Search arXiv⌕ Search

arXiv subjects

Hao Yu

Publications and source records attributed to Hao Yu.

At least 55 records · Page 3Linked to original sources
arXiv2026

PaDoc: Layout-Grounded Parallel Decoding for Document Parsing

End-to-end document parsers provide a unified interface, but serialize page layouts and regional contents into one autoregressive sequence. This formulation forces independent regions onto a decoding path whose length grows with the total content, whereas crop-based two-stage parsers expose region-level parallelism at the cost of repeated visual prefills and fragmented page context. To retain full-page context while removing dependencies, we propose PaDoc, a layout-grounded parser that treats the predicted layout as a branching structure over a shared page representation. Under a region-sufficiency assumption, we derive a prefix-conditioned factorization in which the layout stream and regional content branches advance concurrently, reducing the decoding depth to the longest layout-content path. We realize this factorization within a single MLLM: packed variable-length ancestor attention preserves the visibility under standard next-token training, while masked parallel decoding creates branches that the evaluated vLLM backend serves as concurrent requests with cache-resident shared-prefix reuse. On OmniDocBench Full, PaDoc attains an Overall layout F1 of 91.1 and, among end-to-end parsers, a top-tier Overall score of 94.24 together with the best Text Edit (0.038) and Formula CDM (95.59). On a 384-page subset and one A800 GPU, it is the fastest end-to-end parser at five concurrency levels, improving valid-page throughput by 67.4-118% and reducing P95 latency by 39.2-54.9% relative to a same-backbone Sequential SFT baseline. Code is available at https://github.com/Longin-Yu/Padoc

cs.AI↗
arXiv2026

InfiniSplat: Implicit Gaussian Decoding for Large-Baseline Monocular View Synthesis

Single-image feed-forward 3D Gaussian Splatting (3DGS) aims to directly generate a renderable 3D scene representation from one input image, avoiding the cost of multi-view capture and per-scene optimization. However, existing methods are often constrained by a pixel-aligned representation, where Gaussians are predicted from fixed image-grid locations. Such pixel-aligned primitives can produce promising nearby-view renderings, but they remain weakly coupled to underlying scene surfaces and struggle to preserve coherent structures under large viewpoint shifts. We present InfiniSplat, a feed-forward single-image 3DGS framework that moves from a pixel-aligned representation toward a surface-aligned representation. InfiniSplat constructs this representation by first using geometry-guided sampling to place 2D supports according to depth-induced local surface structure, and then applying a query-conditioned implicit decoder to predict Gaussian attributes from the image features queried at these supports. By grounding support locations in geometry while decoupling Gaussian prediction from fixed pixel centers, InfiniSplat produces Gaussian layouts that better follow scene surfaces and reduce scattered primitives caused by grid discretization. Across multiple cross-dataset NVS evaluations, InfiniSplat achieves state-of-the-art performance compared with single-image feed-forward baselines, and demonstrates zero-shot generalization from Hypersim indoor synthetic training to complex open-world scenes. Project page: https://zju3dv.github.io/InfiniSplat.

cs.CV↗
arXiv2026

Face De-Identification: A Domain-Centric Survey from Capture to Processing

Face de-identification (De-ID) aims to remove or conceal personally identifiable facial features in images or videos to prevent identity recognition while preserving utility for downstream tasks. With the rising emphasis on data privacy and responsible AI, face De-ID has emerged as an active research area spanning computer vision and privacy-preserving communities. Early approaches, and many contemporary ones, operate in the digital domain by modifying pixel-level or appearance-level features through post-capture processing. Recent advances extend face De-ID beyond post-processing by integrating privacy mechanisms directly into sensors during image acquisition, bridging sensing systems and downstream vision algorithms. In parallel, physical-domain methods explore wearable accessories and materials that conceal identity information in real-world environments prior to capture. In this survey, we present the first unified overview that spans the full data acquisition pipeline, encompassing the physical, sensor, and digital domains. Through this domain-centric lens, we systematically analyze current methodologies, technical progress, and the distinct challenges inherent to each stage. We then review and organize existing evaluation protocols, examining current practices and highlighting the critical need for standardized, comprehensive benchmarks. Finally, we identify key open problems and outline emerging research directions to guide future work in this rapidly evolving field. To support ongoing research, we maintain a project page that organizes relevant literature with collected datasets and open source code: https://github.com/CV-AC/Awesome-FaceDe-ID.

cs.CV↗
arXiv2026

Qualitative Shadow Anomaly as a Smoking-gun Signature of Nonmetricity

In general relativity and its metric-formalism extensions, the innermost stable circular orbit and the photon ring of a charged black hole contract as the charge increases. In this paper, we discover a qualitative anomaly that violates this universal behavior, serving as a smoking-gun signature of geometric nonmetricity. By considering a minimal coupling between the bosonic field and the independent affine metric, we demonstrate that nonmetricity induces an effective geometric force that triggers a trajectory expansion of both massless and massive bosons with increasing black hole charge. This qualitative inversion directly imprints onto the black hole shadow, lifting the degeneracy between metric-affine gravity and general relativity. Using Eddington-inspired Born-Infeld gravity as a concrete implementation, we show that next-generation horizon-scale imaging can resolve this signature, probing a complementary high-energy window up to the $10^{5}_{~}\,\text{TeV}$ scale.

gr-qc↗
arXiv2026

TOUR: A Trajectory-Level Unlearning Benchmark for Offline Reinforcement Learning

Offline Reinforcement Learning (RL) agents are trained on fixed behavioral trajectories, which makes trajectory-level deletion important when selected data must be removed after training. Evaluating such deletion is difficult because a lower membership score can reflect trajectory removal, residual memorization visible to another attack, or policy collapse that destroys useful behavior. We introduce Trajectory-level memOrization and Unlearning in offline RL (TOUR), a benchmark that combines trajectory-level partitioning, matched non-member controls, retraining references, retained-performance anchors, and multi-attack privacy auditing. Across D4RL locomotion experiments and an exploratory AntMaze extension, TOUR shows that common deletion baselines have environment-dependent privacy-utility behavior. Retraining and fine-tuning often provide stronger retained-utility references than uniform GA+Refit, while TrajDeleter remains a useful comparator but is not uniformly stronger under the same audit. Reference-model, threshold, deviation, equivalence, action-error, representation-based, and query-limited attacks further show that a single likelihood-based membership score can overstate deletion quality. In the evaluated settings, conclusions about offline RL unlearning are therefore not stable under single-score auditing. They depend on matched non-member construction, retraining-relative calibration, attack family, retained utility, and explicit scope for diagnostic architecture or component-level evidence.

cs.LG↗
arXiv2026

ClawBench: Can AI Agents Complete Everyday Online Tasks?

AI agents may be able to assist with emails and documents, but can they reliably complete everyday online workflows on real websites? Everyday online tasks offer a realistic yet unsolved testbed for evaluating the next generation of AI agents. To this end, we introduce ClawBench, an evaluation framework comprising 153 everyday online tasks that people need to accomplish regularly in their lives and work, spanning 144 platforms across 15 categories, from completing purchases and booking appointments to submitting job applications. These tasks require capabilities beyond existing benchmarks, such as obtaining relevant information from user-provided documents, navigating multi-step workflows across diverse platforms, and write-heavy operations like filling in many detailed forms correctly. Unlike existing benchmarks that evaluate agents in offline sandboxes with static pages, ClawBench operates on production websites, preserving the full complexity, dynamic nature, and interaction challenges of real-world web environments. An interception layer captures and blocks the final submission request, ensuring safe evaluation without real-world side effects. Our evaluations of 8 frontier models show that both proprietary and open-source models complete only a small portion of these tasks. For example, Claude Sonnet 4.6 achieves only 33.3%, which exposes gaps in current AI agents. Progress on ClawBench brings us closer to AI agents that can function as general-purpose assistants.

cs.CL↗
arXiv2026

Search-on-Graph-R1: Training Large Language Models to Search Knowledge Graphs with Reinforcement Learning

Knowledge graph question answering (KGQA) requires navigating from topic entities to an answer several relations away. Recent methods prompt a frontier LLM to explore the graph through a retrieval tool, but their reliance on frontier-scale inference makes them costly to deploy. We present Search-on-Graph-R1 (\sogrone{}), which internalizes this navigation into a compact 8B model through supervised fine-tuning (SFT) followed by reinforcement learning (RL). Our central idea is to scaffold a frontier teacher with each question's gold SPARQL query, so the teacher traverses a known answer-bearing path with a live \texttt{Search} tool rather than having to discover the path itself. Since every call executes against a live Freebase server, the resulting trajectories are grounded in the knowledge graph by construction. On WebQSP, CWQ, and GrailQA, \sogrone{} at 8B surpasses every frozen frontier-LLM system in our comparison and posts the strongest results on CWQ of any system we compare against. It does so using no auxiliary module at inference and no LLM judge during training. Isolating each training stage shows that SFT and RL contribute complementary gains, our approach transfers across model families, and RL learns to reach answers in fewer \texttt{Search} calls than its SFT initialization.

cs.CL↗
arXiv2026

QUADS: Stabilizing NVFP4 Reinforcement Learning for MoE via QUantization-error Alignment across Dual Sides

Rollout generation is a major bottleneck in Reinforcement Learning (RL) for Mixture-of-Experts (MoE) Large Language Models, motivating low-precision rollout acceleration such as FP8. As an emerging low-precision format, NVFP4 combines fine-grained scaling for accuracy preservation with native W4A4 FP4 GEMMs for higher throughput than FP8. However, we find that directly applying NVFP4 to MoE RL rollout is impractical. NVFP4 rollout with BF16 training collapses after roughly 150 steps, accompanied by rapidly growing rollout-trainer log-probability gaps. Through training-inference error analysis and controlled ablations, we identify activation error, rather than weight error, as the dominant source of FP4 RL instability: weights can be synchronized and aligned by a shared quantization-dequantization path, whereas activations are recomputed online and error is amplified by the coarse E2M1 grid. Therefore, to stabilize NVFP4 RL for MoE, we propose QUantization-error Alignment across Dual Sides (QUADS). On the trainer side, we introduce Asymmetric Quantization-Aware Training fake-quantizing weights while keeping activations unquantized for better alignment. On the rollout side, Residual Activation Compensation corrects high-error activation channels while preserving native W4A4 GEMMs. In our MoE RL experiments on several benchmarks, QUADS achieves BF16-level accuracy, improves average pass@1 by 21.49 points over naive NVFP4 RL, and delivers ~16% higher rollout throughput than FP8.

cs.LG↗
arXiv2026

Partitioning-free 3D-IC Floorplanning

3D integration with fine-pitch hybrid bonding offers a promising path to alleviate interconnect bottlenecks in conventional two-dimensional (2D) ICs, yet efficient 3D floorplanning remains challenging due to the enlarged solution space and non-uniform inter-die communication latency. Existing methods either extend 2D representations into 3D, leading to combinatorial complexity, or adopt partitioning-first pipelines that fix block-to-die assignments early and hinder joint optimization of floorplan, die assignment, and vertical connectivity. In this work, we present \textsc{Great3D}, a partitioning-free 3D floorplanning framework that directly optimizes a native 3D floorplan. \textsc{Great3D} formulates a unified objective that couples interconnect cost with a cycles-per-instruction (CPI)-derived latency term to capture the system-level impact of face-to-face (F2F) bonding. Algorithmically, it combines an SDP-based 3D global embedding with a dynamic-programming refinement for die assignment, followed by 2D continuous refinement with practical design constraints. \textcolor{blue}{Experiments on the GSRC and ATPlace benchmark suites show that \textsc{Great3D} consistently achieves strong wirelength and CPI quality against state-of-the-art 3D floorplanners. On GSRC, it reduces total wirelength by up to about $70\%$ (and by $2.40$--$2.74\times$ on average) over competing 3D-native floorplanners, and its dynamic-programming die-assignment stage further improves CPI by $9.5$--$17.8\%$, while maintaining competitive runtime on instances of up to a few hundred blocks.}

cs.ET↗
arXiv2026

Joint Energy Management and Coordinated AIGC Workload Scheduling for Distributed Data Centers: A Diffusion-Aided Reward Shaping Approach

Artificial intelligence-generated content (AIGC) has emerged as a transformative paradigm for automating the creation of diverse and customized content, giving rise to rapidly growing computational workloads in cloud data centers. It is imperative for AIGC service providers (ASPs) to strategically schedule AIGC workloads to reduce data center energy costs while guaranteeing high-quality content generation. However, the distinctive characteristics of AIGC services pose critical challenges, including model heterogeneity across ASPs, implicit service quality evaluation, and complex inference process control. To tackle these challenges, we propose a joint energy management and coordinated AIGC workload scheduling framework, which introduces an explicit mathematical characterization of service quality to promote both job transfer among ASPs and fine-grained inference process configuration. Moreover, various energy resources within data centers are jointly considered to enhance power usage flexibility. Subsequently, a system utility maximization problem is formulated to balance AIGC service revenue with operational penalties and costs. Nevertheless, the strong coupling among job scheduling decisions induces severe reward sparsity, which limits the effectiveness of existing deep reinforcement learning (DRL) algorithms. To address this issue, we develop a diffusion model-aided reward shaping approach to synthesize complementary reward signals through a multi-step denoising process. This approach is seamlessly integrated with DRL to enable efficient learning of scheduling policies under sparse environmental feedback. Experiments based on real-world models and datasets demonstrate that our scheme effectively accommodates electricity price fluctuations and AIGC model heterogeneity, while achieving superior learning convergence and system utility compared with benchmark methods.

cs.LG↗
arXiv2026

The classification of Bidihedral Groups

A group is called bidihedral if it can be expressed as a product of two dihedral subgroups. In this paper, a complete classification for all bidihedral groups is given.

math.GR↗
arXiv2026

Uniform large deviations and long-time dynamics for the 2D anisotropic Navier-Stokes equations on the torus

We study the two-dimensional stochastic Navier-Stokes equations on the torus with horizontal dissipation and additive noise. First, we prove a uniform large deviation principle for the solution paths in the energy space $C([0,T];H).$ The proof combines a large deviation principle for a linear Ornstein-Uhlenbeck process, a new Lipschitz estimate for the nonlinear map, and a contraction principle that avoids any exponential tightness condition. Second, we analyse the long-time dynamics. By exhibiting a deterministic steady state immune to the horizontal dissipation, we show that exponential mixing cannot hold. Finally, we investigate the invariant measures of the system. In the deterministic case, the infinite-dimensional kernel of the horizontal Laplacian contains infinitely many Dirac invariant measures. Under a suitably degenerate noise, however, the kernel supports no invariant measure at all; moreover, any invariant measure of this stochastic system, should it exist, must be orthogonal to this kernel. This sharply distinguishes the anisotropic equations on the torus both from the fully dissipative Navier-Stokes equations and from their counterparts on domains with no-slip boundaries.

math.AP↗
arXiv2026

LightFARM: Model Predictive Lighting Control with Battery-Free IoT for Energy-Efficient Indoor Farming

Lighting is the dominant energy load in indoor farming, yet most deployed systems still rely on fixed rule-based or schedule-based control. We present LightFARM, a predictive lighting control framework that couples crop illumination with battery-free sensing for more energy-efficient indoor farming. LightFARM combines finite-horizon predictive control with compact models of photosynthesis, thermal dynamics, and sensor energy state. The controller adjusts lighting intensity to balance photosynthetic benefit, electrical power consumption, thermal safety, and sensing-energy feasibility. A key design feature is that the same light-emitting diode (LED) fixtures serve both as the photosynthetic light source for crops and as a controllable energy source for self-powered sensor nodes. We implement LightFARM in a real indoor basil cultivation system and evaluate it through two independent 12-day cultivation trials. Compared with a conventional rule-based baseline, LightFARM reduces lighting energy consumption by approximately 41% and improves energy productivity from 36.1 to 52.9 $\mathrm{g\,kWh^{-1}}$ and from 41.1 to 60.2 $\mathrm{g\,kWh^{-1}}$ ($\approx 46.5\%$ on average). These results suggest that energy-cooperative predictive lighting control is a promising approach to improving indoor farming efficiency under practical resource constraints, while explicitly accounting for the trade-off between energy savings and crop yield.

eess.SP↗
arXiv2026

AfriqueLLM: How Data Mixing and Model Architecture Impact Continued Pre-training for African Languages

Large language models (LLMs) are increasingly multilingual, yet open models continue to underperform relative to proprietary systems, with the gap most pronounced for African languages. Continued pre-training (CPT) offers a practical route to language adaptation, but improvements on demanding capabilities such as mathematical reasoning often remain limited. This limitation is driven in part by the uneven domain coverage and missing task-relevant knowledge that characterize many low-resource language corpora. We present \texttt{AfriqueLLM}, a suite of open LLMs adapted to 20 African languages through CPT on 26B tokens. We perform a comprehensive empirical study across five base models spanning sizes and architectures, including Llama 3.1, Gemma 3, and Qwen 3, and systematically analyze how CPT data composition shapes downstream performance. In particular, we vary mixtures that include math, code, and synthetic translated data, and evaluate the resulting models on a range of multilingual benchmarks. Our results identify data composition as the primary driver of CPT gains. Adding math, code, and synthetic translated data yields consistent improvements, including on reasoning-oriented evaluations. Within a fixed architecture, larger models typically improve performance, but architectural choices dominate scale when comparing across model families. Moreover, strong multilingual performance in the base model does not reliably predict post-CPT outcomes; robust architectures coupled with task-aligned data provide a more dependable recipe. Finally, our best models improve long-context performance, including document-level translation. Models and code have been released on [Huggingface](https://huggingface.co/collections/McGill-NLP/afriquellm) and [Github](https://github.com/McGill-NLP/AfriqueLLM).

cs.CL↗
arXiv2026

Beyond Per-Request QoS: Coordinating Industrial Workflows with B5G/6G Network Capabilities

Beyond-5G (B5G) and 6G networks are expected to enable more complex industrial services, which often operate according to multi-phase workflows with phase-specific communication requirements. However, current interaction between applications and networks remains predominantly request-driven: Quality of Service (QoS) is requested at each workflow phase transition and evaluated independently, without explicit consideration of upcoming demand or network's near-term capability. This mismatch limits the ability of both sides to plan ahead, often resulting in foreseeable incompatibilities, even service disruptions. This article presents a capability-aware coordination framework for workflow-based industrial services. Within a bounded planning window, the network exposes the QoS profiles it can sustainably support, while the industrial side maps upcoming workflow phases to these disclosed capabilities and submits the resulting demand trajectory for joint assessment. The framework also supports coordinated updates when network conditions change during execution. An industrial video inspection case study on a real B5G system, complemented by large-scale simulation, illustrates that such coordination can improve service continuity, reduce disruptive rejections, and increase workflow completion under heavy load. The results suggest that future industrial networking should move beyond reactive per-request QoS handling toward forward-looking, capability-aware, workflow-level coordination.

cs.NI↗
arXiv2026

Autonomous Orbit Determination Analysis of a Conceptual Cislunar Navigation Constellation based on Inter-Satellite Range Measurement

With the community's increasing interest in the cislunar space, building a navigation constellation servicing the whole cislunar space has become a pressing need. Previous studies mainly focus on constellations using orbits close to the Moon, which limits the servicing volume of the constellation. In this work, a four-satellite constellation using one L3 orbit, one L4 orbit, one L5 orbit and an orbit close to the Moon is proposed. The orbit determination accuracy is an important factor to be considered when designing parameters of the constellation. In this study, the mode of autonomous orbit determination (AOD) based on inter-satellite range data is considered. With such a model, the out-of-plane design parameters are identified as the main parameters influencing the AOD accuracy. For the AOD based on short arcs, we find that the increase of the out-of-plane amplitude can improve the AOD accuracy, and the out-of-plane initial phases have a more complex influence. A novel relative planarity factor (RPF) $P_\text{r}$, which has negative correlation with the AOD accuracy, is proposed as the metric to evaluate the variation of AOD performance. Using $P_\text{r}$, we demonstrate that the coplanarity of the constellation can significantly reduce the AOD accuracy. For the long arc AOD, the influence of different parameters is insignificant.

astro-ph.EP↗
arXiv2026

Asymptotic Counting of Binary Phylogenetic Networks

Phylogenetic networks provide a general framework for modeling reticulate evolutionary processes such as hybridization, recombination, and horizontal gene transfer. In this paper, we study the asymptotic counting of binary phylogenetic networks with $k$ reticulations on $n$ taxa, where $k$ is allowed to grow with $n$. Using edge insertion, we analyze the local structures that affect the number of possible constructions of such networks. By bounding the contribution of networks with exceptional local configurations and combining these bounds with known asymptotic formulas for tree-child networks, we show that, when $k=o(\sqrt n)$, the number of binary phylogenetic networks with $k$ reticulations on $n$ taxa is asymptotic to \[ \binom{n}{k}2^{n+k-1/2}n^{n+k-1}e^{-n}. \]

q-bio.PE↗
arXiv2026

Where, What, Why, and Importance: Structured Defect Grounding for Text-to-Image Feedback

Despite generating increasingly photorealistic images, text-to-image (T2I) models still exhibit localized, subtle, and structurally complex failures. Diagnosing these failures requires instance-level feedback that answers where a defect occurs, what type it is, why it is defective, and its importance to overall image quality. While recent dense-feedback methods move beyond scalar supervision, their heatmap-centric representations still formulate diagnosis as pixel-field regression, making it difficult to localize variable-cardinality defects and bind semantic reasons to individual failures. To address this representation bottleneck, we propose Structured Defect Grounding (SDG), which casts T2I diagnosis as structured set prediction by modeling each defect as a (location, type, reason, importance) tuple. To make this formulation trainable and measurable, we introduce SDG-30K, a 30K-image dataset with box-grounded annotations across four modern T2I generators, together with a dedicated evaluation protocol, SDG-Eval. Building on this structured representation, we further present a diagnosis-to-alignment framework in which a Vision-Language Model (VLM) serves as the SDG detector, and BoxFlow-GRPO converts predicted defect sets into box-derived, importance-weighted spatial rewards for diffusion model alignment. Extensive experiments show that our SDG detector outperforms leading proprietary VLMs on structured defect grounding, while SDG-guided rewards consistently improve T2I alignment and support localized image refinement. These results establish SDG as a unified, instance-level interface for diagnosing, evaluating, and enhancing modern generative models.

cs.CV↗