arXiv subjects

Hongyu Wu

Publications and source records attributed to Hongyu Wu.

At least 19 recordsLinked to original sources

arXiv2026

SeRV: Semantic-Aligned Residual Vector Quantization for American Sign Language Generation

American Sign Language (ASL) generation remains challenging due to limited paired text-ASL motion data and the difficulty of learning motion representations both precise for reconstruction and predictable from linguistic input. Existing methods rely on motion tokenizers optimized for reconstruction, without explicit semantic supervision from paired text. As a result, the learned tokens remain limited in supporting semantically consistent and fine-grained ASL motion generation. To address this limitation, we propose SeRV (Semantic-Aligned Residual Vector Quantization), a semantic-aligned RVQ tokenizer for ASL generation. SeRV learns a semantically structured residual token space by combining sentence-level motion-text alignment with token-level text-conditioned supervision. Building on this tokenizer, a Hierarchical GPT predicts residual motion tokens in a coarse-to-fine manner, generating structurally coherent and semantically aligned 3D ASL motion. We further construct a large-scale reconstructed 3D ASL motion-text benchmark by recovering paired 3D motion from YouTube-ASL videos. Experiments across 375 hours of ASL video show that SeRV achieves state-of-the-art pose accuracy on both How2Sign and YouTube-ASL datasets, while producing semantically consistent 3D ASL motion directly from text.

cs.CV↗

arXiv2026

Beyond Local Surprise: Grounded Dialogue as Selective Belief Revision under Referential Uncertainty

When a speaker refers to a scene that the listener cannot directly see, the listener must decide whether to preserve its current understanding or revise it as new utterances arrive. Many language systems treat local mismatch as a cue for updating: divergence from the current understanding encourages adjustment. Yet conversational understanding may be more conservative, interpreting mismatching evidence relative to prior understanding rather than immediately revising it. We introduce a controlled, data-driven framework for turn-by-turn preserve/revise decisions in dialogue, where competing revision policies are learned under otherwise identical conditions. We compare four theory-driven revision strategies, each reflecting a different assumption about when listeners should preserve or revise. Two findings stand out. First, a mismatch-driven policy that updates solely based on local divergence reacts strongly to mismatch but destabilizes grounding and degrades retrieval. Second, an uncertainty-sensitive policy extends mismatch-based updating with accumulated evidence, preserving coherent understanding while maintaining strong retrieval performance. Surprisingly, coherent understanding emerges from a counterintuitive pattern: local mismatch promotes preservation, whereas accumulated uncertainty promotes revision, suggesting that listeners maintain prior understanding despite local mismatch and revise only when uncertainty sufficiently accumulates. This pattern is consistent with conceptual pact theory.

cs.CL↗

arXiv2026

VoxStruct3D: Structure-Leading Flow Matching for Voxel-Space 3D MRI Synthesis

High-fidelity 3D MRI synthesis requires both globally coherent anatomy and fine-grained voxel-level detail. Although latent diffusion makes volumetric generation tractable, its image autoencoder introduces a reconstruction bottleneck that can limit the fine detail recoverable in the final volume. We present VoxStruct3D, a voxel-space flow-matching framework that directly models full-resolution MRI volumes using a clean-data prediction objective. Its Volumetric Voxel Generator (VVG) combines factorized 3D patch embedding with overlapping upsampling, time-modulated residual refinement, and skip fusion, enabling neighboring tokens to jointly reconstruct shared voxel regions and suppress patch-boundary artifacts. To complement direct voxel-space modeling with an explicit anatomical prior, we further introduce a Structure-First, Image-Follows (SFIF) strategy. A frozen pretrained 3D medical encoder and a StructVAE extract compact structure tokens that preserve dominant anatomy, while a structure-leading schedule keeps their trajectory ahead of the image trajectory. Patch-Aligned RoPE spatially aligns the unequal token grids, and asymmetric attention enforces one-way guidance from structure to image. Experiments on pathological and healthy T1-weighted brain MRI datasets show that VoxStruct3D achieves the strongest overall performance across feature-distribution alignment, sample diversity, and perceptual quality, producing anatomically coherent and visually realistic volumes.

cs.CV↗

arXiv2026

QuerySplat: Decoupling Geometry and Appearance Representations in 3DGS Prediction

While feed-forward 3D Gaussian Splatting (3DGS) enables efficient 3D reconstruction, achieving high-fidelity rendering remains challenging. Existing pixel-aligned approaches suffer from spatial inflexibility and massive structural redundancy, whereas query-based methods lack 3D priors and entangle geometry with appearance, yielding blurry, pose-dependent results. To overcome these deficiencies, we propose \textbf{QuerySplat}, a feed-forward 3DGS framework driven by geometric priors and explicit appearance decoupling. Specifically, we design a dual-branch query-based decoder: the geometry branch leverages a pretrained Vision Geometric Model for spatial understanding, which intrinsically endows QuerySplat with pose-free modeling capabilities, while the appearance branch recovers high-frequency details through a dedicated pathway separated from geometric attribute regression. Extensive experiments demonstrate that QuerySplat mitigates the blurry rendering issues of early query-based models and consistently outperforms pixel-aligned approaches in rendering fidelity. On the challenging DL3DV benchmark, it achieves state-of-the-art novel view synthesis performance, with average PSNR gains of 2.30 dB and 1.04 dB over the best pose-free and pose-required baselines, respectively. Project Page: https://inspatio.github.io/querysplat.

cs.CV↗

arXiv2026

A Mission-Centric Cyber-Resilience Benchmark for Silent-Watch Operation of Electrified Ground-Platform Power Architectures

Silent-watch operation makes electrified ground platforms depend on supervisory energy management because mission loads must be sustained from stored energy while the engine is off. This paper develops a mission-centric cyber-resilience benchmark for this operating mode. The benchmark connects battery state-of-charge (SOC) spoofing to mission outcomes rather than evaluating the attack only through detector response or control error. It combines a reduced-order DC-bus model, residual-based detection, fallback shedding, and five mission-facing metrics for endurance, critical-load service, priority-weighted loss-of-load cost, unsafe-voltage exposure, and detection delay. The study shows that SOC spoofing creates a structured stealth-versus-impact envelope. Small biases have limited mission effect, intermediate biases produce an endurance deficit well approximated by a first-order expression in bias magnitude, shed power, and average battery draw, and large biases disable the SOC-driven guard. The results also show that defense value depends on fallback depth, not detection alone. An undersized fallback action can leave the Defended case failing to complete the mission despite early detection. MATLAB-to-Simulink parity across five regression scenarios provides a software-verified basis for hardware-in-the-loop testing.

eess.SY↗

arXiv2025

A Rule-Aware Prompt Framework for Structured Numeric Reasoning in Cyber-Physical Systems

Smart grids rely on high-dimensional numeric telemetry and explicit operating rules to maintain reliable and secure operation. Recent large language models (LLMs) are increasingly considered as candidate decision-support components for power system operations, yet most deployments focus on textual logs, alerts, or operator messages and do not directly address rule-grounded reasoning over numeric grid measurements. This paper proposes a rule-aware prompt framework that systematically encodes power system domain context, numeric normalization, and decision rules into a modular prompt architecture for LLMs. The framework decomposes prompts into reusable modules, including role, domain context, numeric normalization, rule-aware reasoning, value block, and output schema, and exposes an interface for plugging in diverse grid operating rules. A key design element separates rule specification from the representation of normalized numeric deviations, enabling concise prompts aligned with power system criteria. To illustrate its behavior, we instantiate the framework on numeric anomaly detection in the IEEE 118-bus transmission network and evaluate several prompting and adaptation regimes. The results show that rule-aware, z-score-based value blocks and a hybrid LLM+DL architecture substantially improve both consistency with grid operating rules and anomaly detection performance while reducing token usage, providing a reusable bridge between grid telemetry and general-purpose LLMs.

eess.SY↗

arXiv2025

Evaluation of Large Language Models for Numeric Anomaly Detection in Power Systems

Large language models (LLMs) have gained increasing attention in power grids for their general-purpose capabilities. Meanwhile, anomaly detection (AD) remains critical for grid resilience, requiring accurate and interpretable decisions based on multivariate telemetry. Yet the performance of LLMs on large-scale numeric data for AD remains largely unexplored. This paper presents a comprehensive evaluation of LLMs for numeric AD in power systems. We use GPT-OSS-20B as a representative model and evaluate it on the IEEE 14-bus system. A standardized prompt framework is applied across zero-shot, few-shot, in-context learning, low rank adaptation (LoRA), fine-tuning, and a hybrid LLM-traditional approach. We adopt a rule-aware design based on the three-sigma criterion, and report detection performance and rationale quality. This study lays the groundwork for further investigation into the limitations and capabilities of LLM-based AD and its integration with classical detectors in cyber-physical power grid applications.

eess.SY↗

arXiv2025

RoboCOIN: An Open-Sourced Bimanual Robotic Data Collection for Integrated Manipulation

Despite the critical role of bimanual manipulation in endowing robots with human-like dexterity, large-scale and diverse datasets remain scarce due to the significant hardware heterogeneity across bimanual robotic platforms. To bridge this gap, we introduce RoboCOIN, a large-scale multi-embodiment bimanual manipulation dataset comprising over 180,000 demonstrations collected from 15 distinct robotic platforms. Spanning 16 diverse environments-including residential, commercial, and industrial settings-the dataset features 421 bimanual tasks systematically categorized by 39 bimanual collaboration actions and 432 objects. A key innovation of our work is the hierarchical capability pyramid, which provides granular annotations ranging from trajectory-level concepts to segment-level subtasks and frame-level kinematics. Furthermore, we present CoRobot, an efficient data processing pipeline powered by the Robot Trajectory Markup Language (RTML), designed to facilitate quality assessment, automated annotation, and unified multi-embodiment and data management. Extensive experiments demonstrate the effectiveness of RoboCOIN in enhancing the performance of various bimanual manipulation models across a wide spectrum of robotic embodiments. The entire dataset and codebase are fully open-sourced, providing a valuable resource for advancing research in bimanual and multi-embodiment manipulation.

cs.RO↗

arXiv2025

Data-model Coevolution as the Architectural Principle for AI-Native Materials Databases

AI-native approaches are reshaping computational materials discovery into iterative data-model coevolution cycles. However, most existing materials databases remain fundamentally data-centric, where predictive models remain external to database state and data growth is decoupled from model updating. Here we formalize data-model coevolution as the architectural basis of AI-native materials databases, where data and predictive models evolve through endogenous generation-evaluation-refinement cycles. Using the Li-P-S ternary as a demonstrative prototype, we generated approximately 70,000 candidate structures, more than 10,000 of which satisfy the stable-unique-novel (S.U.N.) criterion, achieving rapid saturation of local chemical environments together with stabilization of energy distributions. We autonomously found chemically plausible phases and motifs outside the Materials Project (MP) and Alexandria databases, including a stable Li$_2$PS$_3$ phase, the (PS$_3$)$_3^{3-}$ trimer, the (P$_3$S$_8$)$^{3-}$ ring, two isomers of the (P$_2$S$_8$)$^{2-}$ ring, and polymeric (PS$_4$)$_n^{n-}$ chains. Within two to three iterations, the integrated predictive models converged to high precision under a low first-principles cost, and the resulting data-model state can be directly queried for atomistic and electronic-structure properties within the same unified framework. Data-model states can be reused and extended across related chemical systems, enabling scalable and continuous accumulation of computational materials knowledge. These results demonstrate data-model coevolution as a practical architectural principle for AI-era materials data infrastructure.

cond-mat.mtrl-sci↗

arXiv2025

Nested AutoRegressive Models

AutoRegressive (AR) models have demonstrated competitive performance in image generation, achieving results comparable to those of diffusion models. However, their token-by-token image generation mechanism remains computationally intensive and existing solutions such as VAR often lead to limited sample diversity. In this work, we propose a Nested AutoRegressive~(NestAR) model, which proposes nested AutoRegressive architectures in generating images. NestAR designs multi-scale modules in a hierarchical order. These different scaled modules are constructed in an AR architecture, where one larger-scale module is conditioned on outputs from its previous smaller-scale module. Within each module, NestAR uses another AR structure to generate ``patches'' of tokens. The proposed nested AR architecture reduces the overall complexity from $\mathcal{O}(n)$ to $\mathcal{O}(\log n)$ in generating $n$ image tokens, as well as increases image diversities. NestAR further incorporates flow matching loss to use continuous tokens, and develops objectives to coordinate these multi-scale modules in model training. NestAR achieves competitive image generation performance while significantly lowering computational cost.

cs.CV↗

arXiv2025

Gryphon: Scaling Hyperscale Multi-Tenant Gateways Beyond the Petabit-Era via DPU-Augmented Hierarchical Co-Offloading

At ByteDance, cloud gateway clusters orchestrate petabit-scale aggregate traffic. Traditional ASIC-only gateways fail to meet these escalating demands due to severe on-chip resource constraints and limited programmable flexibility, while pure software solutions or alternatives like disaggregated SmartNICs struggle to match terabit-scale line-rate throughput. To bridge this gap, we present Gryphon, a hyperscale cloud gateway built on a hybrid architecture that integrates DPUs directly into the switching ASIC's forwarding path. This design resolves the fundamental tension between capacity and speed, expanding table scale by up to 1000$\times$ and augmenting programmability, while sustaining 1.6 Tbps line-rate throughput at a cost of only ~8$\mu s$ in additional average latency. To manage this hardware heterogeneity, we introduce Hierarchical Co-Offloading (HLCO) in the data plane, achieving >99.9% fast path hit rate, while retaining software fallback for complex operations. In the control plane, we develop an abstraction layer (P4Bridge) that decouples hardware specifics from policy configuration. Gryphon has been operating at production scale for over a year, deployed on hundreds of nodes across multiple Availability Zones. We also share production measurements and operational experiences that serve as the first hyperscale-proven guidelines for next-generation DPU-augmented cloud gateways.

cs.NI↗

arXiv2025

Segment Any 3D-Part in a Scene from a Sentence

This paper aims to achieve the segmentation of any 3D part in a scene based on natural language descriptions, extending beyond traditional object-level 3D scene understanding and addressing both data and methodological challenges. Due to the expensive acquisition and annotation burden, existing datasets and methods are predominantly limited to object-level comprehension. To overcome the limitations of data and annotation availability, we introduce the 3D-PU dataset, the first large-scale 3D dataset with dense part annotations, created through an innovative and cost-effective method for constructing synthetic 3D scenes with fine-grained part-level annotations, paving the way for advanced 3D-part scene understanding. On the methodological side, we propose OpenPart3D, a 3D-input-only framework to effectively tackle the challenges of part-level segmentation. Extensive experiments demonstrate the superiority of our approach in open-vocabulary 3D scene understanding tasks at the part level, with strong generalization capabilities across various 3D scene datasets.

cs.CV↗

arXiv2025

3D Dental Model Segmentation with Geometrical Boundary Preserving

3D intraoral scan mesh is widely used in digital dentistry diagnosis, segmenting 3D intraoral scan mesh is a critical preliminary task. Numerous approaches have been devised for precise tooth segmentation. Currently, the deep learning-based methods are capable of the high accuracy segmentation of crown. However, the segmentation accuracy at the junction between the crown and the gum is still below average. Existing down-sampling methods are unable to effectively preserve the geometric details at the junction. To address these problems, we propose CrossTooth, a boundary-preserving segmentation method that combines 3D mesh selective downsampling to retain more vertices at the tooth-gingiva area, along with cross-modal discriminative boundary features extracted from multi-view rendered images, enhancing the geometric representation of the segmentation network. Using a point network as a backbone and incorporating image complementary features, CrossTooth significantly improves segmentation accuracy, as demonstrated by experiments on a public intraoral scan dataset.

cs.CV↗

arXiv2025

Pre-training, fine-tuning, and distillation (PFD): Automatically generating machine learning force fields from universal models

Universal force fields generalizable across the periodic table represent a new trend in computational materials science. However, the applications of universal force fields in material simulations are limited by their slow inference speed and the lack of first-principles accuracy. Instead of building a single model simultaneously satisfying these characteristics, a strategy that quickly generates material-specific models from the universal model may be more feasible. Here, we propose a new workflow pattern, PFD (Pre-training, Fine-tuning, and Distillation), which automatically generates machine-learning force fields for specific materials from a pre-trained universal model through fine-tuning and distillation. By fine-tuning the pre-trained model, our PFD workflow generates force fields with first-principles accuracy while requiring one to two orders of magnitude less training data compared to traditional methods. The inference speed of the generated force field is further improved through distillation, meeting the requirements of large-scale molecular simulations. Comprehensive testing across diverse materials including complex systems, such as amorphous carbon, interface, etc., reveals marked enhancements in training efficiency, which suggests the PFD workflow a practical and reliable approach for force field generation in computational material sciences.

cond-mat.mtrl-sci↗

arXiv2025

SCoT: Unifying Consistency Models and Rectified Flows via Straight-Consistent Trajectories

Pre-trained diffusion models are commonly used to generate clean data (e.g., images) from random noises, effectively forming pairs of noises and corresponding clean images. Distillation on these pre-trained models can be viewed as the process of constructing advanced trajectories within the pair to accelerate sampling. For instance, consistency model distillation develops consistent projection functions to regulate trajectories, although sampling efficiency remains a concern. Rectified flow method enforces straight trajectories to enable faster sampling, yet relies on numerical ODE solvers, which may introduce approximation errors. In this work, we bridge the gap between the consistency model and the rectified flow method by proposing a Straight Consistent Trajectory~(SCoT) model. SCoT enjoys the benefits of both approaches for fast sampling, producing trajectories with consistent and straight properties simultaneously. These dual properties are strategically balanced by targeting two critical objectives: (1) regulating the gradient of SCoT's mapping to a constant, (2) ensuring trajectory consistency. Extensive experimental results demonstrate the effectiveness and efficiency of SCoT.

cs.CV↗

arXiv2025

A Survey on Pre-Trained Diffusion Model Distillations

Diffusion Models~(DMs) have emerged as the dominant approach in Generative Artificial Intelligence (GenAI), owing to their remarkable performance in tasks such as text-to-image synthesis. However, practical DMs, such as stable diffusion, are typically trained on massive datasets and thus usually require large storage. At the same time, many steps may be required, i.e., recursively evaluating the trained neural network, to generate a high-quality image, which results in significant computational costs during sample generation. As a result, distillation methods on pre-trained DM have become widely adopted practices to develop smaller, more efficient models capable of rapid, few-step generation in low-resource environment. When these distillation methods are developed from different perspectives, there is an urgent need for a systematic survey, particularly from a methodological perspective. In this survey, we review distillation methods through three aspects: output loss distillation, trajectory distillation and adversarial distillation. We also discuss current challenges and outline future research directions in the conclusion.

cs.LG↗

arXiv2024

A compositional ordering-driven morphotropic phase boundary in ferroelectric solid solutions

Ferroelectric solid solutions usually exhibit giant dielectric response and high piezoelectricity in the vicinity of the morphotropic phase boundary (MPB), where the structural phase transitions between the rhombohedral and the tetragonal phases as a result of the composition or strain variation. Here, we propose a compositional ordering-driven MPB in the specified compositional solid solutions. By preforming machine-learning potential based molecular dynamics simulations on lead zirconate titanate, we find a phase transition from the rhombohedral to tetragonal phase with the decrease of compositional ordering, leading to the MPB on temperature-ordering phase diagram. The compositional ordering-driven MPB can enhances the piezoelectricity with a magnitude comparable to that at the composition-driven MPB. Finally, we demonstrate that the mechanism of high piezoelectricity is polarization rotation driven by external field. This work provides an additional degree of freedom, compositional ordering, to design the high-performance piezoelectric materials.

cond-mat.mtrl-sci↗

arXiv2024

Parametric Primitive Analysis of CAD Sketches with Vision Transformer

The design and analysis of Computer-Aided Design (CAD) sketches play a crucial role in industrial product design, primarily involving CAD primitives and their inter-primitive constraints. To address challenges related to error accumulation in autoregressive models and the complexities associated with self-supervised model design for this task, we propose a two-stage network framework. This framework consists of a primitive network and a constraint network, transforming the sketch analysis task into a set prediction problem to enhance the effective handling of primitives and constraints. By decoupling target types from parameters, the model gains increased flexibility and optimization while reducing complexity. Additionally, the constraint network incorporates a pointer module to explicitly indicate the relationship between constraint parameters and primitive indices, enhancing interpretability and performance. Qualitative and quantitative analyses on two publicly available datasets demonstrate the superiority of this method.

cs.CV↗