating that current MLLMs contain some spatial knowledge; however, they still struggle to effectively leverage it for spatial reasoning VQA.

cs.CV↗
arXiv2026

"Refactoring Runaway": Understanding and Mitigating Tangled Refactorings in Coding Agents for Issue Resolution

Recent advances in coding agents have shown remarkable progress in software issue resolution. In practice, real-world issues are typically bug fixes or feature requests in which human developers naturally incorporate refactoring as part of the resolution process, resulting in tangled refactoring. Since LLMs are trained on large-scale open-source repositories, coding agents may inherit such behaviors. In this paper, we conduct an empirical study on Multi-SWE-bench, analyzing 3,691 valid patches generated by three agent frameworks with 12 LLMs. We find that coding agents introduce tangled refactorings less frequently (21.43% vs. 36.72%) and with lower intensity (0.66 vs. 1.75) than human developers, although they exhibit a broader diversity of refactoring types. Logistic regression analysis further shows that tangled refactorings are strongly associated with reduced compilability, while exhibiting no significant association with functional correctness. Based on these findings, we propose a refactoring-aware refinement approach that assesses the necessity and safety of tangled refactorings and selectively removes or repairs problematic operations. Our approach improves compilability from 19.34% to 38.33%, and additionally resolves 2.79% previously unresolved issues. Overall, this work presents the first step towards understanding tangled refactoring practices in agentic issue resolution and opens up avenues for future work.

cs.SE↗
arXiv2026

Partitioning for Intrinsic Model Inversion Resistance in Collaborative Inference

In collaborative inference (CI), transmitting intermediate representations $Z$ from edge devices enables model inversion attacks (MIA) that reconstruct the original inputs $X$, while existing defenses mainly perturb shallow-layer $Z$ at the cost of utility. We instead ask where an edge-cloud model should be partitioned to obtain intrinsic resistance to MIA. We challenge the intuition that depth is the driver of MIA resistance, and show that depth is sufficient only insofar as it enables a representational transition; this transition is necessary for intrinsic resistance and is marked by an abrupt rise in the lower bound of $H(X|Z)$. Correspondingly, the decisive variance term in the entropy bound shifts from a global variance to the intra-class mean-squared radius $R_c^2$ rather than dimensionality alone, yielding an $R_c^2$-based criterion to locate the transition zone, or identify it post hoc from MIA outcomes, which we term the Golden Partition Zone (GPZ). We further explain how $R_c^2$ evolves during training and show that it can be controlled through the label distribution; we refer to this controllable dynamic behavior as the Neural Vortex, an analysis-backed explanatory concept. Across four representative deep vision models, partitioning at the GPZ yields more than 4x higher reconstruction MSE compared to shallow splits; under entropy and inversion-model enhancements, decision-level representations provide 66 percent stronger resistance than feature-level ones, and we further observe that data type affects both the transition boundary and reconstruction.

cs.IT↗
arXiv2026

DisImpact: Quantifying the Physi-Social Impact of Natural Disasters Through Social Media

Natural disasters not only cause large-scale physical destruction, but also cascading social consequences that are difficult to quantify with traditional surveys and reports. Social media platforms offer an alternative perspective that captures multimodal, real-time, and user-generated content that can be leveraged for disaster impacts. In this paper, we introduce DisImpact, a two-stage framework that systematically quantifies the physi-social impacts of disasters via a Multimodal Large Language Model (MLLM). The social media posts are first classified into ten disaster impact categories that cover both physical and social domains. We then construct a disaster impact index that integrates the relative prominence of each category with the intensity of public engagement on a weekly basis. This design provides a unified scale for representing disaster impacts across both individual disaster impact categories and the broader physical and social domains. The unified representation enables direct comparison across categories and allows the impacts to be flexibly aggregated to reveal higher-level patterns and overall trends. We validate the impact indices against authoritative ground-truth data, including FEMA Public Assistance data and NASA FIRMS fire detections, observing consistent lead-lag correlations that demonstrate strong validity across both social and physical impact dimensions. We further conduct temporal and spatial analyses, and the results show that physical impacts are often peak during the disasters and localized in regions that are directly affected by disasters, while social impacts often emerge later and spread more broadly across time and space. To the best of our knowledge, this is the first framework to comprehensively quantify disaster impacts across their physical and social dimensions using multimodal data from multiple social media platforms.

cs.SI↗
arXiv2026

RELO: Reinforcement Learning to Localize for Visual Object Tracking

Conventional visual object trackers localize targets using handcrafted spatial priors, often in the form of heatmaps. Such priors provide only surrogate supervision and are poorly aligned with tracking optimization and evaluation metrics, such as intersection over union (IoU) and area under the success curve (AUC). Here, we introduce RELO, a REinforcement-learning-to-LOcalize method for visual object tracking that formulates target localization as a Markov decision process. Specifically, RELO replaces handcrafted spatial priors with a localization policy learned over spatial positions via reinforcement learning, with rewards combining frame-level IoU and sequence-level AUC. We additionally introduce layer-aligned temporal token propagation to improve semantic consistency across frames, with negligible computational overhead. Across multiple benchmarks, RELO achieves superior results, attaining 57.5% AUC on LaSOText without template updates. This confirms that reward-driven localization provides an effective alternative to prior-driven localization for visual object tracking.

cs.CV↗
arXiv2026

MambaPanoptic: A Vision Mamba-based Structured State Space Framework for Panoptic Segmentation

Panoptic segmentation requires the simultaneous recognition of countable thing instances and amorphous stuff regions, placing joint demands on long-range context modelling, multi-scale feature representation, and efficient dense prediction. Existing convolutional and transformer-based methods struggle to satisfy all three requirements concurrently: convolutional architectures are limited in their capacity to model long-range dependencies, while transformer-based methods incur quadratic computational cost that is prohibitive at high resolutions. In this paper, we propose MambaPanoptic, a fully Mamba-based panoptic segmentation framework that addresses these limitations through two principal contributions. First, we introduce MambaFPN, a top-down feature pyramid that leverages Mamba blocks to generate globally coherent, multi-scale feature representations with linear computational complexity. Second, we adopt a PanopticFCN-style kernel generator that produces unified thing and stuff kernels for proposal-free panoptic prediction, enhanced by a QuadMamba-based feature refinement module applied at multiple network stages. Experiments on the Cityscapes and COCO panoptic segmentation benchmarks demonstrate that MambaPanoptic consistently outperforms PanopticDeepLab and PanopticFCN under comparable model sizes, and matches or surpasses Mask2Former on Cityscapes in PQ and AP while requiring fewer parameters.

cs.CV↗
arXiv2026

Characterizing and Mitigating False-Positive Bug Reports in the Linux Kernel

False-positive bug reports represent a significant yet underexplored challenge in the development and maintenance of the Linux kernel. They occur when correct system behavior is mistakenly flagged as a defect, consuming developer effort without leading to actual code improvements. Such reports can mislead developers, waste debugging resources, and delay the resolution of real bugs. In this paper, we present the first comprehensive empirical study of false-positive bug reports in the Linux kernel. We manually construct a dataset of 2,006 bug reports comprising 1,509 genuine bugs and 497 false positives collected from Bugzilla and Syzkaller. Our analysis indicates that false positives demand effort comparable to real bugs, often requiring extended discussions and non-trivial closure time. They occur in several components, especially File Systems and Drivers, mainly due to external dependencies and semantic misunderstandings. To address this challenge, we evaluate large language models (LLMs) for automated false-positive bug report mitigation. Among various prompting strategies, retrieval-augmented generation (RAG) performs best, achieving 91% recall and an F1 score of 88%. These findings highlight the non-negligible cost of false positive bug reports and show the promise of LLMs for more efficient false positive mitigation in the Linux kernel.

cs.SE↗
arXiv2026

Global Energy Minimization for Simplex Mesh Optimization: A Radius Ratio Approach to Sliver Elimination

This paper constructs an energy function for simplex mesh based on the radius ratio and develops a corresponding mesh optimization method. The method combines vertex relocation and connectivity improvement, and can effectively remove slivers and improve the overall mesh quality. Based on the structure of the gradient of the energy function, we design a preconditioner, which reduces the number of iterations and improves the efficiency of the optimization algorithm. Numerical experiments show that the proposed method is effective in both sliver removal and mesh quality improvement.

math.NA↗
arXiv2026

Learning Decomposed Contextual Token Representations from Pretrained and Collaborative Signals for Generative Recommendation

Recent advances in generative recommenders adopt a two-stage paradigm: items are first tokenized into semantic IDs using a pretrained tokenizer, and then large language models (LLMs) are trained to generate the next item via sequence-to-sequence modeling. However, these two stages are optimized for different objectives: semantic reconstruction during tokenizer pretraining versus user interaction modeling during recommender training. This objective misalignment leads to two key limitations: (i) suboptimal static tokenization, where fixed token assignments fail to reflect diverse usage contexts; and (ii) discarded pretrained semantics, where pretrained knowledge - typically from language model embeddings - is overwritten during recommender training on user interactions. To address these limitations, we propose to learn $\underline{DE}$composed $\underline{CO}$ntextual Token $\underline{R}$epresentations (DECOR), a unified framework that preserves pretrained semantics while enhancing the adaptability of token embeddings. DECOR introduces contextualized token composition to refine token embeddings based on user interaction context, and decomposed embedding fusion that integrates pretrained codebook embeddings with newly learned collaborative embeddings. Experiments on three real-world datasets demonstrate that DECOR consistently outperforms state-of-the-art baselines in recommendation performance.

cs.IR↗