arXiv subjects

Lei Yu

Publications and source records attributed to Lei Yu.

At least 91 records · Page 5Linked to original sources

arXiv2025

MOS: Towards Effective Smart Contract Vulnerability Detection through Mixture-of-Experts Tuning of Large Language Models

Smart contract vulnerabilities pose significant security risks to blockchain systems, potentially leading to severe financial losses. Existing methods face several limitations: (1) Program analysis-based approaches rely on predefined patterns, lacking flexibility for new vulnerability types; (2) Deep learning-based methods lack explanations; (3) Large language model-based approaches suffer from high false positives. We propose MOS, a smart contract vulnerability detection framework based on mixture-of-experts tuning (MOE-Tuning) of large language models. First, we conduct continual pre-training on a large-scale smart contract dataset to provide domain-enhanced initialization. Second, we construct a high-quality MOE-Tuning dataset through a multi-stage pipeline combining LLM generation and expert verification for reliable explanations. Third, we design a vulnerability-aware routing mechanism that activates the most relevant expert networks by analyzing code features and their matching degree with experts. Finally, we extend the feed-forward layers into multiple parallel expert networks, each specializing in specific vulnerability patterns. We employ a dual-objective loss function: one for optimizing detection and explanation performance, and another for ensuring reasonable distribution of vulnerability types to experts through entropy calculation. Experiments show that MOS significantly outperforms existing methods with average improvements of 6.32% in F1 score and 4.80% in accuracy. The vulnerability explanations achieve positive ratings (scores of 3-4 on a 4-point scale) of 82.96%, 85.21% and 94.58% for correctness, completeness, and conciseness through human and LLM evaluation.

cs.SE↗

arXiv2025

NTIRE 2025 Challenge on Event-Based Image Deblurring: Methods and Results

This paper presents an overview of NTIRE 2025 the First Challenge on Event-Based Image Deblurring, detailing the proposed methodologies and corresponding results. The primary goal of the challenge is to design an event-based method that achieves high-quality image deblurring, with performance quantitatively assessed using Peak Signal-to-Noise Ratio (PSNR). Notably, there are no restrictions on computational complexity or model size. The task focuses on leveraging both events and images as inputs for single-image deblurring. A total of 199 participants registered, among whom 15 teams successfully submitted valid results, offering valuable insights into the current state of event-based image deblurring. We anticipate that this challenge will drive further advancements in event-based vision research.

cs.CV↗

arXiv2025

On Average Distance, Level-1 Fourier Weight, and Chang's Lemma

In this paper, we improve the well-known level-1 weight bound, also known as Chang's lemma, by using an induction method. Our bounds are close to optimal no matter when the set is large or small. Our bounds can be seen as bounds on the minimum average distance problem, since maximizing the level-1 weight is equivalent to minimizing the average distance. We apply our new bounds to improve the Friedgut--Kalai--Naor theorem. We also derive the sharp version for Chang's original lemma for $\mathbb{F}_{2}^{n}$. That is, we show that in $\mathbb{F}_{2}^{n}$, Hamming balls maximize the dimension of the space spanned by large Fourier coefficients.

math.CO↗

arXiv2025

Exploring Scene Affinity for Semi-Supervised LiDAR Semantic Segmentation

This paper explores scene affinity (AIScene), namely intra-scene consistency and inter-scene correlation, for semi-supervised LiDAR semantic segmentation in driving scenes. Adopting teacher-student training, AIScene employs a teacher network to generate pseudo-labeled scenes from unlabeled data, which then supervise the student network's learning. Unlike most methods that include all points in pseudo-labeled scenes for forward propagation but only pseudo-labeled points for backpropagation, AIScene removes points without pseudo-labels, ensuring consistency in both forward and backward propagation within the scene. This simple point erasure strategy effectively prevents unsupervised, semantically ambiguous points (excluded in backpropagation) from affecting the learning of pseudo-labeled points. Moreover, AIScene incorporates patch-based data augmentation, mixing multiple scenes at both scene and instance levels. Compared to existing augmentation techniques that typically perform scene-level mixing between two scenes, our method enhances the semantic diversity of labeled (or pseudo-labeled) scenes, thereby improving the semi-supervised performance of segmentation models. Experiments show that AIScene outperforms previous methods on two popular benchmarks across four settings, achieving notable improvements of 1.9% and 2.1% in the most challenging 1% labeled data.

cs.CV↗

arXiv2025

Misaligned Roles, Misplaced Images: Structural Input Perturbations Expose Multimodal Alignment Blind Spots

Multimodal Language Models (MMLMs) typically undergo post-training alignment to prevent harmful content generation. However, these alignment stages focus primarily on the assistant role, leaving the user role unaligned, and stick to a fixed input prompt structure of special tokens, leaving the model vulnerable when inputs deviate from these expectations. We introduce Role-Modality Attacks (RMA), a novel class of adversarial attacks that exploit role confusion between the user and assistant and alter the position of the image token to elicit harmful outputs. Unlike existing attacks that modify query content, RMAs manipulate the input structure without altering the query itself. We systematically evaluate these attacks across multiple Vision Language Models (VLMs) on eight distinct settings, showing that they can be composed to create stronger adversarial prompts, as also evidenced by their increased projection in the negative refusal direction in the residual stream, a property observed in prior successful attacks. Finally, for mitigation, we propose an adversarial training approach that makes the model robust against input prompt perturbations. By training the model on a range of harmful and benign prompts all perturbed with different RMA settings, it loses its sensitivity to Role Confusion and Modality Manipulation attacks and is trained to only pay attention to the content of the query in the input prompt structure, effectively reducing Attack Success Rate (ASR) while preserving the model's general utility.

cs.CR↗

arXiv2025

The Reasoning-Memorization Interplay in Language Models Is Mediated by a Single Direction

Large language models (LLMs) excel on a variety of reasoning benchmarks, but previous studies suggest they sometimes struggle to generalize to unseen questions, potentially due to over-reliance on memorized training examples. However, the precise conditions under which LLMs switch between reasoning and memorization during text generation remain unclear. In this work, we provide a mechanistic understanding of LLMs' reasoning-memorization dynamics by identifying a set of linear features in the model's residual stream that govern the balance between genuine reasoning and memory recall. These features not only distinguish reasoning tasks from memory-intensive ones but can also be manipulated to causally influence model performance on reasoning tasks. Additionally, we show that intervening in these reasoning features helps the model more accurately activate the most relevant problem-solving capabilities during answer generation. Our findings offer new insights into the underlying mechanisms of reasoning and memory in LLMs and pave the way for the development of more robust and interpretable generative AI systems.

cs.CL↗

arXiv2025

Robust LLM safeguarding via refusal feature adversarial training

Large language models (LLMs) are vulnerable to adversarial attacks that can elicit harmful responses. Defending against such attacks remains challenging due to the opacity of jailbreaking mechanisms and the high computational cost of training LLMs robustly. We demonstrate that adversarial attacks share a universal mechanism for circumventing LLM safeguards that works by ablating a dimension in the residual stream embedding space called the refusal feature. We further show that the operation of refusal feature ablation (RFA) approximates the worst-case perturbation of offsetting model safety. Based on these findings, we propose Refusal Feature Adversarial Training (ReFAT), a novel algorithm that efficiently performs LLM adversarial training by simulating the effect of input-level attacks via RFA. Experiment results show that ReFAT significantly improves the robustness of three popular LLMs against a wide range of adversarial attacks, with considerably less computational overhead compared to existing adversarial training methods.

cs.LG↗

arXiv2025

Leveraging Large Language Models for Enhanced Digital Twin Modeling: Trends, Methods, and Challenges

Digital twin technology is a transformative innovation driving the digital transformation and intelligent optimization of manufacturing systems. By integrating real-time data with computational models, digital twins enable continuous monitoring, simulation, prediction, and optimization, effectively bridging the gap between the physical and digital worlds. Recent advancements in communication, computing, and control technologies have accelerated the development and adoption of digital twins across various industries. However, significant challenges remain, including limited data for accurate system modeling, inefficiencies in system analysis, and a lack of explainability in the interactions between physical and digital systems. The rise of large language models (LLMs) offers new avenues to address these challenges. LLMs have shown exceptional capabilities across diverse domains, exhibiting strong generalization and emergent abilities that hold great potential for enhancing digital twins. This paper provides a comprehensive review of recent developments in LLMs and their applications to digital twin modeling. We propose a unified description-prediction-prescription framework to integrate digital twin modeling technologies and introduce a structured taxonomy to categorize LLM functionalities in these contexts. For each stage of application, we summarize the methodologies, identify key challenges, and explore potential future directions. To demonstrate the effectiveness of LLM-enhanced digital twins, we present an LLM-enhanced enterprise digital twin system, which enables automatic modeling and optimization of an enterprise. Finally, we discuss future opportunities and challenges in advancing LLM-enhanced digital twins, offering valuable insights for researchers and practitioners in related fields.

cs.ET↗

arXiv2025

Privacy and Accuracy-Aware AI/ML Model Deduplication

With the growing adoption of privacy-preserving machine learning algorithms, such as Differentially Private Stochastic Gradient Descent (DP-SGD), training or fine-tuning models on private datasets has become increasingly prevalent. This shift has led to the need for models offering varying privacy guarantees and utility levels to satisfy diverse user requirements. However, managing numerous versions of large models introduces significant operational challenges, including increased inference latency, higher resource consumption, and elevated costs. Model deduplication is a technique widely used by many model serving and database systems to support high-performance and low-cost inference queries and model diagnosis queries. However, none of the existing model deduplication works has considered privacy, leading to unbounded aggregation of privacy costs for certain deduplicated models and inefficiencies when applied to deduplicate DP-trained models. We formalize the problems of deduplicating DP-trained models for the first time and propose a novel privacy- and accuracy-aware deduplication mechanism to address the problems. We developed a greedy strategy to select and assign base models to target models to minimize storage and privacy costs. When deduplicating a target model, we dynamically schedule accuracy validations and apply the Sparse Vector Technique to reduce the privacy costs associated with private validation data. Compared to baselines that do not provide privacy guarantees, our approach improved the compression ratio by up to $35\times$ for individual models (including large language models and vision transformers). We also observed up to $43\times$ inference speedup due to the reduction of I/O operations.

cs.CR↗

arXiv2025

Frequency-Adaptive Low-Latency Object Detection Using Events and Frames

Fusing Events and RGB images for object detection leverages the robustness of Event cameras in adverse environments and the rich semantic information provided by RGB cameras. However, two critical mismatches: low-latency Events \textit{vs.}~high-latency RGB frames; temporally sparse labels in training \textit{vs.}~continuous flow in inference, significantly hinder the high-frequency fusion-based object detection. To address these challenges, we propose the \textbf{F}requency-\textbf{A}daptive Low-Latency \textbf{O}bject \textbf{D}etector (FAOD). FAOD aligns low-frequency RGB frames with high-frequency Events through an Align Module, which reinforces cross-modal style and spatial proximity to address the Event-RGB Mismatch. We further propose a training strategy, Time Shift, which enforces the module to align the prediction from temporally shifted Event-RGB pairs and their original representation, that is, consistent with Event-aligned annotations. This strategy enables the network to use high-frequency Event data as the primary reference while treating low-frequency RGB images as supplementary information, retaining the low-latency nature of the Event stream toward high-frequency detection. Furthermore, we observe that these corrected Event-RGB pairs demonstrate better generalization from low training frequency to higher inference frequencies compared to using Event data alone. Extensive experiments on the PKU-DAVIS-SOD and DSEC-Detection datasets demonstrate that our FAOD achieves SOTA performance. Specifically, in the PKU-DAVIS-SOD Dataset, FAOD achieves 9.8 points improvement in terms of the mAP in fully paired Event-RGB data with only a quarter of the parameters compared to SODFormer, and even maintains robust performance (only a 3 points drop in mAP) under 80$\times$ Event-RGB frequency mismatch.

cs.CV↗

arXiv2025

Sparse Brains are Also Adaptive Brains: Cognitive-Load-Aware Dynamic Activation for LLMs

Dense large language models(LLMs) face critical efficiency bottlenecks as they rigidly activate all parameters regardless of input complexity. While existing sparsity methods(static pruning or dynamic activation) address this partially, they either lack adaptivity to contextual or model structural demands or incur prohibitive computational overhead. Inspired by human brain's dual-process mechanisms - predictive coding (N400) for backbone sparsity and structural reanalysis (P600) for complex context - we propose CLADA, a \textit{\textbf{C}ognitive-\textbf{L}oad-\textbf{A}ware \textbf{D}ynamic \textbf{A}ctivation} framework that synergizes statistical sparsity with semantic adaptability. Our key insight is that LLM activations exhibit two complementary patterns: 1) \textit{Global statistical sparsity} driven by sequence-level prefix information, and 2) \textit{Local semantic adaptability} modulated by cognitive load metrics(e.g., surprisal and entropy). CLADA employs a hierarchical thresholding strategy: a baseline from offline error-controlled optimization ensures 40\%+ sparsity, dynamically adjusted by real-time cognitive signals. Evaluations across six mainstream LLMs and nine benchmarks demonstrate that CLADA achieves \textbf{~20\% average speedup with <2\% accuracy drop}, outperforming Griffin (5\%+ degradation) and TT (negligible speedup). Crucially, we establish the first formal connection between neurolinguistic event-related potential (ERP) components and LLM efficiency mechanisms through multi-level regression analysis ($R^2=0.17$ for sparsity-adaptation synergy). Requiring no retraining or architectural changes, CLADA offers a deployable solution for resource-aware LLM inference while advancing biologically-inspired AI design. Our code is available at \href{https://github.com/Oldify/CLADA}{CLADA}.

cs.CL↗

arXiv2025

On the Fundamental Limits of Integrated Sensing and Communications Under Logarithmic Loss

We study a unified information-theoretic framework for integrated sensing and communications (ISAC), applicable to both monostatic and bistatic sensing scenarios. Special attention is given to the case where the sensing receiver (Rx) is required to produce a "soft" estimate of the state sequence, with logarithmic loss serving as the performance metric. We derive lower and upper bounds on the capacity-distortion function, which delineates the fundamental tradeoff between communication rate and sensing distortion. These bounds coincide when the channel between the ISAC transmitter (Tx) and the communication Rx is degraded with respect to the channel between the ISAC Tx and the sensing Rx, or vice versa. Furthermore, we provide a complete characterization of the capacity-distortion function for an ISAC system that simultaneously transmits information over a binary-symmetric channel and senses additive Bernoulli states through another binary-symmetric channel. The Gaussian counterpart of this problem is also explored, which, together with a state-splitting trick, fully determines the capacity-distortion-power function under the squared error distortion measure.

cs.IT↗

arXiv2025

Fresh2comm: Information Freshness Optimized Collaborative Perception

Collaborative perception is a cornerstone of intelligent connected vehicles, enabling them to share and integrate sensory data to enhance situational awareness. However, measuring the impact of high transmission delay and inconsistent delay on collaborative perception in real communication scenarios, as well as improving the effectiveness of collaborative perception under such conditions, remain significant challenges in the field. To address these challenges, we incorporate the key factor of information freshness into the collaborative perception mechanism and develop a model that systematically measures and analyzes the impacts of real-world communication on collaborative perception performance. This provides a new perspective for accurately evaluating and optimizing collaborative perception performance. We propose and validate an Age of Information (AoI)-based optimization framework that strategically allocates communication resources to effectively control the system's AoI, thereby significantly enhancing the freshness of information transmission and the accuracy of perception. Additionally, we introduce a novel experimental approach that comprehensively assesses the varying impacts of different types of delay on perception results, offering valuable insights for perception performance optimization under real-world communication scenarios.

cs.MA↗

arXiv2025

Navigating pollution: A multimodal approach to traffic and exposure management

Few studies quantify how traffic management dynamically reshapes modal split and emission-exposure outcomes over pollution severities. This paper proposes a novel day-to-day assignment model integrating exposure cost, which includes exposure perception and emissions-dispersion-exposure algorithm. Numerical experiments reveal that and various levels of traffic-related measures have an air pollution scenario-dependent effect on the MT system. In light pollution scenarios, vehicle restrictions and reduced fares for buses or ridesharing help lower car usage and reduce emissions and exposure. However, under heavy pollution, higher-level restrictions and ridesharing fares paradoxically increase travelers' exposure by 18% and 6.3%, respectively, due to modal shift. Furthermore, timely pollution information updates could plausibly encourage healthier travel. This paper also proposes practical strategies for both routine and emergency traffic management, considering the trade-offs among travel cost, emission, and exposure, and emphasizes the need for measures tailored to different air pollution contexts to offer deeper insights for urban traffic policies.

physics.soc-ph↗

arXiv2025

MTGA: Multi-View Temporal Granularity Aligned Aggregation for Event-Based Lip-Reading

Lip-reading is to utilize the visual information of the speaker's lip movements to recognize words and sentences. Existing event-based lip-reading solutions integrate different frame rate branches to learn spatio-temporal features of varying granularities. However, aggregating events into event frames inevitably leads to the loss of fine-grained temporal information within frames. To remedy this drawback, we propose a novel framework termed Multi-view Temporal Granularity aligned Aggregation (MTGA). Specifically, we first present a novel event representation method, namely time-segmented voxel graph list, where the most significant local voxels are temporally connected into a graph list. Then we design a spatio-temporal fusion module based on temporal granularity alignment, where the global spatial features extracted from event frames, together with the local relative spatial and temporal features contained in voxel graph list are effectively aligned and integrated. Finally, we design a temporal aggregation module that incorporates positional encoding, which enables the capture of local absolute spatial and global temporal information. Experiments demonstrate that our method outperforms both the event-based and video-based lip-reading counterparts.

cs.CV↗

arXiv2025

DeepCRCEval: Revisiting the Evaluation of Code Review Comment Generation

Code review is a vital but demanding aspect of software development, generating significant interest in automating review comments. Traditional evaluation methods for these comments, primarily based on text similarity, face two major challenges: inconsistent reliability of human-authored comments in open-source projects and the weak correlation of text similarity with objectives like enhancing code quality and detecting defects. This study empirically analyzes benchmark comments using a novel set of criteria informed by prior research and developer interviews. We then similarly revisit the evaluation of existing methodologies. Our evaluation framework, DeepCRCEval, integrates human evaluators and Large Language Models (LLMs) for a comprehensive reassessment of current techniques based on the criteria set. Besides, we also introduce an innovative and efficient baseline, LLM-Reviewer, leveraging the few-shot learning capabilities of LLMs for a target-oriented comparison. Our research highlights the limitations of text similarity metrics, finding that less than 10% of benchmark comments are high quality for automation. In contrast, DeepCRCEval effectively distinguishes between high and low-quality comments, proving to be a more reliable evaluation mechanism. Incorporating LLM evaluators into DeepCRCEval significantly boosts efficiency, reducing time and cost by 88.78% and 90.32%, respectively. Furthermore, LLM-Reviewer demonstrates significant potential of focusing task real targets in comment generation.

cs.SE↗

arXiv2025

Joint Transmission and Deblurring: A Semantic Communication Approach Using Events

Deep learning-based joint source-channel coding (JSCC) is emerging as a promising technology for effective image transmission. However, most existing approaches focus on transmitting clear images, overlooking real-world challenges such as motion blur caused by camera shaking or fast-moving objects. Motion blur often degrades image quality, making transmission and reconstruction more challenging. Event cameras, which asynchronously record pixel intensity changes with extremely low latency, have shown great potential for motion deblurring tasks. However, the efficient transmission of the abundant data generated by event cameras remains a significant challenge. In this work, we propose a novel JSCC framework for the joint transmission of blurry images and events, aimed at achieving high-quality reconstructions under limited channel bandwidth. This approach is designed as a deblurring task-oriented JSCC system. Since RGB cameras and event cameras capture the same scene through different modalities, their outputs contain both shared and domain-specific information. To avoid repeatedly transmitting the shared information, we extract and transmit their shared information and domain-specific information, respectively. At the receiver, the received signals are processed by a deblurring decoder to generate clear images. Additionally, we introduce a multi-stage training strategy to train the proposed model. Simulation results demonstrate that our method significantly outperforms existing JSCC-based image transmission schemes, addressing motion blur effectively.

eess.IV↗

arXiv2025

Healing of the edge magnetic island in the island divertor configuration on J-TEXT

The phenomena of island healing and configuration transition induced by high-power electron cyclotron resonance heating (ECRH) have been investigated in the island divertor configuration on the J-TEXT tokamak. Experimental results reveal that the size of the edge open magnetic island with mode number m/n = 3/1 decreases substantially under specific ECRH conditions. This process, referred to as island healing, occurs when ECRH with a power of 500~600 kW is deposited in the plasma core or when 250 kW of ECRH is deposited at r = 0.5 a, where a is the minor radius. The reduction of the island width makes the island divertor ineffective and transition into the limiter configuration. A model incorporating the influence of ECRH on the scrape-off layer (SOL) thermoelectric current is proposed to explain the observed changes in the edge magnetic topology of the island divertor configuration. These findings suggest that ECRH should be deposited at the plasma core with carefully controlled power to ensure the stable and compatible operation of ECRH and the island divertor configuration in tokamaks. The results can provide insights into achieving robust operation of an island divertor in tokamaks.

physics.plasm-ph↗