Search arXiv⌕ Search

arXiv subjects

Cong Wang

Publications and source records attributed to Cong Wang.

At least 217 records · Page 12Linked to original sources
arXiv2025

Space-time analyticity and refined analyticity radius of the Navier-Stokes equations in the critical Besov spaces

In this paper, we establish the space-time analyticity of global solutions to the incompressible Navier-Stokes equations with small initial data in critical \emph{Besov} spaces $\dot B^{3/p-1}_{p,q}$. Time decay rates of higher order space-time joint derivatives and instantaneous lower bounds of the analyticity radius follow as straightforward consequences. The method employed combines Gevrey-class estimates with iterative derivative techniques. Furthermore, we obtain a logarithmic improvement in the lower bound for the spatial analyticity radius of solutions for arbitrary initial data in critical \emph{Besov} spaces.

math.AP↗
arXiv2025

FPGA-Accelerated SpeckleNN with SNL for Real-time X-ray Single-Particle Imaging

We implement a specialized version of our SpeckleNN model for real-time speckle pattern classification in X-ray Single-Particle Imaging (SPI) using the SLAC Neural Network Library (SNL) on an FPGA. This hardware is optimized for inference near detectors in high-throughput X-ray free-electron laser (XFEL) facilities like the Linac Coherent Light Source (LCLS). To fit FPGA constraints, we optimized SpeckleNN, reducing parameters from 5.6M to 64.6K (98.8% reduction) with 90% accuracy. We also compressed the latent space from 128 to 50 dimensions. Deployed on a KCU1500 FPGA, the model used 71% of DSPs, 75% of LUTs, and 48% of FFs, with an average power consumption of 9.4W. The FPGA achieved 45.015us inference latency at 200 MHz. On an NVIDIA A100 GPU, the same inference consumed ~73W and had a 400us latency. Our FPGA version achieved an 8.9x speedup and 7.8x power reduction over the GPU. Key advancements include model specialization and dynamic weight loading through SNL, eliminating time-consuming FPGA re-synthesis for fast, continuous deployment of (re)trained models. These innovations enable real-time adaptive classification and efficient speckle pattern vetoing, making SpeckleNN ideal for XFEL facilities. This implementation accelerates SPI experiments and enhances adaptability to evolving conditions.

physics.ins-det↗
arXiv2025

Binary Neural Networks for Large Language Model: A Survey

Large language models (LLMs) have wide applications in the field of natural language processing(NLP), such as GPT-4 and Llama. However, with the exponential growth of model parameter sizes, LLMs bring significant resource overheads. Low-bit quantization, as a key technique, reduces memory usage and computational demands by decreasing the bit-width of model parameters, activations, and gradients. Previous quantization methods for LLMs have largely employed Post-Training Quantization (PTQ) and Quantization-Aware Training (QAT). PTQ does not require any retraining of the original model, while QAT involves optimizing precision during training to achieve the best quantization parameters. The BitNet team proposed a radically different approach, where quantization is performed from the start of model training, utilizing low-precision binary weights during the training process. This approach has led to the emergence of many binary quantization techniques for large language models. This paper provides a comprehensive review of these binary quantization techniques. Specifically, we will introduce binary quantization techniques in deep neural networks and further explore their application to LLMs, reviewing their various contributions, implementations, and applications.

cs.CL↗
arXiv2025

Ensembling Diffusion Models via Adaptive Feature Aggregation

The success of the text-guided diffusion model has inspired the development and release of numerous powerful diffusion models within the open-source community. These models are typically fine-tuned on various expert datasets, showcasing diverse denoising capabilities. Leveraging multiple high-quality models to produce stronger generation ability is valuable, but has not been extensively studied. Existing methods primarily adopt parameter merging strategies to produce a new static model. However, they overlook the fact that the divergent denoising capabilities of the models may dynamically change across different states, such as when experiencing different prompts, initial noises, denoising steps, and spatial locations. In this paper, we propose a novel ensembling method, Adaptive Feature Aggregation (AFA), which dynamically adjusts the contributions of multiple models at the feature level according to various states (i.e., prompts, initial noises, denoising steps, and spatial locations), thereby keeping the advantages of multiple diffusion models, while suppressing their disadvantages. Specifically, we design a lightweight Spatial-Aware Block-Wise (SABW) feature aggregator that adaptive aggregates the block-wise intermediate features from multiple U-Net denoisers into a unified one. The core idea lies in dynamically producing an individual attention map for each model's features by comprehensively considering various states. It is worth noting that only SABW is trainable with about 50 million parameters, while other models are frozen. Both the quantitative and qualitative experiments demonstrate the effectiveness of our proposed Adaptive Feature Aggregation method.

cs.CV↗
arXiv2025

FedMobile: Enabling Knowledge Contribution-aware Multi-modal Federated Learning with Incomplete Modalities

The Web of Things (WoT) enhances interoperability across web-based and ubiquitous computing platforms while complementing existing IoT standards. The multimodal Federated Learning (FL) paradigm has been introduced to enhance WoT by enabling the fusion of multi-source mobile sensing data while preserving privacy. However, a key challenge in mobile sensing systems using multimodal FL is modality incompleteness, where some modalities may be unavailable or only partially captured, potentially degrading the system's performance and reliability. Current multimodal FL frameworks typically train multiple unimodal FL subsystems or apply interpolation techniques on the node side to approximate missing modalities. However, these approaches overlook the shared latent feature space among incomplete modalities across different nodes and fail to discriminate against low-quality nodes. To address this gap, we present FedMobile, a new knowledge contribution-aware multimodal FL framework designed for robust learning despite missing modalities. FedMobile prioritizes local-to-global knowledge transfer, leveraging cross-node multimodal feature information to reconstruct missing features. It also enhances system performance and resilience to modality heterogeneity through rigorous node contribution assessments and knowledge contribution-aware aggregation rules. Empirical evaluations on five widely recognized multimodal benchmark datasets demonstrate that FedMobile maintains robust learning even when up to 90% of modality information is missing or when data from two modalities are randomly missing, outperforming state-of-the-art baselines.

cs.LG↗
arXiv2025

Artificially creating emergent interfacial antiferromagnetism and its manipulation in a magnetic van-der-Waals heterostructure

Van der Waals (vdW) magnets, with their two-dimensional (2D) atomic structures, provide a unique platform for exploring magnetism at the nanoscale. Although there have been numerous reports on their diverse quantum properties, the emergent interfacial magnetism--artificially created at the interface between two layered magnets--remains largely unexplored. This work presents observations of such emergent interfacial magnetism at the ferromagnet/antiferromagnet interface in a vdW heterostructure. We report the discovery of an intermediate Hall resistance plateau in the anomalous Hall loop, indicative of emergent interfacial antiferromagnetism fostered by the heterointerface. This plateau can be stabilized and further manipulated under varying pressures but collapses under high pressures over 10 GPa. Our theoretical calculations reveal that charge transfer at the interface is pivotal in establishing the interlayer antiferromagnetic spin-exchange interaction. This work illuminates the previously unexplored emergent interfacial magnetism at a vdW interface comprised of a ferromagnetic metal and an antiferromagnetic insulator, and highlights its gradual evolution under increasing pressure. These findings enrich the portfolio of emergent interfacial magnetism and support further investigations on vdW magnetic interfaces and the development of next-generation spintronic devices.

cond-mat.mtrl-sci↗
arXiv2025

LiveVal: Time-aware Data Valuation via Adaptive Reference Points

Time-aware data valuation enhances training efficiency and model robustness, as early detection of harmful samples could prevent months of wasted computation. However, existing methods rely on model retraining or convergence assumptions or fail to capture long-term training dynamics. We propose LiveVal, an efficient time-aware data valuation method with three key designs: 1) seamless integration with SGD training for efficient data contribution monitoring; 2) reference-based valuation with normalization for reliable benchmark establishment; and 3) adaptive reference point selection for real-time updating with optimized memory usage. We establish theoretical guarantees for LiveVal's stability and prove that its valuations are bounded and directionally aligned with optimization progress. Extensive experiments demonstrate that LiveVal provides efficient data valuation across different modalities and model scales, achieving 180 speedup over traditional methods while maintaining robust detection performance.

cs.LG↗
arXiv2025

Long-Term TalkingFace Generation via Motion-Prior Conditional Diffusion Model

Recent advances in conditional diffusion models have shown promise for generating realistic TalkingFace videos, yet challenges persist in achieving consistent head movement, synchronized facial expressions, and accurate lip synchronization over extended generations. To address these, we introduce the \textbf{M}otion-priors \textbf{C}onditional \textbf{D}iffusion \textbf{M}odel (\textbf{MCDM}), which utilizes both archived and current clip motion priors to enhance motion prediction and ensure temporal consistency. The model consists of three key elements: (1) an archived-clip motion-prior that incorporates historical frames and a reference frame to preserve identity and context; (2) a present-clip motion-prior diffusion model that captures multimodal causality for accurate predictions of head movements, lip sync, and expressions; and (3) a memory-efficient temporal attention mechanism that mitigates error accumulation by dynamically storing and updating motion features. We also release the \textbf{TalkingFace-Wild} dataset, a multilingual collection of over 200 hours of footage across 10 languages. Experimental results demonstrate the effectiveness of MCDM in maintaining identity and motion continuity for long-term TalkingFace generation. Code, models, and datasets will be publicly available.

cs.CV↗
arXiv2025

JBShield: Defending Large Language Models from Jailbreak Attacks through Activated Concept Analysis and Manipulation

Despite the implementation of safety alignment strategies, large language models (LLMs) remain vulnerable to jailbreak attacks, which undermine these safety guardrails and pose significant security threats. Some defenses have been proposed to detect or mitigate jailbreaks, but they are unable to withstand the test of time due to an insufficient understanding of jailbreak mechanisms. In this work, we investigate the mechanisms behind jailbreaks based on the Linear Representation Hypothesis (LRH), which states that neural networks encode high-level concepts as subspaces in their hidden representations. We define the toxic semantics in harmful and jailbreak prompts as toxic concepts and describe the semantics in jailbreak prompts that manipulate LLMs to comply with unsafe requests as jailbreak concepts. Through concept extraction and analysis, we reveal that LLMs can recognize the toxic concepts in both harmful and jailbreak prompts. However, unlike harmful prompts, jailbreak prompts activate the jailbreak concepts and alter the LLM output from rejection to compliance. Building on our analysis, we propose a comprehensive jailbreak defense framework, JBShield, consisting of two key components: jailbreak detection JBShield-D and mitigation JBShield-M. JBShield-D identifies jailbreak prompts by determining whether the input activates both toxic and jailbreak concepts. When a jailbreak prompt is detected, JBShield-M adjusts the hidden representations of the target LLM by enhancing the toxic concept and weakening the jailbreak concept, ensuring LLMs produce safe content. Extensive experiments demonstrate the superior performance of JBShield, achieving an average detection accuracy of 0.95 and reducing the average attack success rate of various jailbreak attacks to 2% from 61% across distinct LLMs.

cs.CR↗
arXiv2025

SpikingSoft: A Spiking Neuron Controller for Bio-inspired Locomotion with Soft Snake Robots

Inspired by the dynamic coupling of moto-neurons and physical elasticity in animals, this work explores the possibility of generating locomotion gaits by utilizing physical oscillations in a soft snake by means of a low-level spiking neural mechanism. To achieve this goal, we introduce the Double Threshold Spiking neuron model with adjustable thresholds to generate varied output patterns. This neuron model can excite the natural dynamics of soft robotic snakes, and it enables distinct movements, such as turning or moving forward, by simply altering the neural thresholds. Finally, we demonstrate that our approach, termed SpikingSoft, naturally pairs and integrates with reinforcement learning. The high-level agent only needs to adjust the two thresholds to generate complex movement patterns, thus strongly simplifying the learning of reactive locomotion. Simulation results demonstrate that the proposed architecture significantly enhances the performance of the soft snake robot, enabling it to achieve target objectives with a 21.6% increase in success rate, a 29% reduction in time to reach the target, and smoother movements compared to the vanilla reinforcement learning controllers or Central Pattern Generator controller acting in torque space.

cs.RO↗
arXiv2025

High-Throughput Discovery of Kagome Materials in Transition Metal Oxide Monolayers

Kagome materials are known for hosting exotic quantum states, including quantum spin liquids, charge density waves, and unconventional superconductivity. The search for kagome monolayers is driven by their ability to exhibit neat and well-defined kagome bands near the Fermi level, which are more easily realized in the absence of interlayer interactions. However, this absence also destabilizes the monolayer forms of many bulk kagome materials, posing significant challenges to their discovery. In this work, we propose a strategy to address this challenge by utilizing oxygen vacancies in transition metal oxides within a "1+3" design framework. Through high-throughput computational screening of 349 candidate materials, we identified 12 thermodynamically stable kagome monolayers with diverse electronic and magnetic properties. These materials were classified into three categories based on their lattice geometry, symmetry, band gaps, and magnetic configurations. Detailed analysis of three representative monolayers revealed kagome band features near their Fermi levels, with orbital contributions varying between oxygen 2p and transition metal d states. This study demonstrates the feasibility of the "1+3" strategy, offering a promising approach to uncovering low-dimensional kagome materials and advancing the exploration of their quantum phenomena.

cond-mat.mes-hall↗
arXiv2025

Two-dimensional Kagome Materials: Theoretical Insights, Experimental Realizations, and Electronic Structures

In recent years, kagome materials have attracted significant attention due to their rich emergent phenomena arising from the quantum interplay of geometry, topology, spin, and correlations. However, in the search for kagome materials, it has been found that bulk compounds with electronic properties related to the kagome lattice are relatively scarce, primarily due to the hybridization of kagome layers with adjacent layers. Therefore, researchers have shown increasing interest in the discovery and construction of two-dimensional (2D) kagome materials, aiming to achieve clean kagome bands near the Fermi level in monolayer or few-layer systems. Substantial advancements have already been made in this area. In this review, we summarize the current progress in the construction and development of 2D kagome materials. We begin by introducing the geometric and electronic structures of the kagome lattice model and its variants, followed by discussions on the experimental realizations and electronic structure characterizations of 2D kagome materials. Finally, we provide an outlook on the future developments of 2D kagome materials.

cond-mat.mtrl-sci↗
arXiv2025

CALM: Curiosity-Driven Auditing for Large Language Models

Auditing Large Language Models (LLMs) is a crucial and challenging task. In this study, we focus on auditing black-box LLMs without access to their parameters, only to the provided service. We treat this type of auditing as a black-box optimization problem where the goal is to automatically uncover input-output pairs of the target LLMs that exhibit illegal, immoral, or unsafe behaviors. For instance, we may seek a non-toxic input that the target LLM responds to with a toxic output or an input that induces the hallucinative response from the target LLM containing politically sensitive individuals. This black-box optimization is challenging due to the scarcity of feasible points, the discrete nature of the prompt space, and the large search space. To address these challenges, we propose Curiosity-Driven Auditing for Large Language Models (CALM), which uses intrinsically motivated reinforcement learning to finetune an LLM as the auditor agent to uncover potential harmful and biased input-output pairs of the target LLM. CALM successfully identifies derogatory completions involving celebrities and uncovers inputs that elicit specific names under the black-box setting. This work offers a promising direction for auditing black-box LLMs. Our code is available at https://github.com/x-zheng16/CALM.git.

cs.AI↗
arXiv2024

A Light-Weight Framework for Open-Set Object Detection with Decoupled Feature Alignment in Joint Space

Open-set object detection (OSOD) is highly desirable for robotic manipulation in unstructured environments. However, existing OSOD methods often fail to meet the requirements of robotic applications due to their high computational burden and complex deployment. To address this issue, this paper proposes a light-weight framework called Decoupled OSOD (DOSOD), which is a practical and highly efficient solution to support real-time OSOD tasks in robotic systems. Specifically, DOSOD builds upon the YOLO-World pipeline by integrating a vision-language model (VLM) with a detector. A Multilayer Perceptron (MLP) adaptor is developed to transform text embeddings extracted by the VLM into a joint space, within which the detector learns the region representations of class-agnostic proposals. Cross-modality features are directly aligned in the joint space, avoiding the complex feature interactions and thereby improving computational efficiency. DOSOD operates like a traditional closed-set detector during the testing phase, effectively bridging the gap between closed-set and open-set detection. Compared to the baseline YOLO-World, the proposed DOSOD significantly enhances real-time performance while maintaining comparable accuracy. The slight DOSOD-S model achieves a Fixed AP of $26.7\%$, compared to $26.2\%$ for YOLO-World-v1-S and $22.7\%$ for YOLO-World-v2-S, using similar backbones on the LVIS minival dataset. Meanwhile, the FPS of DOSOD-S is $57.1\%$ higher than YOLO-World-v1-S and $29.6\%$ higher than YOLO-World-v2-S. Meanwhile, we demonstrate that the DOSOD model facilitates the deployment of edge devices. The codes and models are publicly available at https://github.com/D-Robotics-AI-Lab/DOSOD.

cs.CV↗
arXiv2024

Taming Imbalance and Complexity in WAN Traffic Engineering

The rapid expansion of global cloud infrastructures, coupled with the growing volume and complexity of network traffic, has fueled active research into scalable and resilient Traffic Engineering (TE) solutions for Wide Area Networks (WANs). Despite recent advancements, achieving an optimal balance between solution quality and computational complexity remains a significant challenge, especially for larger WAN topologies under dynamic traffic demands and stringent resource constraints. This paper presents empirical evidence of a critical shortcoming in existing TE solutions: their oversight inadequately accounting for traffic demand heterogeneities and link utilization imbalances. We identify key factors contributing to these issues, including traffic distribution, solver selection, resiliency, and resource overprovisioning. To address these gaps, we propose a holistic solution featuring new performance metrics and a novel resilient TE algorithm. The proposed metrics, critical link set and network criticality, provide a more comprehensive assessment of resilient TE solutions, while the tunnel-based TE algorithm dynamically adapts to changing traffic demands. Through extensive simulations on diverse WAN topologies, we demonstrate that this holistic solution significantly improves network performance, achieving a superior balance across key objectives. This work represents a significant advancement in the development of resilient and scalable TE solutions for WANs.

cs.NI↗
arXiv2024

Non-intrusive and Unconstrained Keystroke Inference in VR Platforms via Infrared Side Channel

Virtual Reality (VR) technologies are increasingly employed in numerous applications across various areas. Therefore, it is essential to ensure the security of interactions between users and VR devices. In this paper, we disclose a new side-channel leakage in the constellation tracking system of mainstream VR platforms, where the infrared (IR) signals emitted from the VR controllers for controller-headset interactions can be maliciously exploited to reconstruct unconstrained input keystrokes on the virtual keyboard non-intrusively. We propose a novel keystroke inference attack named VRecKey to demonstrate the feasibility and practicality of this novel infrared side channel. Specifically, VRecKey leverages a customized 2D IR sensor array to intercept ambient IR signals emitted from VR controllers and subsequently infers (i) character-level key presses on the virtual keyboard and (ii) word-level keystrokes along with their typing trajectories. We extensively evaluate the effectiveness of VRecKey with two commercial VR devices, and the results indicate that it can achieve over 94.2% and 90.5% top-3 accuracy in inferring character-level and word-level keystrokes with varying lengths, respectively. In addition, empirical results show that VRecKey is resilient to several practical impact factors and presents effectiveness in various real-world scenarios, which provides a complementary and orthogonal attack surface for the exploration of keystroke inference attacks in VR platforms.

cs.CR↗
arXiv2024

Interlayer coupling rotatable magnetic easy-axis in MnSe2 mono- and bi-layers

Interlayer coupling plays a critical role in tuning the electronic structures and magnetic ground states of two-dimensional materials, influenced by the number of layers, interlayer distances, and stacking order. However, its effect on the orientation of the magnetic easy axis remains underexplored. In this study, we demonstrate that interlayer coupling can significantly alter the magnetic easy-axis orientation, as shown by the magnetic easy-axis of monolayer 1T-MnSe2 tilting 67° from the z-axis, while aligning with the z-axis in the bilayer. This change results from variations in orbital occupations near the Fermi level, particularly involving nonmetallic Se atoms. Contrary to the traditional focus on magnetic metal atoms, our findings reveal that Se orbitals play a key role in influencing the easy-axis orientation and topological Chern numbers. Furthermore, we validated our conclusions by changing stacking orders, introducing charge doping, applying in-plane biaxial strains, and substituting non-metallic atoms. Our results highlight the pivotal role of interlayer coupling in tuning the magnetic properties of layered materials, with important implications for spintronic applications.

cond-mat.mtrl-sci↗