Automated redesign without manual adjustments marks a key step forward in the design workflow. In this work, we focus on a foundational redesign task termed design layout editing, which seeks to autonomously modify the geometric composition of a design based on user intents. To overcome the ambiguity of user needs expressed in natural language, we introduce four basic and important editing actions and standardize the format of editing operations. The underexplored task presents a unique challenge: satisfying specified editing operations while simultaneously preserving the layout structure of unedited elements. Besides, the scarcity of triplet (original design, editing operation, edited design) samples poses another formidable challenge. To this end, we present ReLayout, a novel framework for versatile and structure-preserving design layout editing that operates without triplet data. Specifically, ReLayout first introduces the relation graph, which contains the position and size relationships among unedited elements, as the constraint for layout structure preservation. Then, relation-aware design reconstruction (RADR) is proposed to bypass the data challenge. By learning to reconstruct a design from its elements, a relation graph, and a synthesized editing operation, RADR effectively emulates the editing process in a self-supervised manner. A multi-modal large language model serves as the backbone for RADR, unifying multiple editing actions within a single model and thus achieving versatile editing after fine-tuning. Qualitative, quantitative results and user studies show that ReLayout significantly outperforms the baseline models in terms of editing quality, accuracy, and layout structure preservation.

Third-Party Library (TPL) detection, which identifies reused libraries in binary code, is critical for software security analysis. At its core, TPL detection depends on binary decomposition-the process of partitioning a monolithic binary into cohesive modules. Existing decomposition methods, whether anchor-based or clustering-based, fundamentally rely on the assumption that reused code exhibits similar function call relationships. However, this assumption is severely undermined by Function Call Graph (FCG) variations introduced by diverse compilation settings, particularly function inlining decisions that drastically alter FCG structures. In this work, we conduct the first systematic empirical study to establish the oracle for optimal binary decomposition under compilation variance. We first develop a labeling method to create precise FCG mappings on a comprehensive dataset compiled with 17 compilers, 6 optimizations, and 4 architectures; then, we identify the minimum semantic-equivalent function regions between FCG variants to derive the ground-truth decomposition. This oracle provides the first rigorous evaluation framework that quantitatively assesses decomposition algorithms under compilation variance. Using this oracle, we evaluate existing methods and expose their critical limitations: they either suffer from under-aggregation failure or over-aggregation failure. Our findings reveal that current decomposition techniques are inadequate for robust TPL detection, highlighting the urgent need for compilation-aware approaches.

The architecture of software systems evolves along with their upgrades and maintenance, inevitably creating a gap between the defact architecture and the designed one. To perceive and fix the discrepancy, clustering-based architecture recovery methods have been developed to re-engineer the real-time system architecture from the code implementation. However, existing solutions still face several limitations. They underutilize both code-level and architecture-level semantics underlying the source code. Moreover, they overlook implicit structural dependencies that complement explicit ones to reflect code interactions. To address these challenges, we propose SemArc, an architecture recovery method that utilizes large language models to comprehend both implementation-level and architecture-level semantics, supported by well-established canonical architectural patterns as a knowledge base. SemArc also incorporates both implicit and explicit dependencies to complete the system behavior representations. Additionally, SemArc introduces a component-as-anchor guided clustering algorithm to improve the clustering process. We evaluated SemArc on 15 software systems written in C/C++, Java, and Python, using five different metrics. The results demonstrate that SemArc outperforms seven baseline methods by an average of 32 percentage points. We also examined how three factors—code semantics, architectural semantics, and implicit dependencies—as well as different levels of architectural semantic descriptions, influence recovery accuracy. A case study on the Bash project indicates that SemArc has the potential to yield even more precise recovery results than those labeled by humans.

Rug pull scams have emerged as a persistent threat to cryptocurrency, causing significant financial losses. A typical scenario involves scammers deploying honeypot contracts to attract investments, restricting token sales, and draining the funds, which leaves investors with worthless tokens. Current methods either rely on predefined patterns to detect code risks or utilize statistical transaction data to train detection models. However, real-world Rug Pull schemes often involve a complex interplay between malicious code and suspicious transaction behaviors. These methods, which solely focus on one aspect, fall short in detecting such schemes effectively. In this paper, we propose RPHunter, a novel technique that integrates code and transaction for Rug Pull detection. First, RPHunter establishes declarative rules and performs flow analysis to extract code risk information, further constructing a semantic risk code graph (SRCG). Meanwhile, to leverage transaction information, RPHunter formulates dynamic token transaction activities as a token flow behavior graph (TFBG) in which nodes and edges are characterized from network structure and market manipulation perspectives. Finally, RPHunter employs graph neural networks to extract complementary features from SRCG and TFBG, integrating them through an attention fusion model to enhance the detection of Rug Pull. We manually analyzed 645 Rug Pull incidents from code and transaction aspects and constructed a ground-truth dataset. We evaluated RPHunter on our dataset, achieving a precision of 95.3

Chain-of-thought (CoT) supervision exposes intermediate rationales, but flat rationale targets usually optimize a single reasoning sequence and provide limited supervision on how local conclusions should support later decisions. We introduce Dependency-Aware Intermediate QA Supervision (DAIS), a training-time framework that converts filtered teacher rationales into stage-level QA records. Each intermediate record predicts a local answer conditioned on the previous states needed for that decision, while the final-answer record keeps the original task format; evaluation therefore uses only the original input and optional context. Across GDPR, AIACT, MedQA, and FOLIO with multiple Qwen backbones, DAIS improves average final-answer accuracy over answer-only, flat chain-of-thought, and independent-QA baselines. On policy-compliance benchmarks, it achieves a largest gain of 5.6

MLLMs are increasingly deployed in user-facing applications, yet they inherit backdoor risks from the pipelines used to construct them: triggers may reside in images, texts, or both. Existing model-level backdoor removal methods, largely designed for conventional classifiers, show limited effectiveness on MLLMs, while MLLM-specific defenses mainly operate at inference time, filtering suspicious inputs without removing the backdoor embedded in the model. To address this gap and eliminate latent backdoors from MLLMs at their source, we present RACER, a model-level repair framework motivated by a key observation: backdoors induce abnormal layer-to-layer evolution in internal representations, which we term the layer-wise inconsistency anomaly. Importantly, this anomaly is modality-dependent, concentrating primarily in the token region encoding the trigger features that the backdoor model actually relies on. RACER therefore decomposes the fused representation into visual and textual token regions, normalizes their layer-wise inconsistency separately, and recomposes them using modality-aware weights over a deep-layer window, yielding a region-aware inconsistency objective that better captures localized backdoor-induced anomalies. Through a min-max optimization, this objective drives worst-case perturbation synthesis and adversarial fine-tuning against the resulting perturbation to repair the model, suppressing the deep representational directional shifts on which backdoor behaviors rely. RACER requires only 100 clean samples and no knowledge of the trigger, attack objective, or even whether the input model contains a backdoor. Evaluations on three open-source MLLMs across 36 backdoor settings spanning image, text, and multimodal triggers show that RACER reduces the average ASR to 1.1

With the rapid advancement of decentralized finance (DeFi), security incidents related to cryptocurrency have become increasingly prevalent. After such incidents, attackers typically attempt to rapidly move stolen assets, concealing the origin of illicit funds and ultimately converting them into fiat currency. However, existing anti-money laundering (AML) methods struggle to cope with the semantic complexity of DeFi transactions. They either rely heavily on low-level token transfers, or perform protocol-agnostic money flow analysis, failing to capture the high-level intent of transactions. In this paper, we propose AMLGuard, a semantic-aware AML framework for account-based blockchains. AMLGuard tracks illicit fund flows from known malicious addresses by performing semantic analysis on complex DeFi transactions, enabling accurate and continuous laundering tracking. Given a complex transaction, AMLGuard combines static rule-based analysis with retrieval-augmented large language model (LLM) reasoning to infer implicit DeFi semantics, transforming raw transaction data into high-level semantic representations. Furthermore, for cross-chain transactions where laundering intent is not explicitly exposed, AMLGuard parses transaction parameters and performs argument parsing to recover cross-chain semantics, enabling seamless tracking across ledgers. Based on the inferred semantics, AMLGuard abstracts each transaction into a DeFi Semantic Unit (DSU). We evaluate the effectiveness of AMLGuard on 82 real-world laundering cases, involving illicit assets worth over $1 billion. Specifically, AMLGuard reconstructs compact illicit fund-flow topologies with destination precision of 94.4% and 87.6%, while achieving the highest address recall of 98.4% and 95.8% and destination recall of 94.1% and 93.8% on single-chain and cross-chain datasets.

Modern Cyber-Physical Power Grids (CPPG) rely heavily on fine-grained data exchange, yet they face the concealed threat of Data Inference Leakage (DIL), where attackers infer sensitive parameters from seemingly non-sensitive data. Existing risk assessment methods typically focus on direct cyber-attacks or employ static binary classifications, failing to capture the dynamic risk propagation caused by complex physical and statistical data couplings. To address this, this paper proposes a unified data sensitivity assessment framework based on a novel Inference Graph Neural Network (IGNN). We first construct an inference relation graph that integrates both model-driven and data-driven correlations. An analytic hierarchy process (AHP) model is constructed to calibrate the intrinsic sensitivity baseline. Subsequently, a novel inference-weighted Graph Attention Network (GAT) is developed to quantify the final sensitivity by aggregating the risk propagating through the inference pathways. Experiments on IEEE 30-, 57-bus, ACTIV 200-bus, and case 1354-pegase systems demonstrate that the proposed method effectively quantifies data sensitivity and identifies high-risk data susceptible to DIL, exhibiting strong robustness across different system scales. Quantitative results indicate that the high-sensitivity data identified by our method yield a highly threatening inference accuracy of up to 0.9145, significantly outperforming GCN-based and rule-based benchmarks.

Model Context Protocol (MCP)-style ecosystems give language-model applications a practical connection layer for tools, resources, prompts, and transports. As agents move from connection to execution, security decisions often remain split across clients, servers, prompts, approval dialogs, OAuth deployments, and logs. This paper asks whether a runtime can make execution-layer invariants explicit and testable while preserving MCP-like workflows. We define eight invariants: metadata non-authority, grant-backed approval, canonical resources, principal binding, scoped capability invocation, source-and-target data-flow authorization, deny-path audit, and explicit protocol state. We implement these invariants in HCP, a Handle-Capability Protocol reference runtime for MCP-style agent execution that represents calls through principals, resources, grants, capabilities, handles, policy decisions, data-pipe checks, and audit entries. We evaluate HCP against two MCP-like baselines: a naive connection-layer runtime and a practice-informed connection-layer mitigation baseline with metadata linting, session checks, and per-call approvals. Across 10 benchmark cases, the naive baseline permits all modeled attacks, the mitigation baseline permits 6 of 10, and HCP blocks all 10 while preserving audit evidence. Ablations identify which runtime components block attacks and preserve forensic evidence. A local in-memory microbenchmark reports sub-millisecond mean latencies for measured policy, invocation, peek, and pipe operations. A bounded GitHub README-screening sample provides ecosystem signals, not vulnerability findings. The results support a narrow claim: MCP-style agent systems need an execution-control layer in addition to connection-layer conventions.

The growing application of large language models (LLMs) in safety-critical domains has raised urgent concerns about their security. Many recent studies have demonstrated the feasibility of backdoor attacks against LLMs. However, existing methods suffer from three key shortcomings: explicit trigger patterns that compromise naturalness, unreliable injection of attacker-specified payloads in long-form generation, and incompletely specified threat models that obscure how backdoors are delivered and activated in practice. To address these gaps, we present BadStyle, a complete backdoor attack framework and pipeline. BadStyle leverages an LLM as a poisoned sample generator to construct natural and stealthy poisoned samples that carry imperceptible style-level triggers while preserving semantics and fluency. To stabilize payload injection during fine-tuning, we design an auxiliary target loss that reinforces the attacker-specified target content in responses to poisoned inputs and penalizes its emergence in benign responses. We further ground the attack in a realistic threat model and systematically evaluate BadStyle under both prompt-induced and PEFT-based injection strategies. Extensive experiments across seven victim LLMs, including LLaMA, Phi, DeepSeek, and GPT series, demonstrate that BadStyle achieves high attack success rates (ASRs) while maintaining strong stealthiness. The proposed auxiliary target loss substantially improves the stability of backdoor activation, yielding an average ASR improvement of around 30

The weighted least squares (WLS) method has been widely used in power systems for state estimation (SE), but in the presence of gross errors, the results of WLS SE can be significantly biased. While various bad data identification (BDI) methods exist, identifying multiple bad data for SE is still challenging, especially when bad data located at leverage points can disproportionately corrupt SE accuracy. This paper proposes an Optimal-Weighted Least Squares State Estimation (OWLS-SE) method with an optimal weight-tuning strategy to address these challenges. The impact of bad data at leverage points on estimation errors is first mathematically demonstrated through a theoretical upper bound of the SE relative error. Then, the optimal dynamic weights are developed to minimize this error bound by quantifying leverage point impacts and integrating them into a multi-objective optimization framework. The proposed OWLS-SE achieves simultaneous SE and BDI within a single optimization model, eliminating conventional two-stage processing. Numerical experiments on the 6- bus system, IEEE 30-, 118- bus systems, and the 2869- bus PEGASE system demonstrate that OWLS-SE achieves superior BDI and SE accuracy with robust performance even when multiple gross errors occur at leverage points, while maintaining computational efficiency comparable to the simplest WLS method.

With the increasing integration of distributed energy resources (DERs) and electric vehicles (EVs), fast and accurate distribution system state estimation (DSSE) has become indispensable for real-time monitoring and optimal control of modern power distribution systems. However, conventional DSSE methods face considerable challenges when addressing frequent topology changes. These challenges include high computational costs due to iterative optimization, sensitivity to parameter uncertainties, and limited adaptability of existing data-driven models to evolving network configurations. To overcome these limitations, this paper proposes a topology-adaptive DSSE framework based on a physics-informed autoencoder. The encoder utilizes LEAPNet to extract measurement features and capture topology-dependent variations, thereby generating accurate and topology-aware state estimates. The decoder reconstructs input measurements by embedding Kirchhoff’s laws, ensuring that estimated states adhere to physical power system constraints. Through joint optimization of the encoder and decoder, the network learns physically consistent mappings between measurements and system states. Extensive simulation studies on distribution networks demonstrate that the proposed model achieves superior accuracy, computational efficiency, and physical consistency compared with conventional DSSE approaches.

Standardized token contracts (e.g., ERC-20) form the foundation of digital assets. However, attackers increasingly abuse this standardization to disguise malicious trap tokens. Unlike obvious violations, these contracts employ a strategy of "deceptive adherence": they strictly adhere to standard protocols to evade detection while embedding covert logic to defraud users. To address this, we first systematize the trap landscape by proposing a novel taxonomy derived from the intrinsic functional lifecycle of tokens (Generation, Circulation, Persistence, and Observation). We then propose TrapHunter, a framework designed to identify these traps and expose covert pathways within these deceptive contracts via intent deviation analysis. Specifically, TrapHunter introduces a unified semantic representation combining Abstract Behavior Trees (ABTs) and Augmented Path Graphs (APGs) to normalize intra-procedural syntax and reveal hidden execution paths driven by inter-procedural state dependencies. Crucially, it bridges the semantic gap by leveraging LLMs to reason about the behavioral intent of deviations from reference implementations, followed by fork-based dynamic validation to confirm exploitability. Experimental evaluation on 269 real-world contracts with three LLMs (DeepSeek, GPT, and Gemini) demonstrates that TrapHunter effectively detects all six categories of traps, achieving an average precision of 81.8

Large Language Models (LLMs) achieve remarkable capabilities through dense Transformers, yet their insatiable energy consumption poses a significant barrier to ubiquitous deployment. Spiking Neural Networks (SNNs) promise ordersof-magnitude energy savings through sparse, event-driven computation, but natively-trained SNNs have historically struggled to scale to general-domain language modeling. We present SymbolicLight, a Pure SNN language model architecture that replaces dense self-attention and continuous activations with spike-gated associative lookup and Leaky Integrate-and-Fire (LIF) dynamics. To overcome the profound instability of deep SNN scaling, we introduce an ATan Surrogate Gradient that guarantees peak gradient magnitudes of exactly 1.0, and adapt Rotary Position Embedding (RoPE) for temporal spike sequence extrapolation. We train a 176M-parameter SymbolicLight model from scratch on 3 billion tokens across 6 diverse English-language domains. Under strictly controlled conditions against a computationally matched GPT-2 baseline, SymbolicLight maintains 90.7% encoder-level activation sparsity and 85.0% global all-layer sparsity throughout training without explicit regularization. While a perplexity gap remains (val PPL 30.8 vs. 13.1 for GPT-2) due to the fundamental information bottleneck of 1-bit quantization, the high sparsity translates to significant energy savings on neuromorphic hardware, where zero-valued spikes require no memory access or MAC operations. To our knowledge, this work establishes the first industrially-aligned, multi-domain evaluation of a Pure SNN language model, providing a concrete baseline for the energy-accuracy Pareto frontier in ultra-low-power artificial general intelligence.

The increasing penetration of renewable energy generation (REG) introduces significant uncertainty into power grids, posing heightened risks for cascading failures. In this paper, a Markov tree model is proposed to assess the risk of cascading failure in power grid with uncertain REG. The model captures the diverse failure paths caused by REG uncertainty, representing the cascading failure process as a sequence of state transitions with probabilities reflecting the likelihood of state transitions. To identify critical tripping branches during cascading failure propagation, a hybrid probability-interval method is introduced. Probabilistic power flow analysis identifies branches with overload risk, while interval positional relationships rank their severity. To improve the efficiency of risk assessment, a risk-based depth-first search (R-DFS) method is proposed. This method uses estimated risk indices to prioritize high-risk failure paths while pruning low-risk paths, significantly reducing simulation time while maintaining assessment accuracy. Compared with existing models, the proposed model balances simulation efficiency and accuracy, effectively identifying high-risk failure paths under REG uncertainty. Simulation results demonstrate the impact of threshold selection on the retention of high-risk paths and simulation performance, providing insights into managing cascading failure risks in power grid with high REG penetration.

Embellishing slides with illustrations is a well-established practice for improving engagement and storytelling. However, this process is challenging, requiring careful consideration of both visual appearance and semantics of illustrations while ensuring they complement rather than overwhelm the slide content. In this paper, we take a pioneering step toward automating this process by introducing the task of Illustration Layout Generation: given a slide and a set of illustrations, automatically determining their optimal sizes and positions to enrich the slide. Existing layout generation approaches struggle with this task as they rely on large-scale layout datasets for training and have limited support for multiple visual inputs. To address these challenges, we propose SlideILG, a method that iteratively optimizes illustration placement using a diffusion-based text-to-image prior. We introduce three key techniques to enhance efficiency and quality: (1) leveraging cross-attention maps from the text-to-image model to initialize illustration placement; (2) employing an over-parameterization strategy to stabilize optimization; and (3) fine-tuning the text-to-image model on high-quality slide thumbnails for more precise guidance. To evaluate SlideILG, we construct IllustrationBench, a benchmark comprising 128 real-world slides, each paired with a set of illustrations for embellishment. Quantitative, qualitative and human-study results demonstrate the effectiveness of our approach. Furthermore, we showcase a real-world application scenario to highlight the significance and practical utility of this task and our method. Code and data are available at https://github.com/jzy124/SlideILG.