arXiv subjects

Hui Zheng

Publications and source records attributed to Hui Zheng.

At least 19 recordsLinked to original sources

arXiv2026

From Blueprint to Reality: Modeling and Applying Putnam's Social Capital Theory with LLM-based Multi-agent Simulations

Putnam's Social Capital Theory is a foundational framework for collective action and community prosperity. However, traditional empirical methods face practical limits on control and replication. Meanwhile, LLM-based social simulations are typically behavior-driven and lack theory-aligned environments for modeling Putnam's core propositions. To address these gaps, we introduce SocaSim, an LLM-based multi-agent simulation framework to study Putnam's Social Capital Theory from theoretical blueprint to simulated reality. Specifically, we build an environment integrating social network evolution, trust dynamics, and norm propagation, where agents engage in repeated collective-action experiments, and then apply the three dimensions to analyze adaptation challenges in smart elderly care. Our simulations reproduce Putnam's macro-level patterns and exhibit strong human-agent alignment at the group level. Unlike traditional methods, SocaSim traces micro-level causal pathways of social network, trust, and norms via round-by-round simulations and counterfactual interventions, enabling process-level interpretability. Taken together, these capabilities establish a research paradigm that leverages LLM agents to bridge social science and computer science.

cs.CL↗

arXiv2026

AtomMem: Building Simple and Effective Memory System for LLM Agents via Atomic Facts

Large language models (LLMs) demonstrate strong reasoning and generation abilities, but their fixed context windows limit long-term information accumulation and reuse across multi-session interactions. Existing memory-augmented systems often construct memory in a coarse and unstable manner, relying on inefficient memory representations or unstable unconstrained updates. To address these challenges, we propose AtomMem, a long-term memory system designed for value-dense storage and stable memory evolution. AtomMem introduces a Fact Executor, which selectively extracts high value atomic facts from long form interactions to serve as highly efficient memory representations. Subsequently, AtomMem organizes these facts into hierarchical event structures and temporal profiles, capturing coherent episodic contexts and tracking dynamically evolving user attributes over time. During retrieval, the system activates an associative memory graph to connect fragmented memories. Experiments on the LoCoMo benchmark confirm that AtomMem achieves state-of-the-art performance across various reasoning tasks, offering a scalable and economically viable solution for deploying intelligent personalized agents.

cs.CL↗

arXiv2026

Precise theoretical prediction on branching fractions and polarizations of $D \to V V$ decays

We present a precise and systematic analysis of $D \to V V$ decays within the factorization-assisted topological-amplitude approach, where $D$ denotes the set $\{D^0, \, D^+,\, D^+_s\}$ and $V$ represents the vector mesons $\rho, K^*, \omega$, and $\phi$. Given the limited current experimental data, the factorization-assisted topological-amplitude approach serves as a available phenomenological framework for predicting charmed meson decays to both vector mesons. In this framework, incorporating flavor SU(3) symmetry breaking effects, we can express nonfactorizable contributions of different modes as a minimal set of universal parameters globally fitted to experimental data. Utilizing 36 experimental data points for $D \to VV$ decays, we precisely extract ten nonfactorizable parameters associated with the $C$ and $E$ topological diagrams with $\chi^2/\mathrm{d.o.f.}=8.43$. We find that a large strong phase in the longitude $E$ amplitude cause strong destructive interference with the $C$ longitudinal component, yielding $f_\parallel >f_L $, contrary to the naive factorization predictions. Additionally, for modes processing exclusively by the $E$ diagram, the amplitude hierarchy $|S|<|D|$ leads to a $D$-wave branching fraction larger than that of the $S$-wave. This explains recent observations that contradict $S$-wave dominance predictions. The predicted branching fractions and polarizations for 28 decay modes are consistent with existing experimental data. Unobserved modes, especially those with branching fractions of order $10^{-3}\sim10^{-2}$, the $D$-wave dominated modes, and modes exhibiting $f_\parallel >f_L $, await measurement by BESIII, STCF, Belle II, and LHCb.

hep-ph↗

arXiv2025

Self-supervised Multiplex Consensus Mamba for General Image Fusion

Image fusion integrates complementary information from different modalities to generate high-quality fused images, thereby enhancing downstream tasks such as object detection and semantic segmentation. Unlike task-specific techniques that primarily focus on consolidating inter-modal information, general image fusion needs to address a wide range of tasks while improving performance without increasing complexity. To achieve this, we propose SMC-Mamba, a Self-supervised Multiplex Consensus Mamba framework for general image fusion. Specifically, the Modality-Agnostic Feature Enhancement (MAFE) module preserves fine details through adaptive gating and enhances global representations via spatial-channel and frequency-rotational scanning. The Multiplex Consensus Cross-modal Mamba (MCCM) module enables dynamic collaboration among experts, reaching a consensus to efficiently integrate complementary information from multiple modalities. The cross-modal scanning within MCCM further strengthens feature interactions across modalities, facilitating seamless integration of critical information from both sources. Additionally, we introduce a Bi-level Self-supervised Contrastive Learning Loss (BSCL), which preserves high-frequency information without increasing computational overhead while simultaneously boosting performance in downstream tasks. Extensive experiments demonstrate that our approach outperforms state-of-the-art (SOTA) image fusion algorithms in tasks such as infrared-visible, medical, multi-focus, and multi-exposure fusion, as well as downstream visual tasks.

cs.CV↗

arXiv2025

DynamicVerse: A Physically-Aware Multimodal Framework for 4D World Modeling

Understanding the dynamic physical world, characterized by its evolving 3D structure, real-world motion, and semantic content with textual descriptions, is crucial for human-agent interaction and enables embodied agents to perceive and act within real environments with human-like capabilities. However, existing datasets are often derived from limited simulators or utilize traditional Structurefrom-Motion for up-to-scale annotation and offer limited descriptive captioning, which restricts the capacity of foundation models to accurately interpret real-world dynamics from monocular videos, commonly sourced from the internet. To bridge these gaps, we introduce DynamicVerse, a physical-scale, multimodal 4D world modeling framework for dynamic real-world video. We employ large vision, geometric, and multimodal models to interpret metric-scale static geometry, real-world dynamic motion, instance-level masks, and holistic descriptive captions. By integrating window-based Bundle Adjustment with global optimization, our method converts long real-world video sequences into a comprehensive 4D multimodal format. DynamicVerse delivers a large-scale dataset consisting of 100K+ videos with 800K+ annotated masks and 10M+ frames from internet videos. Experimental evaluations on three benchmark tasks, namely video depth estimation, camera pose estimation, and camera intrinsics estimation, demonstrate that our 4D modeling achieves superior performance in capturing physical-scale measurements with greater global accuracy than existing methods.

cs.CV↗

arXiv2025

Updated branching ratios and $CP$ asymmetries in $D \to PV$ decays

Motivated by extensive new high-precision experimental data, we present an updated analysis of the two-body charm decays $D \to PV$ (with $P =\pi,K, \eta^{(\prime)}$ and $V =\rho, K^*, \omega, \phi$) within the factorization-assisted topological-amplitude (FAT) approach. In the framework, flavor SU(3) symmetry breaking effect is incorporated into the topological amplitudes, allowing the nonfactorizable contributions from topological diagrams to be expressed as a minimal set of universal parameters determined through a global fit to experimental data. Thanks to the sufficient data with high precision, we are now able to quantify the nonfactorizable contribution of so-called ``factorization" $T$ diagram, which is essential for explaining the observed branching ratios. The parameters for the $C$, $E$, and $A$ diagrams are also updated with significant improved precision, and notably, the resulting strong phases differ substantially from those in the earlier FAT analysis. We find that the $C$ topological amplitude features substantial nonfactorizable effects in the charm sector. These refined parameters enable predictions of significantly improved accuracy, yielding branching ratios in good agreement with current data and the latest results in topological diagram approach, and the predicted direct $\mathit{CP}$ asymmetries differ distinctly form previous FAT results due to updated strong phases. In several modes, the $\mathit{CP}$ asymmetries are predicted to reach $\mathcal{O}(10^{-3})$, thereby making them promising observables for future high-precision experiments at LHCb, BESIII and Belle II. Predictions for unobserved decay modes, especially those with branching fractions of order $10^{-4}\sim10^{-3}$, are also provided for forthcoming experimental tests.

hep-ph↗

arXiv2025

TEACH: Text Encoding as Curriculum Hints for Scene Text Recognition

Scene Text Recognition (STR) remains a challenging task due to complex visual appearances and limited semantic priors. We propose TEACH, a novel training paradigm that injects ground-truth text into the model as auxiliary input and progressively reduces its influence during training. By encoding target labels into the embedding space and applying loss-aware masking, TEACH simulates a curriculum learning process that guides the model from label-dependent learning to fully visual recognition. Unlike language model-based approaches, TEACH requires no external pretraining and introduces no inference overhead. It is model-agnostic and can be seamlessly integrated into existing encoder-decoder frameworks. Extensive experiments across multiple public benchmarks show that models trained with TEACH achieve consistently improved accuracy, especially under challenging conditions, validating its robustness and general applicability.

cs.CV↗

arXiv2025

BrainStratify: Coarse-to-Fine Disentanglement of Intracranial Neural Dynamics

Decoding speech directly from neural activity is a central goal in brain-computer interface (BCI) research. In recent years, exciting advances have been made through the growing use of intracranial field potential recordings, such as stereo-ElectroEncephaloGraphy (sEEG) and ElectroCorticoGraphy (ECoG). These neural signals capture rich population-level activity but present key challenges: (i) task-relevant neural signals are sparsely distributed across sEEG electrodes, and (ii) they are often entangled with task-irrelevant neural signals in both sEEG and ECoG. To address these challenges, we introduce a unified Coarse-to-Fine neural disentanglement framework, BrainStratify, which includes (i) identifying functional groups through spatial-context-guided temporal-spatial modeling, and (ii) disentangling distinct neural dynamics within the target functional group using Decoupled Product Quantization (DPQ). We evaluate BrainStratify on two open-source sEEG datasets and one (epidural) ECoG dataset, spanning tasks like vocal production and speech perception. Extensive experiments show that BrainStratify, as a unified framework for decoding speech from intracranial neural signals, significantly outperforms previous decoding methods. Overall, by combining data-driven stratification with neuroscience-inspired modularity, BrainStratify offers a robust and interpretable solution for speech decoding from intracranial recordings.

eess.SP↗

arXiv2025

Pan-LUT: Efficient Pan-sharpening via Learnable Look-Up Tables

Recently, deep learning-based pan-sharpening algorithms have achieved notable advancements over traditional methods. However, deep learning-based methods incur substantial computational overhead during inference, especially with large images. This excessive computational demand limits the applicability of these methods in real-world scenarios, particularly in the absence of dedicated computing devices such as GPUs and TPUs. To address these challenges, we propose Pan-LUT, a novel learnable look-up table (LUT) framework for pan-sharpening that strikes a balance between performance and computational efficiency for large remote sensing images. Our method makes it possible to process 15K*15K remote sensing images on a 24GB GPU. To finely control the spectral transformation, we devise the PAN-guided look-up table (PGLUT) for channel-wise spectral mapping. To effectively capture fine-grained spatial details, we introduce the spatial details look-up table (SDLUT). Furthermore, to adaptively aggregate channel information for generating high-resolution multispectral images, we design an adaptive output look-up table (AOLUT). Our model contains fewer than 700K parameters and processes a 9K*9K image in under 1 ms using one RTX 2080 Ti GPU, demonstrating significantly faster performance compared to other methods. Experiments reveal that Pan-LUT efficiently processes large remote sensing images in a lightweight manner, bridging the gap to real-world applications. Furthermore, our model surpasses SOTA methods in full-resolution scenes under real-world conditions, highlighting its effectiveness and efficiency.

cs.CV↗

arXiv2025

High-Precision Fabric Defect Detection via Adaptive Shape Convolutions and Large Kernel Spatial Modeling

Detecting fabric defects in the textile industry remains a challenging task due to the diverse and complex nature of defect patterns. Traditional methods often suffer from slow inference speeds, limited accuracy, and inadequate recognition rates, particularly in scenarios involving intricate or subtle defects. To overcome these limitations, we introduce Fab-ASLKS, an advanced fabric defect detection framework built upon the YOLOv8s architecture. Fab-ASLKS incorporates two key modules: (1) the Adaptive Shape Convolution Module (ASCM), which leverages adaptive shape convolution within the Neck to enhance feature fusion and improve efficiency by extending the capabilities of the standard C2f structure, and (2) the Large Kernel Shift Convolution Module (LKSCM), designed to emulate large kernel effects within the Backbone, enabling superior spatial information extraction. These modules collaboratively optimize feature extraction and information integration across the network. Extensive experiments conducted on the Tianchi fabric defect detection dataset demonstrate that Fab-ASLKS achieves a 5% improvement in mAP@50 over the baseline, showcasing its capability to deliver high precision and efficiency.

cs.CV↗

arXiv2024

Training-Free Large Model Priors for Multiple-in-One Image Restoration

Image restoration aims to reconstruct the latent clear images from their degraded versions. Despite the notable achievement, existing methods predominantly focus on handling specific degradation types and thus require specialized models, impeding real-world applications in dynamic degradation scenarios. To address this issue, we propose Large Model Driven Image Restoration framework (LMDIR), a novel multiple-in-one image restoration paradigm that leverages the generic priors from large multi-modal language models (MMLMs) and the pretrained diffusion models. In detail, LMDIR integrates three key prior knowledges: 1) global degradation knowledge from MMLMs, 2) scene-aware contextual descriptions generated by MMLMs, and 3) fine-grained high-quality reference images synthesized by diffusion models guided by MMLM descriptions. Standing on above priors, our architecture comprises a query-based prompt encoder, degradation-aware transformer block injecting global degradation knowledge, content-aware transformer block incorporating scene description, and reference-based transformer block incorporating fine-grained image priors. This design facilitates single-stage training paradigm to address various degradations while supporting both automatic and user-guided restoration. Extensive experiments demonstrate that our designed method outperforms state-of-the-art competitors on multiple evaluation benchmarks.

cs.CV↗

arXiv2024

SI-SD: Sleep Interpreter through awake-guided cross-subject Semantic Decoding

Understanding semantic content from brain activity during sleep represents a major goal in neuroscience. While studies in rodents have shown spontaneous neural reactivation of memories during sleep, capturing the semantic content of human sleep poses a significant challenge due to the absence of well-annotated sleep datasets and the substantial differences in neural patterns between wakefulness and sleep. To address these challenges, we designed a novel cognitive neuroscience experiment and collected a comprehensive, well-annotated electroencephalography (EEG) dataset from 134 subjects during both wakefulness and sleep. Leveraging this benchmark dataset, we developed SI-SD that enhances sleep semantic decoding through the position-wise alignment of neural latent sequence between wakefulness and sleep. In the 15-way classification task, our model achieves 24.12% and 21.39% top-1 accuracy on unseen subjects for NREM 2/3 and REM sleep, respectively, surpassing all other baselines. With additional fine-tuning, decoding performance improves to 30.32% and 31.65%, respectively. Besides, inspired by previous neuroscientific findings, we systematically analyze how the "Slow Oscillation" event impacts decoding performance in NREM 2/3 sleep -- decoding performance on unseen subjects further improves to 40.02%. Together, our findings and methodologies contribute to a promising neuro-AI framework for decoding brain activity during sleep.

cs.LG↗

arXiv2024

Du-IN: Discrete units-guided mask modeling for decoding speech from Intracranial Neural signals

Invasive brain-computer interfaces with Electrocorticography (ECoG) have shown promise for high-performance speech decoding in medical applications, but less damaging methods like intracranial stereo-electroencephalography (sEEG) remain underexplored. With rapid advances in representation learning, leveraging abundant recordings to enhance speech decoding is increasingly attractive. However, popular methods often pre-train temporal models based on brain-level tokens, overlooking that brain activities in different regions are highly desynchronized during tasks. Alternatively, they pre-train spatial-temporal models based on channel-level tokens but fail to evaluate them on challenging tasks like speech decoding, which requires intricate processing in specific language-related areas. To address this issue, we collected a well-annotated Chinese word-reading sEEG dataset targeting language-related brain networks from 12 subjects. Using this benchmark, we developed the Du-IN model, which extracts contextual embeddings based on region-level tokens through discrete codex-guided mask modeling. Our model achieves state-of-the-art performance on the 61-word classification task, surpassing all baselines. Model comparisons and ablation studies reveal that our design choices, including (i) temporal modeling based on region-level tokens by utilizing 1D depthwise convolution to fuse channels in the ventral sensorimotor cortex (vSMC) and superior temporal gyrus (STG) and (ii) self-supervision through discrete codex-guided mask modeling, significantly contribute to this performance. Overall, our approach -- inspired by neuroscience findings and capitalizing on region-level representations from specific brain regions -- is suitable for invasive brain modeling and represents a promising neuro-inspired AI approach in brain-computer interfaces.

eess.SP↗

arXiv2024

The ab initio amorphous materials database: Empowering machine learning to decode diffusivity

Amorphous materials exhibit unique properties that make them suitable for various applications in science and technology, ranging from optical and electronic devices and solid-state batteries to protective coatings. However, data-driven exploration and design of amorphous materials is hampered by the absence of a comprehensive database covering a broad chemical space. In this work, we present the largest computed amorphous materials database to date, generated from systematic and accurate \textit{ab initio} molecular dynamics (AIMD) calculations. We also show how the database can be used in simple machine-learning models to connect properties to composition and structure, here specifically targeting ionic conductivity. These models predict the Li-ion diffusivity with speed and accuracy, offering a cost-effective alternative to expensive density functional theory (DFT) calculations. Furthermore, the process of computational quenching amorphous materials provides a unique sampling of out-of-equilibrium structures, energies, and force landscape, and we anticipate that the corresponding trajectories will inform future work in universal machine learning potentials, impacting design beyond that of non-crystalline materials.

cond-mat.mtrl-sci↗

arXiv2024

Rethinking Impersonation and Dodging Attacks on Face Recognition Systems

Face Recognition (FR) systems can be easily deceived by adversarial examples that manipulate benign face images through imperceptible perturbations. Adversarial attacks on FR encompass two types: impersonation (targeted) attacks and dodging (untargeted) attacks. Previous methods often achieve a successful impersonation attack on FR, however, it does not necessarily guarantee a successful dodging attack on FR in the black-box setting. In this paper, our key insight is that the generation of adversarial examples should perform both impersonation and dodging attacks simultaneously. To this end, we propose a novel attack method termed as Adversarial Pruning (Adv-Pruning), to fine-tune existing adversarial examples to enhance their dodging capabilities while preserving their impersonation capabilities. Adv-Pruning consists of Priming, Pruning, and Restoration stages. Concretely, we propose Adversarial Priority Quantification to measure the region-wise priority of original adversarial perturbations, identifying and releasing those with minimal impact on absolute model output variances. Then, Biased Gradient Adaptation is presented to adapt the adversarial examples to traverse the decision boundaries of both the attacker and victim by adding perturbations favoring dodging attacks on the vacated regions, preserving the prioritized features of the original perturbations while boosting dodging performance. As a result, we can maintain the impersonation capabilities of original adversarial examples while effectively enhancing dodging capabilities. Comprehensive experiments demonstrate the superiority of our method compared with state-of-the-art adversarial attack methods.

cs.CV↗

arXiv2022

Efficient photocatalytic nitrogen fixation from air under sunlight via iron-doped WO$_3$

Photocatalytic nitrogen fixation from air directly under sunlight can contribute significantly to carbon neutralization. It is an ideal pathway to replace the industrial Haber Bosch process in future. A Fe-doped layered WO$_3$ photocatalyst containing oxygen vacancies was developed which can fix nitrogen from air directly under sunlight at atmospheric pressure. The iron doping enhances the transport efficiency of photogenerated electrons. The photocatalytic efficiency is around 4 times higher than that of pure WO$_3$. The optimum nitrogen fixation conditions were examined by orthogonal experiments and its nitrogen fixation performance could reach up to 477 $μ\text{g} \cdot \text{g}_{\text{cat}}^{-1} \cdot \text{h}^{-1}$ under sunlight. In addition, the process of nitrogen fixation was detected by situ infrared, which confirmed the reliability of nitrogen fixation. Also, modelling on the interactions between light and the photocatalyst was carried out to study the distribution of surface charge and validate the light absorption of the photocatalyst. This work provides a simple and cheap strategy for photocatalytic nitrogen fixation from air under mild conditions.

cond-mat.mtrl-sci↗

arXiv2022

Efficient sunlight promoted nitrogen fixation from air under room temperature and ambient pressure via Ti/Mo composites

Photocatalytic nitrogen fixation is an important pathway for carbon neutralization and sustainable development. Inspired by nitrogenase, the participation of molybdenum can effectively activate nitrogen. A novel Ti/Mo composites photocatalyst is designed by sintering the molybdenum acetylacetonate precursor with TiO$_{2}$. The special carbon-coated hexagonal photocatalyst is obtained which photocatalytic nitrogen fixation performance is enhanced 16 times compared to pure TiO$_{2}$ at room temperature and ambient pressure. The abundant surface defects in this composite were confirmed to be the key factor for nitrogen fixation. The $^{15}$N$_{2}$ isotope labeling experiment was used to demonstrate the feasibility of nitrogen to ammonia conversion. Also, modelling on the interactions between light and the synthesized photocatalyst particle was examined for the light absorption. The optimum nitrogen fixation conditions have been examined, and the nitrogen fixation performance can reach up to 432 $μ$g$\cdot$g$_{\text{cat}}^{-1}\cdot$h$^{-1}$. Numerical simulations via the field-only surface integral method were also carried out to study the interactions between light and the photocatalytic particles to further confirm that it can be a useful material for photocatalyst. This newly developed Ti/Mo composites provide a simple and effective strategy for photocatalytic nitrogen fixation from air directly under ambient conditions.

cond-mat.mtrl-sci↗

arXiv2022

Multi-scale Investigation of Chemical Short-Range Order and Dislocation Glide in the MoNbTi and TaNbTi Refractory Multi-Principal Element Alloys

Refractory multi-principal element alloys (RMPEAs) are promising materials for high-temperature structural applications. Here, we investigate the role of chemical short-range ordering (CSRO) on dislocation glide in two model RMPEAs - TaNbTi and MoNbTi - using a multi-scale modeling approach. A highly accurate machine learning interatomic potential was developed for the Mo-Ta-Nb-Ti system and used to demonstrate that MoNbTi exhibits a much greater degree of SRO than TaNbTi and the local composition has a direct effect on the unstable stacking fault energies (USFE). From mesoscale phase-field dislocation dynamics simulations, we find that increasing SRO leads to higher mean USFEs, thereby increasing the stress required for dislocation glide. The gliding dislocations experience significant hardening due to pinning and depinning caused by random compositional fluctuations, with higher SRO decreasing the degree of USFE dispersion and hence, amount of hardening. Finally, we show how the morphology of an expanding dislocation loop is affected by the applied stress, with higher SRO requiring higher applied stresses to achieve smooth screw dislocation glide.

cond-mat.mtrl-sci↗