arXiv subjects

Jun Zhang

Publications and source records attributed to Jun Zhang.

At least 307 records · Page 17Linked to original sources

arXiv2025

DeepMEL: A Multi-Agent Collaboration Framework for Multimodal Entity Linking

Multimodal Entity Linking (MEL) aims to associate textual and visual mentions with entities in a multimodal knowledge graph. Despite its importance, current methods face challenges such as incomplete contextual information, coarse cross-modal fusion, and the difficulty of jointly large language models (LLMs) and large visual models (LVMs). To address these issues, we propose DeepMEL, a novel framework based on multi-agent collaborative reasoning, which achieves efficient alignment and disambiguation of textual and visual modalities through a role-specialized division strategy. DeepMEL integrates four specialized agents, namely Modal-Fuser, Candidate-Adapter, Entity-Clozer and Role-Orchestrator, to complete end-to-end cross-modal linking through specialized roles and dynamic coordination. DeepMEL adopts a dual-modal alignment path, and combines the fine-grained text semantics generated by the LLM with the structured image representation extracted by the LVM, significantly narrowing the modal gap. We design an adaptive iteration strategy, combines tool-based retrieval and semantic reasoning capabilities to dynamically optimize the candidate set and balance recall and precision. DeepMEL also unifies MEL tasks into a structured cloze prompt to reduce parsing complexity and enhance semantic comprehension. Extensive experiments on five public benchmark datasets demonstrate that DeepMEL achieves state-of-the-art performance, improving ACC by 1%-57%. Ablation studies verify the effectiveness of all modules.

cs.CL↗

arXiv2025

DreamSwapV: Mask-guided Subject Swapping for Any Customized Video Editing

With the rapid progress of video generation, demand for customized video editing is surging, where subject swapping constitutes a key component yet remains under-explored. Prevailing swapping approaches either specialize in narrow domains--such as human-body animation or hand-object interaction--or rely on some indirect editing paradigm or ambiguous text prompts that compromise final fidelity. In this paper, we propose DreamSwapV, a mask-guided, subject-agnostic, end-to-end framework that swaps any subject in any video for customization with a user-specified mask and reference image. To inject fine-grained guidance, we introduce multiple conditions and a dedicated condition fusion module that integrates them efficiently. In addition, an adaptive mask strategy is designed to accommodate subjects of varying scales and attributes, further improving interactions between the swapped subject and its surrounding context. Through our elaborate two-phase dataset construction and training scheme, our DreamSwapV outperforms existing methods, as validated by comprehensive experiments on VBench indicators and our first introduced DreamSwapV-Benchmark.

cs.CV↗

arXiv2025

Quantum synchronization at the critical point of Floquet driven Rydberg atoms

The criticality enhanced correlations and susceptibility allow weak periodic driving to induce collective synchronization due to critical slowing down, providing a unique platform to study non-equilibrium order emergence. This establishes a powerful paradigm for investigating non-equilibrium order formation, yet the fundamental mechanisms of critical-point synchronization remain poorly understood. Here, we utilize a microwave pulse as a seed to induce quantum synchronization at the critical point of Floquet driven Rydberg atoms. In the experiment, the microwave periodic driving on Rydberg states acts as a seeded temporal order in subspace, which triggers synchronization across the entire ensemble. The behavior of the emergent synchronized oscillation is elaborately linked to alterations in the seed, such as the relative phase shift and the frequency difference, which result in phase-dependent seeding and embryonic synchronization. This result opens up new possibilities for studying and harnessing time-dependent quantum many-body phenomena, offering insight into the behavior of complex many-body systems under seeding.

cond-mat.quant-gas↗

arXiv2025

ProtTeX-CC: Activating In-Context Learning in Protein LLM via Two-Stage Instruction Compression

Recent advances in protein large language models, such as ProtTeX, represent both side-chain amino acids and backbone structure as discrete token sequences of residue length. While this design enables unified modeling of multimodal protein information, it suffers from two major limitations: (1) The concatenation of sequence and structure tokens approximately doubles the protein length and breaks the intrinsic residue-level alignment between modalities. (2) Constrained by the training corpus and limited context window, ProtTeX is typically trained on single-protein inputs, rendering it incompatible with in-context learning (ICL) and thus limiting its generalization capability. To address these issues, we propose ProtTeX-CC, a lightweight two-stage compression framework designed to enhance ProtTeX under few-shot settings. We first design a joint embedding compression mechanism that fuses sequence and structure representations at the residue level, effectively reducing the protein input length by half without sacrificing performance. Then we propose a self-compression module that aggregates each full demonstration into the latent space of the last few linguistic tokens, reducing the average demonstration length from 751 tokens to less than 16 tokens. Compared to the original ProtTeX, our self-compression approach achieves a compression ratio of approximately 93.68% in the total prompt length under the 16-shot setting. Without modifying the backbone model, ProtTeX-CC introduces only a small number of additional parameters through PEFT-based tuning in the joint embedding compression stage and a single trainable projection layer in the self-compression stage. Extensive experiments on protein function prediction show that ProtTeX-CC improves performance on the in-domain benchmark by 2%, and generalizes well to the out-of-domain dataset with a performance gain of 11%.

cs.LG↗

arXiv2025

Efficient and fair trading mechanisms for resource exchange in market design

We develop a method using parameterized linear equations to define trading mechanisms in market design models. Our method adeptly addresses challenges arising from factors such as complex endowments or coarse priorities, while offering flexibility to incorporate fairness concerns through the selection of equation parameters. Applying this method to models including fractional endowment exchange, priority-based allocation, and house allocation with existing tenants, we obtain new mechanisms that are both efficient and fair in these models.

econ.TH↗

arXiv2025

Efficient and fair trading mechanisms on the full preference domain

Yu and Zhang (2025) introduce a new method for defining trading mechanisms in market design and apply it to develop new mechanisms that achieve efficiency and fairness in various models. However, their assumption of strict preferences limits its applicability. This paper extends their approach to the full preference domain, allowing for indifferences while preserving key efficiency and fairness properties. As a corollary, we generalize the probabilistic serial mechanism in the house allocation model to the full preference domain without relying on operations research tools.

econ.TH↗

arXiv2025

The core in the housing market model with fractional endowments

We explore the core concept in a generalization of the housing market model where agents own fractional endowments while maintaining ordinal preferences. Recognizing that individuals are easier than coalitions to block an allocation, we adopt a definition in which individuals block an allocation if their received assignments do not first-order stochastically dominate their endowment, while a non-singleton coalition blocks an allocation if they can reallocate their endowments to obtain new assignments that first-order stochastically dominate their original assignments. Our findings show that, unlike the original model, the strong core may be empty, while the weak core is nonempty. The weak core always contains elements that satisfy equal treatment of equals, but it may not contain elements satisfying equal-endowment no envy.

econ.TH↗

arXiv2025

Representation Quantization for Collaborative Filtering Augmentation

As the core algorithm in recommendation systems, collaborative filtering (CF) algorithms inevitably face the problem of data sparsity. Since CF captures similar users and items for recommendations, it is effective to augment the lacking user-user and item-item homogeneous linkages. However, existing methods are typically limited to connecting through overlapping interacted neighbors or through similar attributes and contents. These approaches are constrained by coarse-grained, sparse attributes and fail to effectively extract behavioral characteristics jointly from interaction sequences and attributes. To address these challenges, we propose a novel two-stage collaborative recommendation algorithm, DQRec: Decomposition-based Quantized Variational AutoEncoder (DQ-VAE) for Recommendation. DQRec augments features and homogeneous linkages by extracting the behavior characteristics jointly from interaction sequences and attributes, namely patterns, such as user multi-aspect interests. Inspired by vector quantization (VQ) technology, we propose a new VQ algorithm, DQ-VAE, which decomposes the pre-trained representation embeddings into distinct dimensions, and quantize them to generates semantic IDs. We utilize the generated semantic IDs as the extracted patterns mentioned above. By integrating these semantic ID patterns into the recommendation process through feature and linkage augmentation, the system enriches both latent and explicit user and item features, identifies pattern-similar neighbors, and thereby improves the efficiency of information diffusion. Experimental comparisons with baselines across multiple datasets demonstrate the superior performance of the proposed DQRec method.

cs.IR↗

arXiv2025

Yan: Foundational Interactive Video Generation

We present Yan, a foundational framework for interactive video generation, covering the entire pipeline from simulation and generation to editing. Specifically, Yan comprises three core modules. AAA-level Simulation: We design a highly-compressed, low-latency 3D-VAE coupled with a KV-cache-based shift-window denoising inference process, achieving real-time 1080P/60FPS interactive simulation. Multi-Modal Generation: We introduce a hierarchical autoregressive caption method that injects game-specific knowledge into open-domain multi-modal video diffusion models (VDMs), then transforming the VDM into a frame-wise, action-controllable, real-time infinite interactive video generator. Notably, when the textual and visual prompts are sourced from different domains, the model demonstrates strong generalization, allowing it to blend and compose the style and mechanics across domains flexibly according to user prompts. Multi-Granularity Editing: We propose a hybrid model that explicitly disentangles interactive mechanics simulation from visual rendering, enabling multi-granularity video content editing during interaction through text. Collectively, Yan offers an integration of these modules, pushing interactive video generation beyond isolated capabilities toward a comprehensive AI-driven interactive creation paradigm, paving the way for the next generation of creative tools, media, and entertainment. The project page is: https://greatx3.github.io/Yan/.

cs.CV↗

arXiv2025

Anomalous Transport of Elongated Particles in Oscillatory Vortical Flows

We investigate the transport dynamics of elongated particles in cellular vortical flows that undergo spatial oscillations over time. Experimental flow visualizations reveal mixed flow fields with chaotic and elliptic regions coexisting. Surprisingly, the particle transport rate does not increase monotonically with particle length, even though longer particles are expected to explore neighboring vortices more easily. Numerical simulations in a much larger system produce similar transport anomalies, characterized by subdiffusion due to frequent long-time trapping in vortices at certain lengths, but normal diffusion at others. At moderate oscillation frequencies, these long-time trapping events occur within the chaotic region; at high frequencies, they occur in the elliptic regions, but only for particles whose lengths match these regions. In the latter case, subdiffusion is robust against random noise. Our results reveal new mechanisms for controlling particle diffusion in fluid flows.

physics.flu-dyn↗

arXiv2025

Exploring Large Language Model Agents for Piloting Social Experiments

Computational social experiments, which typically employ agent-based modeling to create testbeds for piloting social experiments, not only provide a computational solution to the major challenges faced by traditional experimental methods, but have also gained widespread attention in various research fields. Despite their significance, their broader impact is largely limited by the underdeveloped intelligence of their core component, i.e., agents. To address this limitation, we develop a framework grounded in well-established social science theories and practices, consisting of three key elements: (i) large language model (LLM)-driven experimental agents, serving as "silicon participants", (ii) methods for implementing various interventions or treatments, and (iii) tools for collecting behavioral, survey, and interview data. We evaluate its effectiveness by replicating three representative experiments, with results demonstrating strong alignment, both quantitatively and qualitatively, with real-world evidence. This work provides the first framework for designing LLM-driven agents to pilot social experiments, underscoring the transformative potential of LLMs and their agents in computational social science

cs.CY↗

arXiv2025

Traction and Stress Control Formation and Motion of +1/2 Topological Defects in Epithelial Cell Monolayers

In confluent cell monolayers, patterns of cell forces and motion are systematically altered near topological defects in cell shape. In turn, defects have been proposed to alter cell density, extrusion, and invasion, but it remains unclear how the defects form and how they affect cell forces and motion. Here, we studied +1/2 defects, and, in contrast to prior studies, we observed the concurrent occurrence of both tail-to-head and head-to-tail defect motion in the same cell monolayer. We quantified the cell velocities, the tractions at the cell-substrate interface, and the stresses within the cell layer near +1/2 defects. Results revealed that both traction and stress are sources of activity and dissipation within the epithelial cell monolayer, with the direction of motion of +1/2 defects depending on whether energy is injected by stresses or tractions. Interestingly, patterns of motion, traction, stress, and energy injection near +1/2 defects existed before defect formation, suggesting that defects form as a result of spatially coordinated patterns in cell forces and motion. These findings introduce a new focus, on coordinated patterns of force and motion that lead to defect formation and motion.

physics.bio-ph↗

arXiv2025

p-MoD: Building Mixture-of-Depths MLLMs via Progressive Ratio Decay

Despite the remarkable performance of multimodal large language models (MLLMs) across diverse tasks, the substantial training and inference costs impede their advancement. In this paper, we propose p-MoD, an efficient MLLM architecture that significantly reduces training and inference costs while maintaining model performance. The majority of computation in MLLMs stems from the overwhelming volume of vision tokens processed by the transformer-based LLM. Accordingly, we leverage the Mixture-of-Depths (MoD) mechanism, where each LLM layer selects essential vision tokens to process while skipping redundant ones. However, integrating MoD into MLLMs is non-trivial. To address the challenges of training and inference stability as well as limited training data, we adapt the MoD module with two novel designs: tanh-gated weight normalization (TanhNorm) and symmetric token reweighting (STRing). Moreover, we observe that vision tokens exhibit higher redundancy in deeper layers and thus design a progressive ratio decay (PRD) strategy, which gradually reduces the token retention ratio layer by layer, employing a shifted cosine schedule. This crucial design fully unleashes the potential of MoD, significantly boosting the efficiency and performance of our models. Extensive experiments on two baseline models across 15 benchmarks show that our model matches or even surpasses the performance of corresponding baselines, while requiring only 55.6% TFLOPs and 53.7% KV cache storage during inference, and 77.7% GPU hours during training.

cs.CV↗

arXiv2025

FinMMR: Make Financial Numerical Reasoning More Multimodal, Comprehensive, and Challenging

We present FinMMR, a novel bilingual multimodal benchmark tailored to evaluate the reasoning capabilities of multimodal large language models (MLLMs) in financial numerical reasoning tasks. Compared to existing benchmarks, our work introduces three significant advancements. (1) Multimodality: We meticulously transform existing financial reasoning benchmarks, and construct novel questions from the latest Chinese financial research reports. FinMMR comprises 4.3K questions and 8.7K images spanning 14 categories, including tables, bar charts, and ownership structure charts. (2) Comprehensiveness: FinMMR encompasses 14 financial subdomains, including corporate finance, banking, and industry analysis, significantly exceeding existing benchmarks in financial domain knowledge breadth. (3) Challenge: Models are required to perform multi-step precise numerical reasoning by integrating financial knowledge with the understanding of complex financial images and text. The best-performing MLLM achieves only 53.0% accuracy on Hard problems. We believe that FinMMR will drive advancements in enhancing the reasoning capabilities of MLLMs in real-world scenarios.

cs.CV↗

arXiv2025

Vector-valued Reproducing Kernel Banach Spaces with Group Lasso Norms

Focusing on establishing a mathematical basis for kernel methods in sparse multi-task learning, we explore the theory of vector-valued reproducing kernel Banach spaces (RKBSs) endowed with $\ell_{p,1}$-norms ($1\le p\le +\infty$), encompassing both the sparse learning case when $p=1$ and the group lasso when $p=2$. We develop RKBSs equipped with these group lasso norms that support the linear representer theorem for regularized learning frameworks. Additionally, we introduce reproducing kernels admissible for this construction. Such reproducing kernels are applicable to sparse multi-task learning with group lasso norms.

math.FA↗

arXiv2025

PHM-Bench: A Domain-Specific Benchmarking Framework for Systematic Evaluation of Large Models in Prognostics and Health Management

With the rapid advancement of generative artificial intelligence, large language models (LLMs) are increasingly adopted in industrial domains, offering new opportunities for Prognostics and Health Management (PHM). These models help address challenges such as high development costs, long deployment cycles, and limited generalizability. However, despite the growing synergy between PHM and LLMs, existing evaluation methodologies often fall short in structural completeness, dimensional comprehensiveness, and evaluation granularity. This hampers the in-depth integration of LLMs into the PHM domain. To address these limitations, this study proposes PHM-Bench, a novel three-dimensional evaluation framework for PHM-oriented large models. Grounded in the triadic structure of fundamental capability, core task, and entire lifecycle, PHM-Bench is tailored to the unique demands of PHM system engineering. It defines multi-level evaluation metrics spanning knowledge comprehension, algorithmic generation, and task optimization. These metrics align with typical PHM tasks, including condition monitoring, fault diagnosis, RUL prediction, and maintenance decision-making. Utilizing both curated case sets and publicly available industrial datasets, our study enables multi-dimensional evaluation of general-purpose and domain-specific models across diverse PHM tasks. PHM-Bench establishes a methodological foundation for large-scale assessment of LLMs in PHM and offers a critical benchmark to guide the transition from general-purpose to PHM-specialized models.

cs.AI↗

arXiv2025

Towed Movable Antenna (ToMA) Array for Ultra Secure Airborne Communications

This paper proposes a novel towed movable antenna (ToMA) array architecture to enhance the physical layer security of airborne communication systems. Unlike conventional onboard arrays with fixed-position antennas (FPAs), the ToMA array employs multiple subarrays mounted on flexible cables and towed by distributed drones, enabling agile deployment in three-dimensional (3D) space surrounding the central aircraft. This design significantly enlarges the effective array aperture and allows dynamic geometry reconfiguration, offering superior spatial resolution and beamforming flexibility. We consider a secure transmission scenario where an airborne transmitter communicates with multiple legitimate users in the presence of potential eavesdroppers. To ensure security, zero-forcing beamforming is employed to nullify signal leakage toward eavesdroppers. Based on the statistical distributions of locations of users and eavesdroppers, the antenna position vector (APV) of the ToMA array is optimized to maximize the users' ergodic achievable rate. Analytical results for the case of a single user and a single eavesdropper reveal the optimal APV structure that minimizes their channel correlation. For the general multiuser scenario, we develop a low-complexity alternating optimization algorithm by leveraging Riemannian manifold optimization. Simulation results confirm that the proposed ToMA array achieves significant performance gains over conventional onboard FPA arrays, especially in scenarios where eavesdroppers are closely located to users under line-of-sight (LoS)-dominant channels.

cs.IT↗

arXiv2025

Dependence of halo properties on central-satellite magnitude gaps through weak lensing measurements

The magnitude gap between the central and satellite galaxies encodes information about the mass accretion history of a dark matter halo, and serves as a useful observational probe for the mass distribution in a halo. In this work, we perform the first weak lensing test of the connections between the magnitude gap and the halo profile. We measure the halo profiles of isolated central galaxies (ICGs) selected primarily from the SDSS Main Galaxy Sample. Halo mass and concentration are inferred by fitting stacked lensing profiles in bins of central luminosity, $L_\mathrm{c}$, and the central-satellite magnitude gap, $L_\mathrm{gap}$. We detect dependence on the magnitude gap in both halo properties. The dependence is the strongest in the ICG luminosity range of $10^{10.3}<L_\mathrm{c}[h^{-2}L_\odot]\leq 10^{10.7}$, where halos with smaller gaps have higher masses and lower concentrations. When $10^{10.7} <L_c[h^{-2}L_\odot] \leq 10^{11.1}$, however, no significant gap dependence is detected. In the range of $10^{9.9}<L_\mathrm{c}[h^{-2}L_\odot] \leq 10^{10.3}$, a disordering of the gap dependence is marginally observable. We compare the observational results with predictions by two lightcone catalogs built from the Illustris TNG300 and the Millennium simulations. The gap dependence in the two mock samples show overall consistency with observations, but neither matches them in all $L_\mathrm{c}$ bins to a quantitative level. We also compare the significance of the gap dependence on halo mass and concentration and find that our measurement prefers gap dependence in both parameters, while the halo mass dependence is preferred over the concentration if only one of the two dependencies is allowed.

astro-ph.CO↗