arXiv subjects

Ziqi Wang

Publications and source records attributed to Ziqi Wang.

At least 73 records · Page 4Linked to original sources

arXiv2026

Graph2Eval: Automatic Multimodal Task Generation for Agents via Knowledge Graphs

As multimodal LLM-driven agents advance in autonomy and generalization, traditional static datasets face inherent scalability limitations and are insufficient for fully assessing their capabilities in increasingly complex and diverse tasks. Existing studies have attempted to generate agent tasks using LLMs, but due to the inherent hallucinations of LLMs and the lack of internal data relationship modeling, these tasks often exhibit semantic inconsistencies and solvability issues. To address these challenges, we introduce Graph2Eval, a knowledge-graph-driven framework for automated, scalable, and semantically grounded agent task generation. At its core, Graph2Eval leverages a knowledge graph built from heterogeneous external data sources as a structured task space, generating multimodal agent tasks through subgraph sampling and task construction guided by task templates and meta-path strategies. To further ensure task reliability, a multi-stage filtering pipeline based on node reachability analysis, LLM scoring, and similarity analysis ensures the diversity and solvability of the generated tasks. By unifying both RAG Agent and Web Agent scenarios, Graph2Eval enables efficient generation of multimodal document understanding tasks and multi-step web interaction tasks. We instantiate the framework with Graph2Eval-Bench, a curated dataset of 1,319 tasks spanning document understanding and web interaction scenarios. Extensive experiments show that, on average, Graph2Eval improves task semantic consistency by 20% and solvability by 17% over baselines, while Graph2Eval-Bench effectively distinguishes agent performance, offering a new perspective on agent evaluation.

cs.CL↗

arXiv2026

Kelix Technical Report

Autoregressive large language models (LLMs) scale well by expressing diverse tasks as sequences of discrete natural-language tokens and training with next-token prediction, which unifies comprehension and generation under self-supervision. Extending this paradigm to multimodal data requires a shared, discrete representation across modalities. However, most vision-language models (VLMs) still rely on a hybrid interface: discrete text tokens paired with continuous Vision Transformer (ViT) features. Because supervision is largely text-driven, these models are often biased toward understanding and cannot fully leverage large-scale self-supervised learning on non-text data. Recent work has explored discrete visual tokenization to enable fully autoregressive multimodal modeling, showing promising progress toward unified understanding and generation. Yet existing discrete vision tokens frequently lose information due to limited code capacity, resulting in noticeably weaker understanding than continuous-feature VLMs. We present Kelix, a fully discrete autoregressive unified model that closes the understanding gap between discrete and continuous visual representations.

cs.CV↗

arXiv2026

Adaptive Dual-Weighting Framework for Federated Learning via Out-of-Distribution Detection

Federated Learning (FL) enables collaborative model training across large-scale distributed service nodes while preserving data privacy, making it a cornerstone of intelligent service systems in edge-cloud environments. However, in real-world service-oriented deployments, data generated by heterogeneous users, devices, and application scenarios are inherently non-IID. This severe data heterogeneity critically undermines the convergence stability, generalization ability, and ultimately the quality of service delivered by the global model. To address this challenge, we propose FLood, a novel FL framework inspired by out-of-distribution (OOD) detection. FLood dynamically counteracts the adverse effects of heterogeneity through a dual-weighting mechanism that jointly governs local training and global aggregation. At the client level, it adaptively reweights the supervised loss by upweighting pseudo-OOD samples, thereby encouraging more robust learning from distributionally misaligned or challenging data. At the server level, it refines model aggregation by weighting client contributions according to their OOD confidence scores, prioritizing updates from clients with higher in-distribution consistency and enhancing the global model's robustness and convergence stability. Extensive experiments across multiple benchmarks under diverse non-IID settings demonstrate that FLood consistently outperforms state-of-the-art FL methods in both accuracy and generalization. Furthermore, FLood functions as an orthogonal plug-in module: it seamlessly integrates with existing FL algorithms to boost their performance under heterogeneity without modifying their core optimization logic. These properties make FLood a practical and scalable solution for deploying reliable intelligent services in real-world federated environments.

cs.LG↗

arXiv2026

WinFLoRA: Incentivizing Client-Adaptive Aggregation in Federated LoRA under Privacy Heterogeneity

Large Language Models (LLMs) increasingly underpin intelligent web applications, from chatbots to search and recommendation, where efficient specialization is essential. Low-Rank Adaptation (LoRA) enables such adaptation with minimal overhead, while federated LoRA allows web service providers to fine-tune shared models without data sharing. However, in privacy-sensitive deployments, clients inject varying levels of differential privacy (DP) noise, creating privacy heterogeneity that misaligns individual incentives and global performance. In this paper, we propose WinFLoRA, a privacy-heterogeneous federated LoRA that utilizes aggregation weights as incentives with noise awareness. Specifically, the noises from clients are estimated based on the uploaded LoRA adapters. A larger weight indicates greater influence on the global model and better downstream task performance, rewarding lower-noise contributions. By up-weighting low-noise updates, WinFLoRA improves global accuracy while accommodating clients' heterogeneous privacy requirements. Consequently, WinFLoRA aligns heterogeneous client utility in terms of privacy and downstream performance with global model objectives without third-party involvement. Extensive evaluations demonstrate that across multiple LLMs and datasets, WinFLoRA achieves up to 52.58% higher global accuracy and up to 2.56x client utility than state-of-the-art benchmarks. Source code is publicly available at https://github.com/koums24/WinFLoRA.git.

cs.LG↗

arXiv2026

Intermittency suppression in turbulence via forced light particles

We investigate how turbulence is reshaped by the presence of externally forced light particles, using high-resolution direct numerical simulations with four-way coupling. The particles are subject to an oscillatory force that in turn locally affects the fluid flow through momentum exchange at the position of the particles. Since the light particles preferentially concentrate in high vorticity regions, this leads to an intricate preferential turbulence modulation. We show that through this modulation, the forced light particles strongly reduce the intermittency of the flow, shedding new light on the delicate relationship between vortex filaments and turbulence intermittency.

physics.flu-dyn↗

arXiv2026

Idea2Story: An Automated Pipeline for Transforming Research Concepts into Complete Scientific Narratives

Autonomous scientific discovery with large language model (LLM)-based agents has recently made substantial progress, demonstrating the ability to automate end-to-end research workflows. However, existing systems largely rely on runtime-centric execution paradigms, repeatedly reading, summarizing, and reasoning over large volumes of scientific literature online. This on-the-spot computation strategy incurs high computational cost, suffers from context window limitations, and often leads to brittle reasoning and hallucination. We propose Idea2Story, a pre-computation-driven framework for autonomous scientific discovery that shifts literature understanding from online reasoning to offline knowledge construction. Idea2Story continuously collects peer-reviewed papers together with their review feedback, extracts core methodological units, composes reusable research patterns, and organizes them into a structured methodological knowledge graph. At runtime, underspecified user research intents are aligned to established research paradigms, enabling efficient retrieval and reuse of high-quality research patterns instead of open-ended generation and trial-and-error. By grounding research planning and execution in a pre-built knowledge graph, Idea2Story alleviates the context window bottleneck of LLMs and substantially reduces repeated runtime reasoning over literature. We conduct qualitative analyses and preliminary empirical studies demonstrating that Idea2Story can generate coherent, methodologically grounded, and novel research patterns, and can produce several high-quality research demonstrations in an end-to-end setting. These results suggest that offline knowledge construction provides a practical and scalable foundation for reliable autonomous scientific discovery.

cs.CE↗

arXiv2025

TrackTeller: Temporal Multimodal 3D Grounding for Behavior-Dependent Object References

Understanding natural-language references to objects in dynamic 3D driving scenes is essential for interactive autonomous systems. In practice, many referring expressions describe targets through recent motion or short-term interactions, which cannot be resolved from static appearance or geometry alone. We study temporal language-based 3D grounding, where the objective is to identify the referred object in the current frame by leveraging multi-frame observations. We propose TrackTeller, a temporal multimodal grounding framework that integrates LiDAR-image fusion, language-conditioned decoding, and temporal reasoning in a unified architecture. TrackTeller constructs a shared UniScene representation aligned with textual semantics, generates language-aware 3D proposals, and refines grounding decisions using motion history and short-term dynamics. Experiments on the NuPrompt benchmark demonstrate that TrackTeller consistently improves language-grounded tracking performance, outperforming strong baselines with a 70% relative improvement in Average Multi-Object Tracking Accuracy and a 3.15-3.4 times reduction in False Alarm Frequency.

cs.CV↗

arXiv2025

AIAuditTrack: A Framework for AI Security system

The rapid expansion of AI-driven applications powered by large language models has led to a surge in AI interaction data, raising urgent challenges in security, accountability, and risk traceability. This paper presents AiAuditTrack (AAT), a blockchain-based framework for AI usage traffic recording and governance. AAT leverages decentralized identity (DID) and verifiable credentials (VC) to establish trusted and identifiable AI entities, and records inter-entity interaction trajectories on-chain to enable cross-system supervision and auditing. AI entities are modeled as nodes in a dynamic interaction graph, where edges represent time-specific behavioral trajectories. Based on this model, a risk diffusion algorithm is proposed to trace the origin of risky behaviors and propagate early warnings across involved entities. System performance is evaluated using blockchain Transactions Per Second (TPS) metrics, demonstrating the feasibility and stability of AAT under large-scale interaction recording. AAT provides a scalable and verifiable solution for AI auditing, risk management, and responsibility attribution in complex multi-agent environments.

cs.AI↗

arXiv2025

Avalanches of choice: how stranger-to-stranger interactions shape crowd dynamics

Pedestrian routing choices play a crucial role in shaping collective crowd dynamics, yet the influence of interactions among unfamiliar individuals remains poorly understood. In this study, we analyze real-world pedestrian behavior at a route split within a busy train station using high-resolution trajectory data collected over a three-year time frame. We disclose a striking tendency for individuals to follow the same path as the person directly in front of them, even in the absence of social ties and even when such a choice leads to a longer travel time. This tendency leads to bursty dynamics, where sequences of pedestrians make identical decisions in succession, leading to strong patterns in collective movement. We employ a stochastic model that includes route costs, randomness, and social imitation to accurately reproduce the observed behavior, highlighting that local imitation behavior is the dominant driver of collective routing choices. These findings highlight how brief, low-level interactions between strangers can scale up to influence large-scale pedestrian movement, with strong implications for crowd management, urban design, and the broader understanding of social behavior in public spaces.

physics.soc-ph↗

arXiv2025

Inertial rotation of a small oblate spheroid in a simple shear flow

We compare experiments and fully-resolved particle simulations designed to match the experimental conditions of a weakly inertial, neutrally buoyant, moderately oblate spheroid in shear flow under confinement. Experimental and numerical results are benchmarked against theory valid for asymptotically small particle Reynolds numbers and for unconfined systems. By considering the combined effects of confinement and inertia, sensitivity to initial conditions, and the time span of observation, we reconcile the findings of theory, experiments, and numerical simulations. Furthermore, we demonstrate that confinement significantly influences the orientational stability of log-rolling spheroids compared to weak inertia, with the primary consequence being a reduced drift rate towards the stable log-rolling orbit.

physics.flu-dyn↗

arXiv2025

Recorded Versus Synthetic Spectral-compatible Ground Motions: A Comparative Analysis of Structural Seismic Responses

This paper presents a comparative analysis of structural seismic responses under two types of ground motion inputs: (i) synthetic motions generated by stochastic spectral-compatible ground motion models and (ii) recorded motions from an earthquake database. Both ground motion datasets are calibrated to a shared target response spectrum to ensure consistent spectral median, variance, and correlation structure. Five key stochastic response metrics-probability distributions, statistical moments, correlations, tail indices, and variance-based global sensitivity indices-are systematically evaluated for two representative structures: a medium-period building and a limiting case of a long-period tower. The comparison accounts for uncertainties both from ground motion and structural parameters. The results reveal that synthetic motions closely replicate recorded motions in terms of global response behavior-including distributions, mean and variance, correlation structure, and dominant uncertainty sources-indicating their suitability for routine seismic design and parametric studies. However, substantial differences emerge in response extremes for long-period structures, particularly in metrics governed by rare events, such as higher-order moments and tail behavior. These differences, which often exceed 50%, can be attributed to the non-Gaussian features and complex characteristics inherent in recorded motions, which are less pronounced in synthetic datasets. The findings support the use of synthetic ground motions for evaluating global seismic response characteristics, while highlighting their limitations in capturing rare-event behavior and long-period structural dynamics.

physics.geo-ph↗

arXiv2025

Tracking the rotation of light magnetic particles in turbulence

Particle-laden turbulence involves complex interactions between the dispersed and continuous phases. Given that particles can exhibit a wide range of properties, such as varying density, size, and shape, their interplay with the flow can lead to various modifications of the turbulence. Therefore, understanding the dynamics of particles is a necessary first step toward revealing the behavior of the multiphase system. Within the context of particle dynamics, accurately resolving rotational motion presents a significantly greater challenge compared to translational motion. We propose an experimental method to track the rotational motion of spherical, light, and magnetic particles with sizes significantly smaller than the Taylor microscale, typically an order of magnitude larger than the Kolmogorov scale of the turbulence in which they are suspended. The method fully resolves all three components of the particle angular velocity using only 2D images acquired from a single camera. This technique enables a detailed investigation of the rotational dynamics of magnetic particles subjected simultaneously to small-scale turbulent structures and external magnetic forcing. Beyond advancing the study of particle dynamics in turbulence, this approach opens new possibilities for actively modulating turbulence through externally applied magnetic fields.

physics.flu-dyn↗

arXiv2025

Shape-invariant Potentials and Singular Spaces

In this work, we present two brane-world-type solutions in a two-dimensional (2D) dilaton gravity model with singular space-time backgrounds. By employing a first-order superpotential formalism, we first construct the 2D analogues of the thick brane solution previously given by Gremm and analyze the corresponding linear scalar perturbations. We show that for a model with canonical scalar matter fields, the effective potential of the linear perturbation equation is a singular Pöschl--Teller~II type, which does not admit bound states. However, for a model with non-canonical scalar fields, the effective potential becomes an exactly solvable Pöschl--Teller~I potential, which has an infinite tower of normalizable bound states. We also present a second analytic solution inspired by the work of Girardello \emph{et al.}, but with non-canonical scalar field. In this case, the linear perturbation equation is a Schrödinger equation with the Eckart potential, which is also exactly solvable.

hep-th↗

arXiv2025

OmniFuser: Adaptive Multimodal Fusion for Service-Oriented Predictive Maintenance

Accurate and timely prediction of tool conditions is critical for intelligent manufacturing systems, where unplanned tool failures can lead to quality degradation and production downtime. In modern industrial environments, predictive maintenance is increasingly implemented as an intelligent service that integrates sensing, analysis, and decision support across production processes. To meet the demand for reliable and service-oriented operation, we present OmniFuser, a multimodal learning framework for predictive maintenance of milling tools that leverages both visual and sensor data. It performs parallel feature extraction from high-resolution tool images and cutting-force signals, capturing complementary spatiotemporal patterns across modalities. To effectively integrate heterogeneous features, OmniFuser employs a contamination-free cross-modal fusion mechanism that disentangles shared and modality-specific components, allowing for efficient cross-modal interaction. Furthermore, a recursive refinement pathway functions as an anchor mechanism, consistently retaining residual information to stabilize fusion dynamics. The learned representations can be encapsulated as reusable maintenance service modules, supporting both tool-state classification (e.g., Sharp, Used, Dulled) and multi-step force signal forecasting. Experiments on real-world milling datasets demonstrate that OmniFuser consistently outperforms state-of-the-art baselines, providing a dependable foundation for building intelligent industrial maintenance services.

cs.AI↗

arXiv2025

Learning Intractable Multimodal Policies with Reparameterization and Diversity Regularization

Traditional continuous deep reinforcement learning (RL) algorithms employ deterministic or unimodal Gaussian actors, which cannot express complex multimodal decision distributions. This limitation can hinder their performance in diversity-critical scenarios. There have been some attempts to design online multimodal RL algorithms based on diffusion or amortized actors. However, these actors are intractable, making existing methods struggle with balancing performance, decision diversity, and efficiency simultaneously. To overcome this challenge, we first reformulate existing intractable multimodal actors within a unified framework, and prove that they can be directly optimized by policy gradient via reparameterization. Then, we propose a distance-based diversity regularization that does not explicitly require decision probabilities. We identify two diversity-critical domains, namely multi-goal achieving and generative RL, to demonstrate the advantages of multimodal policies and our method, particularly in terms of few-shot robustness. In conventional MuJoCo benchmarks, our algorithm also shows competitive performance. Moreover, our experiments highlight that the amortized actor is a promising policy model class with strong multimodal expressivity and high performance. Our code is available at https://github.com/PneuC/DrAC

cs.LG↗

arXiv2025

High thermal conductivity of rutile-GeO$_2$ films grown by MOCVD: $52.9~\mathrm{W\,m^{-1}\,K^{-1}}$

Rutile germanium dioxide (r-GeO2) has recently emerged as a promising ultrawide-bandgap (UWBG) semiconductor owing to its wide bandgap (~4.4-5.1 eV), ambipolar doping potential, and high theoretical thermal conductivity. However, experimental data on the thermal conductivity of r-GeO2 epitaxial layers have not been reported, primarily due to challenges in phase control and surface roughness. Here, we report a high thermal conductivity of 52.9 +/- 6.6 W m^-1 K^-1 for high-quality (002) r-GeO2 films grown by metal-organic chemical vapor deposition (MOCVD) and characterized using time-domain thermoreflectance (TDTR). The phase control was achieved through a seed-driven stepwise crystallization (SDSC) approach, and the surface roughness was significantly reduced from 76 nm to 16 nm (locally as low as 1 A) via chemical mechanical polishing (CMP). These results highlight the promise of r-GeO2 as a UWBG oxide platform for power electronics applications.

cond-mat.mtrl-sci↗

arXiv2025

Dimensionality reduction can be used as a surrogate model for high-dimensional forward uncertainty quantification

We introduce a method to construct a stochastic surrogate model from the results of dimensionality reduction in forward uncertainty quantification. The hypothesis is that the high-dimensional input augmented by the output of a computational model admits a low-dimensional representation. This assumption can be met by numerous uncertainty quantification applications with physics-based computational models. The proposed approach differs from a sequential application of dimensionality reduction followed by surrogate modeling, as we "extract" a surrogate model from the results of dimensionality reduction in the input-output space. This feature becomes desirable when the input space is genuinely high-dimensional. The proposed method also diverges from the Probabilistic Learning on Manifold, as a reconstruction mapping from the feature space to the input-output space is circumvented. The final product of the proposed method is a stochastic simulator that propagates a deterministic input into a stochastic output, preserving the convenience of a sequential "dimensionality reduction + Gaussian process regression" approach while overcoming some of its limitations. The proposed method is demonstrated through two uncertainty quantification problems characterized by high-dimensional input uncertainties.

stat.AP↗

arXiv2025

A composition of simplified physics-based model with neural operator for trajectory-level seismic response predictions of structural systems

Accurate prediction of nonlinear structural responses is essential for earthquake risk assessment and management. While high-fidelity nonlinear time history analysis provides the most comprehensive and accurate representation of the responses, it becomes computationally prohibitive for complex structural system models and repeated simulations under varying ground motions. To address this challenge, we propose a composite learning framework that integrates simplified physics-based models with a Fourier neural operator to enable efficient and accurate trajectory-level seismic response prediction. In the proposed architecture, a simplified physics-based model, obtained from techniques such as linearization, modal reduction, or solver relaxation, serves as a preprocessing operator to generate structural response trajectories that capture coarse dynamic characteristics. A neural operator is then trained to correct the discrepancy between these initial approximations and the true nonlinear responses, allowing the composite model to capture hysteretic and path-dependent behaviors. Additionally, a linear regression-based postprocessing scheme is introduced to further refine predictions and quantify associated uncertainty with negligible additional computational effort. The proposed approach is validated on three representative structural systems subjected to synthetic or recorded ground motions. Results show that the proposed approach consistently improves prediction accuracy over baseline models, particularly in data-scarce regimes. These findings demonstrate the potential of physics-guided operator learning for reliable and data-efficient modeling of nonlinear structural seismic responses.

stat.AP↗