Search arXiv⌕ Search

arXiv subjects

Hao Zhang

Publications and source records attributed to Hao Zhang.

At least 325 records · Page 18Linked to original sources
arXiv2026

Audio Outperforms Text for Visual Decoding

Decoding visual semantic representations from human brain activity is a significant challenge. While recent zero-shot decoding approaches have improved performance by leveraging aligned image-text datasets, they overlook a fundamental aspect of human cognition: semantic understanding is inherently anchored in the auditory modality of speech, not text. To address this, our study introduces the first comparative framework for evaluating auditory versus textual semantic modalities in zero-shot visual neural decoding. We propose a novel brain-visual-auditory multimodal alignment model that directly utilizes auditory representations to encapsulate semantics, serving as a substitute for traditional textual descriptors. Our experimental results demonstrate that the auditory modality not only surpasses the textual modality in decoding accuracy but also achieves higher computational efficiency. These findings indicate that auditory semantic representations are more closely aligned with neural activity patterns during visual processing. This work reveals the critical and previously underestimated role of auditory semantics in decoding visual cognition and provides new insights for developing brain-computer interfaces that are more congruent with natural human cognitive mechanisms.

q-bio.NC↗
arXiv2026

Average first-passage times for character sums

Let $\varepsilon>0$ and, for an odd prime $p$, set $$ S_\ell(p):=\sum_{n\le \ell}\left(\frac{n}{p}\right). $$ Define the first-passage time $$ f_\varepsilon(p):=\min\{\ell\ge 1:\ S_\ell(p)<\varepsilon\ell\}. $$ We prove that there exists a constant $c_\varepsilon>0$ such that, as $x\to\infty$, $$ \sum_{p\le x} f_\varepsilon(p)\sim c_\varepsilon \frac{x}{\log x}. $$

math.NT↗
arXiv2026

Electronic procrystalline state in moire structures

Solid state materials can display varieties of atomic structural orders ranging from crystalline to amorphous, underlying their properties and diverse functionalities. Procrystal has emerged as a new category of solids, featuring a long-range ordered lattice framework tiled with disordered atomic or molecular structures on the lattice sites, arousing great interest due to its novel structural and physical properties. However, the electronic analogue of a procrystal, dubbed as an electronic procrystalline (EPC) state, has never been experimentally observed. Here, we report the observation of an EPC state in a moire superstructure formed between a monolayer metallic NiTe2 and a superconductor NbSe2 with incommensurate lattice wavevectors. The observed EPC state exhibits a long-range periodic charge modulation at the moire scale inlaid with short-range irregular orders within each moire cell. Strikingly, the short-range charge orders inside the moire unit cells have proximately root3*root3 quasi-period, which is absent in pristine NiTe2. Intriguingly, the EPC order is also observed in the superconducting state of the moire superstructure. Furthermore, the emergent EPC state and short-range charge order, coexisting with the proximity induced superconductivity, can be precisely modulated with the thickness of NiTe2. Our findings uncover the potential of moire platform for understanding and tuning novel correlated quantum phases with this exotic procrystalline order.

cond-mat.mtrl-sci↗
arXiv2026

Arctic-Text2SQL-R1: Simple Rewards, Strong Reasoning in Text-to-SQL

Translating natural language into SQL (Test2SQL) is a longstanding challenge at the intersection of natural language understanding and structured data access. While large language models (LLMs) have significantly improved fluency in SQL generation, producing correct and executable SQL--particularly for complex queries--remains a bottleneck. We present Arctic-Text2SQL-R1, a reinforcement learning (RL) framework and model family designed to generate accurate, executable SQL using a lightweight reward signal based solely on execution correctness. Our approach avoids brittle intermediate supervision and complex reward shaping, promoting stable training and alignment with the end task. Combined with carefully curated data, strong supervised initialization, and effective training practices, Arctic-Text2SQL-R1 achieves state-of-the-art execution accuracy across six diverse Test2SQL benchmarks, including the top position on the BIRD leaderboard. Notably, our 7B model outperforms prior 70B-class systems, highlighting the framework's scalability and efficiency. We further demonstrate inference-time robustness through simple extensions like value retrieval and majority voting. Extensive experiments and ablation studies offer both positive and negative insights, providing practical guidance for future Test2SQL research.

cs.CL↗
arXiv2026

MI-PRUN: Optimize Large Language Model Pruning via Mutual Information

Large Language Models (LLMs) have become indispensable across various domains, but this comes at the cost of substantial computational and memory resources. Model pruning addresses this by removing redundant components from models. In particular, block pruning can achieve significant compression and inference acceleration. However, existing block pruning methods are often unstable and struggle to attain globally optimal solutions. In this paper, we propose a mutual information based pruning method MI-PRUN for LLMs. Specifically, we leverages mutual information to identify redundant blocks by evaluating transitions in hidden states. Additionally, we incorporate the Data Processing Inequality (DPI) to reveal the relationship between the importance of entire contiguous blocks and that of individual blocks. Moreover, we develop the Fast-Block-Select algorithm, which iteratively updates block combinations to achieve a globally optimal solution while significantly improving the efficiency. Extensive experiments across various models and datasets demonstrate the stability and effectiveness of our method.

cs.CL↗
arXiv2026

The optimal error analysis of nonuniform L1 method for the variable-exponent subdiffusion model

This work investigates the optimal error estimate of the fully discrete scheme for the variable-exponent subdiffusion model under the nonuniform temporal mesh. We apply the perturbation method to reformulate the original model into its equivalent form, and apply the L1 scheme as well as the interpolation quadrature rule to discretize the Caputo derivative term and the convolution term in the reformulated model, respectively. We then prove the temporal convergence rates $O(N^{-\min\{2-α(0), rα(0)\}})$ under the nonuniform mesh, which improves the existing convergence results in [Zheng, CSIAM T. Appl. Math. 2025] for $r\geq \frac{2-α(0)}{α(0)}$. Numerical results are presented to substantiate the theoretical findings.

math.NA↗
arXiv2026

RigMo: Unifying Rig and Motion Learning for Generative Animation

Despite significant progress in 4D generation, rig and motion, the core structural and dynamic components of animation are typically modeled as separate problems. Existing pipelines rely on ground-truth skeletons and skinning weights for motion generation and treat auto-rigging as an independent process, undermining scalability and interpretability. We present RigMo, a unified generative framework that jointly learns rig and motion directly from raw mesh sequences, without any human-provided rig annotations. RigMo encodes per-vertex deformations into two compact latent spaces: a rig latent that decodes into explicit Gaussian bones and skinning weights, and a motion latent that produces time-varying SE(3) transformations. Together, these outputs define an animatable mesh with explicit structure and coherent motion, enabling feed-forward rig and motion inference for deformable objects. Beyond unified rig-motion discovery, we introduce a Motion-DiT model operating in RigMo's latent space and demonstrate that these structure-aware latents can naturally support downstream motion generation tasks. Experiments on DeformingThings4D, Objaverse-XL, and TrueBones demonstrate that RigMo learns smooth, interpretable, and physically plausible rigs, while achieving superior reconstruction and category-level generalization compared to existing auto-rigging and deformation baselines. RigMo establishes a new paradigm for unified, structure-aware, and scalable dynamic 3D modeling.

cs.GR↗
arXiv2026

Gecko: An Efficient Neural Architecture Inherently Processing Sequences with Arbitrary Lengths

Designing a unified neural network to efficiently and inherently process sequential data with arbitrary lengths is a central and challenging problem in sequence modeling. The design choices in Transformer, including quadratic complexity and weak length extrapolation, have limited their ability to scale to long sequences. In this work, we propose Gecko, a neural architecture that inherits the design of Mega and Megalodon (exponential moving average with gated attention), and further introduces multiple technical components to improve its capability to capture long range dependencies, including timestep decay normalization, sliding chunk attention mechanism, and adaptive working memory. In a controlled pretraining comparison with Llama2 and Megalodon in the scale of 7 billion parameters and 2 trillion training tokens, Gecko achieves better efficiency and long-context scalability. Gecko reaches a training loss of 1.68, significantly outperforming Llama2-7B (1.75) and Megalodon-7B (1.70), and landing close to Llama2-13B (1.67). Notably, without relying on any context-extension techniques, Gecko exhibits inherent long-context processing and retrieval capabilities, stably handling sequences of up to 4 million tokens and retrieving information from contexts up to $4\times$ longer than its attention window. Code: https://github.com/XuezheMax/gecko-llm

cs.LG↗
arXiv2026

Kitaev interactions in the van der Waals antiferromagnet VBr3

Van der Waals materials hosting Kitaev interactions are promising platforms for exploring exotic quantum phenomena. Here, we report inelastic neutron scattering investigations of the van der Waals antiferromagnet VBr3, which forms a honeycomb lattice structure at room temperature and exhibits zigzag-type magnetic order below 26.5 K. Our observations reveal distinctive low-energy spin excitations arising from gamma, gamma', and M' points, each featuring a spin gap of around 2.5 meV. The overall spin excitation spectra can be effectively described by a spin Hamiltonian incorporating substantial nearest-neighbor Kitaev and biquadratic interactions, along with Heisenberg interactions. Our findings not only establish VBr3 as a new Kitaev magnet but also suggest that ligand engineering may provide a promising strategy to modulate Kitaev interactions, offering new opportunities for designing Kitaev materials with tailored quantum properties.

cond-mat.str-el↗
arXiv2026

Upstream Laser-based Longitudinal Enhancement of Relativistic Photoelectrons

Controlling the longitudinal phase space of high-brightness relativistic electron beams is crucial for advancing a broad spectrum of charged-particle-based instrumentation and scientific frontiers. A generalized method for achieving this control involves manipulating the photoemission laser's temporal distribution at the picosecond level, a long-standing technical challenge. Recent developments in laser shaping have enabled the creation of high-power, picosecond-scale symmetrical and asymmetrical temporal profiles, capable of fine-tuning complex space-charge dynamics and external field effects in relativistic charged-particle beams. Here, we demonstrate that rather than deviations from theorized, idealized laser distributions, a controlled asymmetry can be harnessed to counteract accelerator-induced distortions. By implementing spatiotemporal shaping of the ultraviolet photocathode laser at the LCLS-II superconducting injector, we achieve deterministic control over the longitudinal phase space without downstream corrections. We find that this optical asymmetry induces a self-linearizing effect across both low (40 pC) and high (80 pC) charge regimes, effectively suppressing nonlinear compression and energy chirp. Consequently, this approach is expected to preserve a low emittance comparable to that of ideal flattop or regular Gaussian profiles, while delivering superior current uniformity and shot-to-shot stability. These results establish spatiotemporal laser shaping as a compact, generalizable tool for directly optimizing beam brightness at the source.

physics.optics↗
arXiv2026

Iterative Structured Pruning for Large Language Models with Multi-Domain Calibration

Large Language Models (LLMs) have achieved remarkable success across a wide spectrum of natural language processing tasks. However, their ever-growing scale introduces significant barriers to real-world deployment, including substantial computational overhead, memory footprint, and inference latency. While model pruning presents a viable solution to these challenges, existing unstructured pruning techniques often yield irregular sparsity patterns that necessitate specialized hardware or software support. In this work, we explore structured pruning, which eliminates entire architectural components and maintains compatibility with standard hardware accelerators. We introduce a novel structured pruning framework that leverages a hybrid multi-domain calibration set and an iterative calibration strategy to effectively identify and remove redundant channels. Extensive experiments on various models across diverse downstream tasks show that our approach achieves significant compression with minimal performance degradation.

cs.CL↗
arXiv2026

Self-Supervised Learning with Noisy Dataset for Rydberg Microwave Sensors Denoising

We report a self-supervised deep learning framework for Rydberg sensors that enables single-shot noise suppression matching the accuracy of multi-measurement averaging. The framework eliminates the need for clean reference signals (hardly required in quantum sensing) by training on two sets of noisy signals with identical statistical distributions. When evaluated on Rydberg sensing datasets, the framework outperforms wavelet transform and Kalman filtering, achieving a denoising effect equivalent to 10,000-set averaging while reducing computation time by three orders of magnitude. We further validate performance across diverse noise profiles and quantify the complexity-performance trade-off of U-Net and Transformer architectures, providing actionable guidance for optimizing deep learning-based denoising in Rydberg sensor systems.

quant-ph↗
arXiv2026

Phonon-informed Crystal Structure Classification via Precision-Adaptive ResNet-based Confidence Ensemble

Accurate description of crystal structures is a prerequisite for predicting the physicochemical properties of materials. However, conventional X-ray diffraction (XRD) characterization often encounters intrinsic bottlenecks when applied to complex multiphase systems, necessitating the integration of complementary optical measurement. In this study, we developed a multi-descriptor framework by integrating key parameters including space groups, Pearson symbols, and Wyckoff sequences, to categorize the dataset of over 19,000 crystals into several dozen structural prototypes. Then, an accuracy-adaptive ensemble network based on residual architectures was implemented to capture structural ``fingerprints" within phonon vibration modes and Raman spectra. The ensemble algorithm demonstrates exceptional robustness when processing various crystals of varying lengths and quality. This data-driven classification strategy not only overcomes the reliance of traditional characterization on ideal data but also provides a high-throughput tool for the automated analysis of material structures in large-scale experimental workflows.

cond-mat.mtrl-sci↗
arXiv2026

Real-Time Forecasting of Pathological Gait via IMU Navigation: A Few-Shot and Generative Learning Framework for Wearable Devices

Current gait analysis faces challenges in various aspects, including limited and poorly labeled data within existing wearable electronics databases, difficulties in collecting patient data due to privacy concerns, and the inadequacy of the Zero-Velocity Update Technique (ZUPT) in accurately analyzing pathological gait patterns. To address these limitations, we introduce GaitMotion, a novel machine-learning framework that employs few-shot learning on a multitask dataset collected via wearable IMU sensors for real-time pathological gait analysis. GaitMotion enhances data quality through detailed, ground-truth-labeled sequences and achieves accurate step and stride segmentation and stride length estimation, which are essential for diagnosing neurological disorders. We incorporate a generative augmentation component, which synthesizes rare or underrepresented pathological gait patterns. GaitMotion achieves a 65\% increase in stride length estimation accuracy compared to ZUPT. In addition, its application to real patient datasets via transfer learning confirms its robust predictive capability. By integrating generative AI into wearable gait analysis, GaitMotion not only refines the precision of pathological gait forecasting but also demonstrates a scalable framework for leveraging synthetic data in biomechanical pattern recognition, paving the way for more personalized and data-efficient digital health services.

eess.SP↗
arXiv2025

TrimTokenator-LC: Towards Adaptive Visual Token Pruning for Large Multimodal Models with Long Contexts

Large Multimodal Models (LMMs) have proven effective on various tasks. They typically encode visual inputs into Original Model sequences of tokens, which are then concatenated with textual tokens and jointly processed by the language model. However, the growing number of visual tokens greatly increases inference cost. Visual token pruning has emerged as a promising solution. However, existing methods often overlook scenarios involving long context inputs with multiple images. In this paper, we analyze the challenges of visual token pruning in long context, multi-image settings and introduce an adaptive pruning method tailored for such scenarios. We decompose redundancy into intra-image and inter-image components and quantify them through intra-image diversity and inter-image variation, which jointly guide dynamic budget allocation. Our approach consists of two stages. The intra-image stage allocates each image a content-aware token budget and greedily selects its most representative tokens. The inter-image stage performs global diversity filtering to form a candidate pool and then applies a Pareto selection procedure that balances diversity with text alignment. Extensive experiments show that our approach can reduce up to 80% of visual tokens while maintaining performance in long context settings.

cs.CV↗
arXiv2025

Bidirectional Sparse Attention for Faster Video Diffusion Training

Video diffusion Transformer (DiT) models excel in generative quality but hit major computational bottlenecks when producing high-resolution, long-duration videos. The quadratic complexity of full attention leads to prohibitively high training and inference costs. Full attention inefficiency stems from two key challenges: excessive computation due to the inherent sparsity of Queries and Key-Value pairs, and redundant computation as fixed sparse patterns fail to leverage DiT's dynamic attention. To overcome this limitation, we propose a Bidirectional Sparse Attention (BSA) framework for faster video DiT training, the first to dynamically sparsify both Queries and Key-Value pairs within 3D full attention, thereby substantially improving training and inference efficiency. BSA addresses these issues through two key components. Query sparsity is optimized by selecting the most informative query tokens via semantic similarity and with a dynamic spatial-time training strategy, while KV sparsity is achieved by computing a statistical dynamic threshold to retain only the most salient KV blocks for computation. Extensive experiments demonstrate that BSA significantly accelerates DiT training across long sequences, reducing FLOPs by up to 20x and achieving 17.79x faster attention training, while preserving or even surpassing the generative quality of full attention.

cs.CV↗
arXiv2025

Complexity and convergence analysis of a single-loop SDCAM for Lipschitz composite optimization and beyond

We develop and analyze a single-loop algorithm for minimizing the sum of a Lipschitz differentiable function $f$, a prox-friendly proper closed function $g$ (with a closed domain on which $g$ is continuous) and the composition of another prox-friendly proper closed function $h$ (whose domain is closed on which $h$ is continuous) with a continuously differentiable mapping $c$ (that is Lipschitz continuous and Lipschitz differentiable on the convex closure of the domain of $g$). Such models arise naturally in many contemporary applications, where $f$ is the loss function for data misfit, and $g$ and $h$ are nonsmooth functions for inducing desirable structures in $x$ and $c(x)$. Existing single-loop algorithms mainly focus either on the case where $h$ is Lipschitz continuous or the case where $h$ is an indicator function of a closed convex set. In this paper, we develop a single-loop algorithm for more general possibly non-Lipschitz $h$. Our algorithm is a single-loop variant of the successive difference-of-convex approximation method (SDCAM) proposed in [22]. We show that when $h$ is Lipschitz, our algorithm exhibits an iteration complexity that matches the best known complexity result for obtaining an $(ε_1,ε_2,0)$-stationary point. Moreover, we show that, by assuming additionally that dom $g$ is compact, our algorithm exhibits an iteration complexity of $\tilde{O}(ε^{-4})$ for obtaining an $(ε,ε,ε)$-stationary point when $h$ is merely continuous and real-valued. Furthermore, we consider a scenario where $h$ does not have full domain and establish vanishing bounds on successive changes of iterates. Finally, in all three cases mentioned above, we show that one can construct a subsequence such that any accumulation point $x^*$ satisfies $c(x^*)\in$ dom $h$, and if a standard constraint qualification holds at $x^*$, then $x^*$ is a stationary point.

math.OC↗