arXiv subjects

Jun Zhang

Publications and source records attributed to Jun Zhang.

At least 271 records · Page 15Linked to original sources

arXiv2025

Feed-Forward 3D Gaussian Splatting Compression with Long-Context Modeling

3D Gaussian Splatting (3DGS) has emerged as a revolutionary 3D representation. However, its substantial data size poses a major barrier to widespread adoption. While feed-forward 3DGS compression offers a practical alternative to costly per-scene per-train compressors, existing methods struggle to model long-range spatial dependencies, due to the limited receptive field of transform coding networks and the inadequate context capacity in entropy models. In this work, we propose a novel feed-forward 3DGS compression framework that effectively models long-range correlations to enable highly compact and generalizable 3D representations. Central to our approach is a large-scale context structure that comprises thousands of Gaussians based on Morton serialization. We then design a fine-grained space-channel auto-regressive entropy model to fully leverage this expansive context. Furthermore, we develop an attention-based transform coding model to extract informative latent priors by aggregating features from a wide range of neighboring Gaussians. Our method yields a $20\times$ compression ratio for 3DGS in a feed-forward inference and achieves state-of-the-art performance among generalizable codecs.

cs.CV↗

arXiv2025

Local unanimity in Shapley-Scarf housing markets

In the housing market model introduced by Shapley and Scarf (1974), we propose a new axiom, local unanimity, that extends the unanimity condition widely used in social choice theory. It applies the unanimity condition to any subset of agents in the model who unanimously agree on the best exchange of their endowments. Building on this axiom, we provide several concise characterizations of the Top Trading Cycles (TTC) mechanism under both strict and weak preference domains.

econ.TH↗

arXiv2025

RemedyGS: Defend 3D Gaussian Splatting against Computation Cost Attacks

As a mainstream technique for 3D reconstruction, 3D Gaussian splatting (3DGS) has been applied in a wide range of applications and services. Recent studies have revealed critical vulnerabilities in this pipeline and introduced computation cost attacks that lead to malicious resource occupancies and even denial-of-service (DoS) conditions, thereby hindering the reliable deployment of 3DGS. In this paper, we propose the first effective and comprehensive black-box defense framework, named RemedyGS, against such computation cost attacks, safeguarding 3DGS reconstruction systems and services. Our pipeline comprises two key components: a detector to identify the attacked input images with poisoned textures and a purifier to recover the benign images from their attacked counterparts, mitigating the adverse effects of these attacks. Moreover, we incorporate adversarial training into the purifier to enforce distributional alignment between the recovered and original natural images, thereby enhancing the defense efficacy. Experimental results demonstrate that our framework effectively defends against white-box, black-box, and adaptive attacks in 3DGS systems, achieving state-of-the-art performance in both safety and utility.

cs.CV↗

arXiv2025

Bundled School Choice

This paper proposes a novel school choice system where schools are grouped into hierarchical bundles and offered to students as options for preference reports. By listing a bundle, a student seeks admission to any school within it without ranking the schools. This approach helps students who struggle to rank schools precisely and expands options on limited preference lists, potentially improving match outcomes. We design a modified deferred acceptance mechanism to handle bundle reports while preserving stability. Two laboratory experiments support our theory, showing that well-constructed bundles aligned with student preferences enhance welfare and match rates without compromising fairness. Practical applications are discussed.

econ.TH↗

arXiv2025

PaddleOCR-VL: Boosting Multilingual Document Parsing via a 0.9B Ultra-Compact Vision-Language Model

In this report, we propose PaddleOCR-VL, a SOTA and resource-efficient model tailored for document parsing. Its core component is PaddleOCR-VL-0.9B, a compact yet powerful vision-language model (VLM) that integrates a NaViT-style dynamic resolution visual encoder with the ERNIE-4.5-0.3B language model to enable accurate element recognition. This innovative model efficiently supports 109 languages and excels in recognizing complex elements (e.g., text, tables, formulas, and charts), while maintaining minimal resource consumption. Through comprehensive evaluations on widely used public benchmarks and in-house benchmarks, PaddleOCR-VL achieves SOTA performance in both page-level document parsing and element-level recognition. It significantly outperforms existing solutions, exhibits strong competitiveness against top-tier VLMs, and delivers fast inference speeds. These strengths make it highly suitable for practical deployment in real-world scenarios. Code is available at https://github.com/PaddlePaddle/PaddleOCR .

cs.CV↗

arXiv2025

VIL2C: Value-of-Information Aware Low-Latency Communication for Multi-Agent Reinforcement Learning

Inter-agent communication serves as an effective mechanism for enhancing performance in collaborative multi-agent reinforcement learning(MARL) systems. However, the inherent communication latency in practical systems induces both action decision delays and outdated information sharing, impeding MARL performance gains, particularly in time-critical applications like autonomous driving. In this work, we propose a Value-of-Information aware Low-latency Communication(VIL2C) scheme that proactively adjusts the latency distribution to mitigate its effects in MARL systems. Specifically, we define a Value of Information (VOI) metric to quantify the importance of delayed message transmission based on each delayed message's importance. Moreover, we propose a progressive message reception mechanism to adaptively adjust the reception duration based on received messages. We derive the optimized VoI aware resource allocation and theoretically prove the performance advantage of the proposed VIL2C scheme. Extensive experiments demonstrate that VIL2C outperforms existing approaches under various communication conditions. These gains are attributed to the low-latency transmission of high-VoI messages via resource allocation and the elimination of unnecessary waiting periods via adaptive reception duration.

cs.MA↗

arXiv2025

LumiTex: Towards High-Fidelity PBR Texture Generation with Illumination Context

Physically-based rendering (PBR) provides a principled standard for realistic material-lighting interactions in computer graphics. Despite recent advances in generating PBR textures, existing methods fail to address two fundamental challenges: 1) materials decomposition from image prompts under limited illumination cues, and 2) seamless and view-consistent texture completion. To this end, we propose LumiTex, an end-to-end framework that comprises three key components: (1) a multi-branch generation scheme that disentangles albedo and metallic-roughness under shared illumination priors for robust material understanding, (2) a lighting-aware material attention mechanism that injects illumination context into the decoding process for physically grounded generation of albedo, metallic, and roughness maps, and (3) a geometry-guided inpainting module based on a large view synthesis model that enriches texture coverage and ensures seamless, view-consistent UV completion. Extensive experiments demonstrate that LumiTex achieves state-of-the-art performance in texture quality, surpassing both existing open-source and commercial methods.

cs.CV↗

arXiv2025

Observation of Magnetostatic Surface Spin Wave Solitons in Yttrium Iron Garnet Thin Film

Magnetostatic surface spin wave (MSSW) solitons hold great promise for magnonic information processing, but their existence has long been debated. In this work, we resolve this issue by advanced time-resolved Brillouin light scattering (TR-BLS) spectroscopy. We observe long-period MSSW soliton trains in yttrium iron garnet (YIG) thin films by demonstrating their quasiparticle behavior, mode beating, and periodic modulations. We reveal that the MSSW soliton originates from the dipole gap mechanism with a unique comb-like frequency spectrum caused by the four-magnon process. By varying the microwave power, we find significant changes in soliton periods linked to the spin wave dispersion renormalization. Additionally, we report exotic transverse solitons across the YIG thickness. These findings deepen the understanding of nonlinear physics, and pave the way for spin-wave-soliton-based information technologies.

cond-mat.mes-hall↗

arXiv2025

Reg-DPO: SFT-Regularized Direct Preference Optimization with GT-Pair for Improving Video Generation

Recent studies have identified Direct Preference Optimization (DPO) as an efficient and reward-free approach to improving video generation quality. However, existing methods largely follow image-domain paradigms and are mainly developed on small-scale models (approximately 2B parameters), limiting their ability to address the unique challenges of video tasks, such as costly data construction, unstable training, and heavy memory consumption. To overcome these limitations, we introduce a GT-Pair that automatically builds high-quality preference pairs by using real videos as positives and model-generated videos as negatives, eliminating the need for any external annotation. We further present Reg-DPO, which incorporates the SFT loss as a regularization term into the DPO loss to enhance training stability and generation fidelity. Additionally, by combining the FSDP framework with multiple memory optimization techniques, our approach achieves nearly three times higher training capacity than using FSDP alone. Extensive experiments on both I2V and T2V tasks across multiple datasets demonstrate that our method consistently outperforms existing approaches, delivering superior video generation quality.

cs.CV↗

arXiv2025

Constraining $σ_8$ with Lensing Statistics in Low and High Density Regions

Lensing studies are typically carried out around high density regions, such as groups and clusters, where the lensing signals are significant and indicative of rich density structures. However, a more comprehensive test of the cosmological model should also include the lensing effect in low density regions. In this work, we incorporate the stacked weak lensing signals around the low density positions, alongside galaxy-galaxy lensing and galaxy-galaxy two point correlation function to perform a joint cosmological analysis on $σ_8$. The low density positions are constructed from the DR9 data release of the DESI legacy imaging survey, using galaxies with r-band absolute magnitude cut M$<$-21.5 and in the redshift range of 0.18$<$z$<$0.28. In doing so, we simultaneously parameterize photometric redshift errors and halo mass uncertainties while building mock catalogs from simulations using the method of SubHalo Abundance Matching (SHAM). For the weak lensing measurements, we use the shear estimators derived from the DECaLS DR8 imaging data, processed by the Fourier_Quad pipeline. The survey boundaries and masks are fully taken into account. Our analysis achieves a total significance of $31.1σ$ detection for lensing in the low density positions, which significantly improve the $σ_8$ constraint compared to galaxy-galaxy lensing and galaxy-galaxy two point correlation function by 14$\%$. For flat $Λ$CDM model, we constrain $σ_8$ =$0.824^{+0.015}_{-0.015}$, which shows a good agreement with the PLANCK result. Additionally, the halo mass uncertainty $σ_{\lg M}$ and photometric redshift error $σ_z$ are constrained to be $0.565^{+0.086}_{-0.070}$ and $0.004^{+0.004}_{-0.003}$ respectively, which are somewhat different from our expectations due to the significant degeneracy of the two parameters.

astro-ph.CO↗

arXiv2025

Towards General Auditory Intelligence: Large Multimodal Models for Machine Listening and Speaking

In the era of large language models (LLMs) and artificial general intelligence (AGI), computer audition must evolve beyond traditional paradigms to fully leverage the capabilities of foundation models, towards more comprehensive understanding, more natural generation and more human-like interaction. Audio, as a modality rich in semantic, emotional, and contextual cues, plays a vital role in achieving naturalistic and embodied machine intelligence. This survey provides a comprehensive review of recent progress in integrating audio into LLMs, with a focus on four key areas: audio comprehension, audio generation, speech-based interaction, and audio-visual understanding. We analyze how LLMs are reshaping audio perception and reasoning, enabling systems to understand sound at a deeper semantic level, generate expressive audio outputs, and engage in human-like spoken interaction. Furthermore, we explore how the fusion of audio and visual modalities enhances situational awareness and cross-modal reasoning, pushing the boundaries of multimodal intelligence. This survey not only synthesizes existing research but also identifies critical challenges and future directions for building audio-native AGI systems capable of perceiving, understanding, and interacting through sound as naturally as humans do.

eess.AS↗

arXiv2025

Resistance switch in ferromagnet/spin glass/ferromagnet spin valves

We obtain in single van-der-Waals layer of MnBi2Te4 the spin-valve-like ferromagnet/spin glass (SG)/ferromagnet architecture, where the switch of individual spin states in the SG-like layer appears as the resistance switch behavior. The characteristic temperature of SG can be effectively tuned by fine-control of Bi-doping in the SG layer. A doping- and temperature-dependent phase diagram is established. We demonstrate the remote manipulation and detection of the states of individual Mn-layer spins by tunneling electrons in favor of the electron-phonon, spin-phonon, spin-transfer torque and spin-flip interactions among hot electrons, lattice and local spins, where the spin valve layer is even buried below two-quintuple-layer Bi2Te3. The integration of the SG state into spin valves opens the opportunity of realizing atomic-scale spintronics by integrating different degrees of freedom in two dimensional materials.

cond-mat.mtrl-sci↗

arXiv2025

The spin-switch scanning tunneling microscopy: an architecture to probe electron-phonon interactions in the atomic scale

On the spin-valve-like ferromagnet/spin glass/ferromagnet (FM/SG/FM) structure, the tunneling current is dominated by resistance switch (RS) instead of the local density of states according to the conventional tunneling theory. Here we show lattice-site dependent RS behaviors in one-quintuple-layer Bi2Te3 deposited on single MnBi2Te4 septuple layer, which comes from the difference in the efficiency of tunneling electrons to induce focused current or phonons at different sites, switching remotely the spin valve by spin-transfer torque or spin-phonon interactions. These lead to the observation of the dynamic 2-state lattice when the tip scans across the surface as well as the ability of scanning tunneling microscope (STM) to reveal atomic-scale features of electron-phonon (EP) interactions. Our work demonstrates the possibility of the spin-switch STM to image lattice-site dependent EP interactions of any materials deposited on the FM/SG/FM structure.

cond-mat.mtrl-sci↗

arXiv2025

CAS-Spec: Cascade Adaptive Self-Speculative Decoding for On-the-Fly Lossless Inference Acceleration of LLMs

Speculative decoding has become a widely adopted as an effective technique for lossless inference acceleration when deploying large language models (LLMs). While on-the-fly self-speculative methods offer seamless integration and broad utility, they often fall short of the speed gains achieved by methods relying on specialized training. Cascading a hierarchy of draft models promises further acceleration and flexibility, but the high cost of training multiple models has limited its practical application. In this paper, we propose a novel Cascade Adaptive Self-Speculative Decoding (CAS-Spec) method which constructs speculative draft models by leveraging dynamically switchable inference acceleration (DSIA) strategies, including layer sparsity and activation quantization. Furthermore, traditional vertical and horizontal cascade algorithms are inefficient when applied to self-speculative decoding methods. We introduce a Dynamic Tree Cascade (DyTC) algorithm that adaptively routes the multi-level draft models and assigns the draft lengths, based on the heuristics of acceptance rates and latency prediction. Our CAS-Spec method achieves state-of-the-art acceleration compared to existing on-the-fly speculative decoding methods, with an average speedup from $1.1\times$ to $2.3\times$ over autoregressive decoding across various LLMs and datasets. DyTC improves the average speedup by $47$\% and $48$\% over cascade-based baseline and tree-based baseline algorithms, respectively. CAS-Spec can be easily integrated into most existing LLMs and holds promising potential for further acceleration as self-speculative decoding techniques continue to evolve.

cs.LG↗

arXiv2025

Semantic Communications with World Models

Semantic communication is a promising technique for emerging wireless applications, which reduces transmission overhead by transmitting only task-relevant features instead of raw data. However, existing methods struggle under extremely low bandwidth and varying channel conditions, where corrupted or missing semantics lead to severe reconstruction errors. To resolve this difficulty, we propose a world foundation model (WFM)-aided semantic video transmission framework that leverages the predictive capability of WFMs to generate future frames based on the current frame and textual guidance. This design allows transmissions to be omitted when predictions remain reliable, thereby saving bandwidth. Through WFM's prediction, the key semantics are preserved, yet minor prediction errors tend to amplify over time. To mitigate issue, a lightweight depth-based feedback module is introduced to determine whether transmission of the current frame is needed. Apart from transmitting the entire frame, a segmentation-assisted partial transmission method is proposed to repair degraded frames, which can further balance performance and bandwidth cost. Furthermore, an active transmission strategy is developed for mobile scenarios by exploiting camera trajectory information and proactively scheduling transmissions before channel quality deteriorates. Simulation results show that the proposed framework significantly reduces transmission overhead while maintaining task performances across varying scenarios and channel conditions.

eess.IV↗

arXiv2025

Task-Oriented Feature Compression for Multimodal Understanding via Device-Edge Co-Inference

With the rapid development of large multimodal models (LMMs), multimodal understanding applications are emerging. As most LMM inference requests originate from edge devices with limited computational capabilities, the predominant inference pipeline involves directly forwarding the input data to an edge server which handles all computations. However, this approach introduces high transmission latency due to limited uplink bandwidth of edge devices and significant computation latency caused by the prohibitive number of visual tokens, thus hindering delay-sensitive tasks and degrading user experience. To address this challenge, we propose a task-oriented feature compression (TOFC) method for multimodal understanding in a device-edge co-inference framework, where visual features are merged by clustering and encoded by a learnable and selective entropy model before feature projection. Specifically, we employ density peaks clustering based on K nearest neighbors to reduce the number of visual features, thereby minimizing both data transmission and computational complexity. Subsequently, a learnable entropy model with hyperprior is utilized to encode and decode merged features, further reducing transmission overhead. To enhance compression efficiency, multiple entropy models are adaptively selected based on the characteristics of the visual features, enabling a more accurate estimation of the probability distribution. Comprehensive experiments on seven visual question answering benchmarks validate the effectiveness of the proposed TOFC method. Results show that TOFC achieves up to 52% reduction in data transmission overhead and 63% reduction in system latency while maintaining identical task performance, compared with neural compression ELIC.

eess.SP↗

arXiv2025

FAN: Fourier Analysis Networks

Despite the remarkable successes of general-purpose neural networks, such as MLPs and Transformers, we find that they exhibit notable shortcomings in modeling and reasoning about periodic phenomena, achieving only marginal performance within the training domain and failing to generalize effectively to out-of-domain (OOD) scenarios. Periodicity is ubiquitous throughout nature and science. Therefore, neural networks should be equipped with the essential ability to model and handle periodicity. In this work, we propose FAN, a novel neural network that effectively addresses periodicity modeling challenges while offering broad applicability similar to MLP with fewer parameters and FLOPs. Periodicity is naturally integrated into FAN's structure and computational processes by introducing the Fourier Principle. Unlike existing Fourier-based networks, which possess particular periodicity modeling abilities but face challenges in scaling to deeper networks and are typically designed for specific tasks, our approach overcomes this challenge to enable scaling to large-scale models and maintains general-purpose modeling capability. Through extensive experiments, we demonstrate the superiority of FAN in periodicity modeling tasks and the effectiveness and generalizability of FAN across a range of real-world tasks. Moreover, we reveal that compared to existing Fourier-based networks, FAN accommodates both periodicity modeling and general-purpose modeling well.

cs.LG↗

arXiv2025

NTIRE 2025 Challenge on Low Light Image Enhancement: Methods and Results

This paper presents a comprehensive review of the NTIRE 2025 Low-Light Image Enhancement (LLIE) Challenge, highlighting the proposed solutions and final outcomes. The objective of the challenge is to identify effective networks capable of producing brighter, clearer, and visually compelling images under diverse and challenging conditions. A remarkable total of 762 participants registered for the competition, with 28 teams ultimately submitting valid entries. This paper thoroughly evaluates the state-of-the-art advancements in LLIE, showcasing the significant progress.

cs.CV↗