arXiv subjects

Weidong Shi

Publications and source records attributed to Weidong Shi.

At least 19 recordsLinked to original sources

arXiv2026

Reachability-Based Capability Confinement for LLM Agents under Indirect Prompt Injection

Large language model agents place outputs from external skills into their execution context, allowing attacker-controlled data to influence later privileged actions. Existing defenses mainly classify untrusted content or authorize proposed operations. They do not directly address how an agent's future authority should change once untrusted data enters its state. We present SkillGuard, a harness-level enforcement layer that treats this event as contamination and restricts future capabilities to disconnect the resulting state from deployer-defined forbidden states. Given sound skill summaries and policies, SkillGuard represents security-relevant transitions with a Skill Impact Graph, specifies admissible control over skill parameters via steerability signatures, and mediates invocations with an inline reference monitor. Following contamination, it computes weighted capability restrictions using binary, fractional, or fractional-flow strategies without auxiliary language-model inference. We evaluate SkillGuard on four AgentDojo suites with two backend LLMs, Gemini 2.5 Flash and Llama3.3-70B, against an LLM-only No Defense baseline and three defenses at different system layers: Spotlighting, CaMeL, and AttriGuard. We construct a compositional attack benchmark in which each attack combines observations individually insufficient to induce target violation and evaluate the same baselines on it. Under AgentDojo's Tool Knowledge attacks, SkillGuard eliminates attack success on three of four suites for both backends and reduces it to 4.8% and 14.3% on Slack. Against compositional attacks, it outperforms every baseline on Llama and matches the strongest baseline on Gemini at higher benign utility. Fractional-flow restriction preserves substantially more capabilities than binary restriction at the same attack success rate. Across both settings, SkillGuard adds no model calls or token overhead.

cs.CR↗

arXiv2026

Allocation Before Ranking: Decoupled Token Compression for OmniLLMs

Token compression in OmniLLMs is typically posed as a single saliency-ranking problem: score each multimodal token, keep the top-K. We argue this abstraction is mis-specified. The same attention score simultaneously decides two things: how much retained capacity each modality receives, and which tokens within a modality are kept. A shared top-K rule therefore inherits this audio-favoring allocation prior, spending retained capacity on audio before video tokens have a chance to compete. We propose Macer, a training-free compressor that first assigns explicit audio and video budgets, then performs allocation-normalized ranking within each modality at modality-specific shallow layers. Macer significantly reduces token cost while preserving accuracy across audio-grounded, audio--video joint, visual-dominant, and video-centric benchmarks. At 25 % retention, Macer preserves 98.7 % of full-token performance on Qwen2.5-Omni-7B and 97.3 % on Qwen2.5-Omni-3B. On Qwen2.5-Omni-7B, this 25 % setting reaches OmniZip-level performance at 45 % retention while using lower FLOPs. On OmniVinci-9B, the same allocation-before-ranking principle improves over shared top-K ranking by up to 12.9 points.

cs.AI↗

arXiv2026

Context Contamination in LLM Analysis of Network Security Logs: Poison with Passive Prompt Injection and Mitigation Evaluation

Large Language Models are increasingly deployed in Security Operations Centers for log analysis tasks including summarization, alert triage, and threat investigation. These systems ingest logs from external-facing services and process network logs as natural language contexts to generate security insights. We demonstrate that this architectural pattern introduces a critical vulnerability: adversaries can embed prompt injection payloads in log-generating fields that persist in storage and are executed when analysts query the LLM, achieving what we term passive prompt injection. We present LogInject, a systematic framework for evaluating these threats. Using LogInject-1.0, a benchmark of 12,847 log entries including 2,569 adversarial samples, we evaluate three production LLMs across four attack objectives: activity concealment, false positive generation, information exfiltration, and output hijacking. Our findings reveal an up to 88.2% attack success rate (83.4% average across models) under the baseline conditions. We introduce Context Stitching, a novel technique that fragments payloads across multiple log entries to evade stateless filters while exploiting LLM long-context reasoning, achieving a 76.4% success rate. As mitigation, we evaluate layered defenses by combining input filtering, prompt hardening, and output validation, demonstrating a 90.4% attack reduction, although 8.4% residual vulnerability persists. Our results establish that LLM-based log analysis creates an inherent confused deputy vulnerability where untrusted data and trusted instructions compete indistinguishably for model attention, requiring defense in-depth architectures and continued human oversight for security-critical decisions.

cs.CR↗

arXiv2026

CausalGraphX: A Counterfactual Graph Neural Network Framework for Explainable Systemic Risk Assessment

The interconnected nature of global financial systems makes them vulnerable to systemic risks, where the failure of a few institutions can trigger catastrophic cascading defaults. Traditional risk models often fail to capture the complex, non-linear dynamics of these networks. While Graph Neural Networks (GNNs) have shown promise in modeling relational data, they primarily learn correlative patterns and function as black boxes, offering little insight into the causal mechanisms of shock propagation. This limitation is critical for regulators who require explainable models to perform stress tests and devise effective interventions. We introduce CausalGraphX, a novel framework that integrates GNNs with counterfactual reasoning to provide explainable assessments of systemic risk. CausalGraphX employs a Graph Attention mechanism to learn representations of institutional vulnerability and uses an adversarial regularization technique to ensure these representations capture causal drivers rather than spurious correlations. Furthermore, we propose an optimization-based approach to generate counterfactual explanations, answering questions such as, "What minimum capital injection would have prevented Bank A's default under a specific stress scenario?" We validate CausalGraphX on large-scale synthetic financial networks. Our results demonstrate that CausalGraphX significantly outperforms traditional and deep learning baselines in predicting cascading defaults while providing sparse, plausible, and actionable counterfactual explanations.

cs.AI↗

arXiv2026

Rethinking Publication: A Certification Framework for AI-Enabled Research

AI research pipelines can now generate academic work that may satisfy existing peer review standards for quality, novelty, and methodological rigor. However, the publication system was built around the assumption that research is produced by human authors. It therefore lacks a clear way to evaluate work when the knowledge claim may be valid but the producer is partly or fully automated. This paper proposes a two-layer certification framework for AI-generated research. The first layer evaluates whether the knowledge claim is sound. The second layer evaluates the level of human contribution. This separation allows journals and conferences to assess pipeline-generated work more consistently without creating new institutions. The framework uses normative analysis, conceptual design, and dry-run validation against representative submission cases. It classifies human contribution into three categories: Category A, where the work is reachable by an automated pipeline; Category B, where human direction is required at identifiable stages; and Category C, where the work goes beyond current pipeline capability, especially at the problem-formulation stage. The paper also proposes dedicated benchmark slots for fully disclosed automated research. These slots would provide a transparent publication path and help reviewers calibrate judgments over time. The key argument is that publication has historically certified two things at once: that the knowledge is valid and that a human produced it. AI research pipelines separate these two claims. By decoupling knowledge certification from authorship attribution, the proposed framework responds to a structural change already underway. It can be implemented within existing editorial systems, works even when attribution is uncertain, and recognizes human frontier contribution based on epistemic value rather than human origin alone.

cs.AI↗

arXiv2026

zkRansomware: Proof-of-Data Recoverability and Multi-round Game Theoretic Modeling of Ransomware Decisions

Ransomware is still one of the most serious cybersecurity threats. Victims often pay but fail to regain access to their data, while also facing the danger of losing data privacy. These uncertainties heavily shape the attacker-victim dynamics in decision-making. In this paper, we introduce and analyze zkRansomware. This new ransomware model integrates zero-knowledge proofs to enable verifiable data recovery and uses smart contracts to enforce multi-round payments while mitigating the risk of data disclosure and privacy loss. We show that zkRansomware is technically feasible using existing cryptographic and blockchain tools and, perhaps counterintuitively, can align incentives between the attacker and the victim. Finally, we develop a theoretical decision-making framework for zkRansomware that distinguishes it from known ransomware decision models and discusses its implications for ransomware risk analysis and response decision support.

cs.CR↗

arXiv2025

Securing Smart Contract Languages with a Unified Agentic Framework for Vulnerability Repair in Solidity and Move

The rapid growth of the blockchain ecosystem and the increasing value locked in smart contracts necessitate robust security measures. While languages like Solidity and Move aim to improve smart contract security, vulnerabilities persist. This paper presents Smartify, a novel multi-agent framework leveraging Large Language Models (LLMs) to automatically detect and repair vulnerabilities in Solidity and Move smart contracts. Unlike traditional methods that rely solely on vast pre-training datasets, Smartify employs a team of specialized agents working on different specially fine-tuned LLMs to analyze code based on underlying programming concepts and language-specific security principles. We evaluated Smartify on a dataset for Solidity and a curated dataset for Move, demonstrating its effectiveness in fixing a wide range of vulnerabilities. Our results show that Smartify (Gemma2+codegemma) achieves state-of-the-art performance, surpassing existing LLMs and enhancing general-purpose models' capabilities, such as Llama 3.1. Notably, Smartify can incorporate language-specific knowledge, such as the nuances of Move, without requiring massive language-specific pre-training datasets. This work offers a detailed analysis of various LLMs' performance on smart contract repair, highlighting the strengths of our multi-agent approach and providing a blueprint for developing more secure and reliable decentralized applications in the growing blockchain landscape. We also provide a detailed recipe for extending this to other similar use cases.

cs.CR↗

arXiv2025

QuCoWE Quantum Contrastive Word Embeddings with Variational Circuits for NearTerm Quantum Devices

We present QuCoWE a framework that learns quantumnative word embeddings by training shallow hardwareefficient parameterized quantum circuits PQCs with a contrastive skipgram objective Words are encoded by datareuploading circuits with controlled ring entanglement similarity is computed via quantum state fidelity and passed through a logitfidelity head that aligns scores with the shiftedPMI scale of SGNSNoiseContrastive Estimation To maintain trainability we introduce an entanglementbudget regularizer based on singlequbit purity that mitigates barren plateaus On Text8 and WikiText2 QuCoWE attains competitive intrinsic WordSim353 SimLex999 and extrinsic SST2 TREC6 performance versus 50100d classical baselines while using fewer learned parameters per token All experiments are run in classical simulation we analyze depolarizingreadout noise and include errormitigation hooks zeronoise extrapolation randomized compiling to facilitate hardware deployment

quant-ph↗

arXiv2025

HPCAgentTester: A Multi-Agent LLM Approach for Enhanced HPC Unit Test Generation

Unit testing in High-Performance Computing (HPC) is critical but challenged by parallelism, complex algorithms, and diverse hardware. Traditional methods often fail to address non-deterministic behavior and synchronization issues in HPC applications. This paper introduces HPCAgentTester, a novel multi-agent Large Language Model (LLM) framework designed to automate and enhance unit test generation for HPC software utilizing OpenMP and MPI. HPCAgentTester employs a unique collaborative workflow where specialized LLM agents (Recipe Agent and Test Agent) iteratively generate and refine test cases through a critique loop. This architecture enables the generation of context-aware unit tests that specifically target parallel execution constructs, complex communication patterns, and hierarchical parallelism. We demonstrate HPCAgentTester's ability to produce compilable and functionally correct tests for OpenMP and MPI primitives, effectively identifying subtle bugs that are often missed by conventional techniques. Our evaluation shows that HPCAgentTester significantly improves test compilation rates and correctness compared to standalone LLMs, offering a more robust and scalable solution for ensuring the reliability of parallel software systems.

cs.DC↗

arXiv2025

Evaluating the Quality of Randomness and Entropy in Tasks Supported by Large Language Models

The rapid advancement of large language model (LLM) technology has led to diverse applications, many of which inherently require randomness, such as stochastic decision-making, gaming, scheduling, AI agents, and cryptography-related tasks. However, the capabilities of LLMs in handling randomness, particularly in generating and utilizing random numbers effectively, remain unclear. This paper investigates the capacity of LLMs for handling tasks that involve randomness through a series of experiments. We designed a set of experiments that consider various factors that can influence an LLM's performance in tasks involving randomness, such as accessibility to external tools, types of tasks, model states (fresh vs. non-fresh), and prompting strategies. The experiments cover a range of tasks, including generating random numbers, generating random strings such as passwords, shuffling items, and evaluating the quality of randomness using entropy and the NIST randomness test-suite. Our findings reveal that while LLMs can generate outputs that exhibit some degree of randomness, their performance is inconsistent and often deviates significantly from the expected behavior. The analysis of the experimental results highlights key limitations and areas where improvement is needed for the LLMs to effectively handle tasks involving randomness

cs.AI↗

arXiv2025

Adding All Flavors: A Hybrid Random Number Generator for dApps and Web3

Random numbers play a vital role in many decentralized applications (dApps), such as gaming and decentralized finance (DeFi) applications. Existing random number provision mechanisms can be roughly divided into two categories, on-chain, and off-chain. On-chain approaches usually rely on the blockchain as the major input and all computations are done by blockchain nodes. The major risk for this type of method is that the input itself is susceptible to the adversary's influence. Off-chain approaches, as the name suggested, complete the generation without the involvement of blockchain nodes and share the result directly with a dApp. These mechanisms usually have a strong security assumption and high complexity. To mitigate these limitations and provide a framework that allows a dApp to balance different factors involved in random number generation, we propose a hybrid random number generation solution that leverages IoT devices equipped with trusted execution environment (TEE) as the randomness sources, and then utilizes a set of cryptographic tools to aggregate the multiple sources and obtain the final random number that can be consumed by the dApp. The new approach only needs one honest random source to guarantee the unbiasedness of the final random number and a user can configure the system to tolerate malicious participants who can refuse to respond to avoid unfavored results. We also provide a concrete construction that can further reduce the on-chain computation complexity to lower the cost of the solution in practice. We evaluate the computation and gas costs to demonstrate the effectiveness of the improvement.

cs.CR↗

arXiv2025

Thermodynamically Consistent Modeling and Stable ALE Approximations of Reactive Semi-Permeable Interfaces

Reactive, semi-permeable interfaces play important roles in key biological processes such as targeted drug delivery, lipid metabolism, and signal transduction. These systems involve coupled surface reactions, transmembrane transport, and interfacial deformation, often triggered by local biochemical signals. The strong mechanochemical couplings complicate the modeling of such interfacial dynamics. We propose a thermodynamically consistent continuum framework that integrates bulk fluid motion, interfacial dynamics, surface chemistry, and selective solute exchange, derived via an energy variation approach to ensure mass conservation and energy dissipation. To efficiently solve the resulting coupled system, we develop a finite element scheme within an Arbitrary Lagrangian-Eulerian (ALE) framework, incorporating the Barrett-Garcke-Nurnberg (BGN) strategy to maintain mesh regularity and preserve conservation laws. Numerical experiments verify the convergence and conservation properties of the scheme and demonstrate its ability in capturing complex interfacial dynamics. Two biologically inspired examples showcase the model's versatility: cholesterol efflux via the ABCG1 pathway, involving multistage interfacial reactions and HDL uptake; and a self-propelled droplet system with reaction-activated permeability, mimicking drug release in pathological environments. This work provides a unified computational platform for studying strongly coupled biochemical and mechanical interactions at interfaces, offering new insights into reactive transport processes in both biological and industrial contexts.

math.NA↗

arXiv2025

From General Reasoning to Domain Expertise: Uncovering the Limits of Generalization in Large Language Models

Recent advancements in Large Language Models (LLMs) have demonstrated remarkable capabilities in various domains. However, effective decision-making relies heavily on strong reasoning abilities. Reasoning is the foundation for decision-making, providing the analytical and logical framework to make sound choices. Reasoning involves analyzing information, drawing inferences, and reaching conclusions based on logic or evidence. Decision-making builds on this foundation by applying the insights from reasoning to select the best course of action among alternatives. Together, these processes create a continuous cycle of thought and action aimed at achieving goals effectively. As AI technology evolves, there is a growing trend to train LLMs to excel in general reasoning. This study explores how the general reasoning capabilities of LLMs connect to their performance in domain-specific reasoning tasks.

cs.CL↗

arXiv2025

Governance Challenges in Reinforcement Learning from Human Feedback: Evaluator Rationality and Reinforcement Stability

Reinforcement Learning from Human Feedback (RLHF) is central in aligning large language models (LLMs) with human values and expectations. However, the process remains susceptible to governance challenges, including evaluator bias, inconsistency, and the unreliability of feedback. This study examines how the cognitive capacity of evaluators, specifically their level of rationality, affects the stability of reinforcement signals. A controlled experiment comparing high-rationality and low-rationality participants reveals that evaluators with higher rationality scores produce significantly more consistent and expert-aligned feedback. In contrast, lower-rationality participants demonstrate considerable variability in their reinforcement decisions ($p < 0.01$). To address these challenges and improve RLHF governance, we recommend implementing evaluator pre-screening, systematic auditing of feedback consistency, and reliability-weighted reinforcement aggregation. These measures enhance the fairness, transparency, and robustness of AI alignment pipelines.

cs.CY↗

arXiv2025

zkVC: Fast Zero-Knowledge Proof for Private and Verifiable Computing

In the context of cloud computing, services are held on cloud servers, where the clients send their data to the server and obtain the results returned by server. However, the computation, data and results are prone to tampering due to the vulnerabilities on the server side. Thus, verifying the integrity of computation is important in the client-server setting. The cryptographic method known as Zero-Knowledge Proof (ZKP) is renowned for facilitating private and verifiable computing. ZKP allows the client to validate that the results from the server are computed correctly without violating the privacy of the server's intellectual property. Zero-Knowledge Succinct Non-Interactive Argument of Knowledge (zkSNARKs), in particular, has been widely applied in various applications like blockchain and verifiable machine learning. Despite their popularity, existing zkSNARKs approaches remain highly computationally intensive. For instance, even basic operations like matrix multiplication require an extensive number of constraints, resulting in significant overhead. In addressing this challenge, we introduce \textit{zkVC}, which optimizes the ZKP computation for matrix multiplication, enabling rapid proof generation on the server side and efficient verification on the client side. zkVC integrates optimized ZKP modules, such as Constraint-reduced Polynomial Circuit (CRPC) and Prefix-Sum Query (PSQ), collectively yielding a more than 12-fold increase in proof speed over prior methods. The code is available at https://github.com/UCF-Lou-Lab-PET/zkformer

cs.CR↗

arXiv2025

Synthesizing Public Opinions with LLMs: Role Creation, Impacts, and the Future to eDemorcacy

This paper investigates the use of Large Language Models (LLMs) to synthesize public opinion data, addressing challenges in traditional survey methods like declining response rates and non-response bias. We introduce a novel technique: role creation based on knowledge injection, a form of in-context learning that leverages RAG and specified personality profiles from the HEXACO model and demographic information, and uses that for dynamically generated prompts. This method allows LLMs to simulate diverse opinions more accurately than existing prompt engineering approaches. We compare our results with pre-trained models with standard few-shot prompts. Experiments using questions from the Cooperative Election Study (CES) demonstrate that our role-creation approach significantly improves the alignment of LLM-generated opinions with real-world human survey responses, increasing answer adherence. In addition, we discuss challenges, limitations and future research directions.

cs.CL↗

arXiv2025

Collaboration is all you need: LLM Assisted Safe Code Translation

This paper introduces UniTranslator, a visionary framework that re-imagines code translation as a collaborative endeavor among multiple, compact LLMs. By orchestrating the interaction of specialized agents, each focused on different aspects of the translation process and grounded in a deep understanding of programming concepts, UniTranslator achieves a level of accuracy and efficiency that rivals larger, monolithic models. Our preliminary evaluation demonstrates the potential of UniTranslator to overcome the limitations of existing approaches and unlock the power of smaller LLMs for complex code translation tasks. We explore the effectiveness of this dynamic multi-agent paradigm in handling diverse language pairs, including low-resource languages, and in mitigating common issues such as code artifacts and hallucinations through the use of Natural Language Inference (NLI) grounding and iterative feedback mechanisms

cs.AI↗

arXiv2025

Weaving the Cosmos: WASM-Powered Interchain Communication for AI Enabled Smart Contracts

In this era, significant transformations in industries and tool utilization are driven by AI/Large Language Models (LLMs) and advancements in Machine Learning. There's a growing emphasis on Machine Learning Operations(MLOps) for managing and deploying these AI models. Concurrently, the imperative for richer smart contracts and on-chain computation is escalating. Our paper introduces an innovative framework that integrates blockchain technology, particularly the Cosmos SDK, to facilitate on-chain AI inferences. This system, built on WebAssembly (WASM), enables interchain communication and deployment of WASM modules executing AI inferences across multiple blockchain nodes. We critically assess the framework from feasibility, scalability, and model security, with a special focus on its portability and engine-model agnostic deployment. The capability to support AI on-chain may enhance and expand the scope of smart contracts, and as a result enable new use cases and applications.

cs.SE↗