arXiv subjects

Weiyi Wang

Publications and source records attributed to Weiyi Wang.

At least 19 recordsLinked to original sources

arXiv2026

How Powerful are LLMs in Generating Formal Program Specifications?

Formal verification provides strong guarantees of software correctness, but its adoption is limited by the high cost of writing precise formal specifications. While recent large language models (LLMs) have shown strong capabilities in theorem proving and verified code generation, their true ability to generate program specifications remains unclear. Existing evaluations require either verifying implementation conformance or proving semantic equivalence between specifications, both of which are formidably difficult and may conflate proof difficulty with specification quality. To address this problem, we introduce Coins, a Rocq based evaluation framework that assesses specification quality by instantiating specifications under evaluation on trusted test cases and generating concrete proof obligations. This design aligns with the asymmetric nature of formal reasoning, where successful proofs provide reliable evidence while proof failures are inherently ambiguous. Using Coins, we conduct a large scale study on HumanEval with a curated set of human written Rocq specifications. Our results show that specification generation remains a formidable challenge, and that verification complexity can obscure genuine differences in specification quality. Overall, we find that accurate specification evaluation, rather than model scaling alone, is central to understanding the power of LLMs for specification synthesis, and that test case based formal reasoning offers a more faithful and discriminative measure of progress.

cs.SE↗

arXiv2026

Rethinking Scientific Discovery in the Agentic Era

Artificial intelligence has advanced scientific discovery, but most AI4Science systems remain fragmented tools that rely on humans to coordinate problem formulation, literature grounding, model use, simulation, validation, and knowledge reuse. This paper presents \textbf{SCION (Scientific Collaborative Innovation with Agentic Organizational Nexus)}, an agentic scientific operating system that acts as an \textbf{organizational nexus}. Through a Science Agent serving as a \textbf{Meta-Harness}, SCION connects scientific tasks, tools, agents, artifacts, and memory, transforming research into an executable, auditable, and reusable operational process. At its core is the \textbf{Research Execution Plan (REP)}, which compiles high-level scientific intent into staged objectives, dependencies, verification checkpoints, tool requirements, expected artifacts, and fallback conditions. SCION further integrates hierarchical multi-agent execution, profile-driven specialization, selective context construction, governed delegation, and layered epistemic memory to support long-horizon scientific work. We formulate discovery under SCION as \textbf{Target-conditioned Inverse Search} and extend it to hidden-target settings through batch active search under finite experimental budgets. Applications in materials analysis, molecule design, and protein or antibody screening, together with experiments on scientific reading, idea generation, molecule generation, and antibody screening, show that SCION outperforms existing autonomous research-agent baselines, especially in decomposition, verification, refinement, and memory reuse. Overall, SCION shifts AI from isolated tools toward a coordinated operational layer for traceable and reusable scientific innovation.

cs.CL↗

arXiv2026

ClawGuard: Out-of-Band Detection of LLM Agent Workflow Hijacking via EM Side Channel

Autonomous LLM agents face a critical security risk known as workflow hijacking, where attackers subtly alter tool and skill invocations. Existing defenses rely on host-internal telemetry (such as audit logs), which can be forged if the host OS is compromised. To solve this, we introduce ClawGuard, a passive, out-of-band monitor that audits LLM-agent workflows using electromagnetic (EM) emanations. Because distinct agent skills create unique hardware usage patterns (computation, DRAM, network blocking), they emit measurable, macroscopic EM envelopes. External software-defined radios (SDRs) capture these physical signals. Using a drift-aware pipeline with 320-dimensional features, ClawGuard converts RF streams into physical evidence. Evaluated on a 7.82TB RF corpus, ClawGuard achieved an AUC of 0.9945, detecting attacks with a 100% true-positive rate and a 1.16% false-positive rate. This proves passive EM sensing is a practical, forge-resistant physical check against compromised host software.

cs.CR↗

arXiv2026

Quasi-one-dimensional spin excitations in the iron pnictide NaFe$_{0.53}$Cu$_{0.47}$As

Spectroscopic measurements in model one-dimensional (1D) correlated systems offer insights for understanding their two-dimensional counterparts, which include the cuprate and iron pnictide/chalcogenide superconductors. A major challenge is the identification of such correlated systems with dominantly 1D physics. In this work, inelastic neutron scattering measurements on NaFe$_{0.53}$Cu$_{0.47}$As single crystal directly reveal quasi-1D spin excitations, resulting from atomic order that lead to magnetic Fe and nonmagnetic Cu chains. The dominant exchange interaction is antiferromagnetic along the chain ($SJ_{\rm \parallel}\approx90.1(3)$~meV), whereas the inter-chain couplings are much weaker ($SJ_{\rm \perp}\approx-2.4(1)$~meV and $SJ_{\rm c}\approx0.15(5)$~meV). The quasi-1D spin excitations in NaFe$_{0.53}$Cu$_{0.47}$As stem from both the Néel and stripe vectors, with Néel excitations sensitive to Fe impurities on the Cu site. The spin excitations in quasi-1D NaFe$_{0.53}$Cu$_{0.47}$As and quasi-2D FeSe exhibit a striking resemblance, suggesting a common origin for their coexistent stripe and Néel excitations. Our findings demonstrate magnetic dilution in NaFeAs leads to dimension reduction of its magnetic degree of freedom, presenting a strategy for discovering low-dimensional quantum materials.

cond-mat.str-el↗

arXiv2026

AstroReason-Bench: Evaluating Unified Agentic Planning across Heterogeneous Space Planning Problems

Recent advances in agentic Large Language Models (LLMs) have positioned them as generalist planners capable of reasoning and acting across diverse tasks. However, existing agent benchmarks largely focus on symbolic or weakly grounded environments, leaving their performance in physics-constrained real-world domains underexplored. We introduce AstroReason-Bench, a comprehensive benchmark for evaluating agentic planning in Space Planning Problems (SPP), a family of high-stakes problems with heterogeneous objectives, strict physical constraints, and long-horizon decision-making. AstroReason-Bench integrates multiple scheduling regimes, including ground station communication and agile Earth observation, and provides a unified agent-oriented interaction protocol. Evaluating on a range of state-of-the-art open- and closed-source agentic LLM systems, we find that current agents substantially underperform specialized solvers, highlighting key limitations of generalist planning under realistic constraints. AstroReason-Bench offers a challenging and diagnostic testbed for future agentic research.

cs.AI↗

arXiv2026

Strain Engineering of Intrinsic Anomalous Hall and Nernst Effects in Altermagnetic MnTe at Realistic Doping Levels

Hexagonal MnTe has emerged as a prototypical g-wave altermagnet, hosting time-reversal symmetry breaking in momentum space despite a vanishing net magnetization. While this symmetry breaking theoretically allows for an intrinsic anomalous Hall effect, experimentally observed signals have remained weak. In this work, we investigate the origin of this suppression and demonstrate a strategy to amplify anomalous transport responses within the experimentally accessible doping regime. Using a $\bm{k}\cdot\bm{p}$ effective model, we reveal that near the valence band maximum, which corresponds to the energy window relevant for typical hole doping ($\sim10^{19}cm^{-3}$), the intrinsic Hall effect is suppressed due to a symmetry-enforced cancellation of opposing Berry curvature contributions. We propose that breaking the crystalline symmetry via volume-conserving biaxial strain lifts this cancellation, resulting in a significant enhancement of the anomalous Hall conductivity by orders of magnitude. This strain-induced Fermi surface distortion also amplifies the anomalous Nernst effect. Furthermore, the analysis of the spin texture confirms that these strain-enabled anomalous transport signatures emerge while preserving the zero net magnetization.

cond-mat.mtrl-sci↗

arXiv2025

The Wireless Charger as a Gesture Sensor: A Novel Approach to Ubiquitous Interaction

Advancements in information technology have increased demand for natural human-computer interaction in areas such as gaming, smart homes, and vehicles. However, conventional approaches like physical buttons or cameras are often limited by contact requirements, privacy concerns, and high costs.Motivated by the observation that these EM signals are not only strong and measurable but also rich in gesture-related information, we propose EMGesture, a novel contactless interaction technique that leverages the electromagnetic (EM) signals from Qi wireless chargers for gesture recognition. EMGesture analyzes the distinctive EM features and employs a robust classification model. The end-to-end framework enables it capable of accurately interpreting user intent. Experiments involving 30 participants, 10 mobile devices, and 5 chargers showed that EMGesture achieves over 97% recognition accuracy. Corresponding user studies also confirmed higher usability and convenience, which demonstrating that EMGesture is a practical, privacy-conscious, and cost-effective solution for pervasive interaction.

cs.HC↗

arXiv2025

EmbeddingGemma: Powerful and Lightweight Text Representations

We introduce EmbeddingGemma, a new lightweight, open text embedding model based on the Gemma 3 language model family. Our innovative training recipe strategically captures knowledge from larger models via encoder-decoder initialization and geometric embedding distillation. We improve model robustness and expressiveness with a spread-out regularizer, and ensure generalizability by merging checkpoints from varied, optimized mixtures. Evaluated on the Massive Text Embedding Benchmark (MTEB) across multilingual, English, and code domains, EmbeddingGemma (300M) achieves state-of-the-art results. Notably, it outperforms prior top models, both proprietary and open, with fewer than 500M parameters, and provides performance comparable to models double its size, offering an exceptional performance-to-cost ratio. Remarkably, this lead persists when quantizing model weights or truncating embedding outputs. This makes EmbeddingGemma particularly well-suited for low-latency and high-throughput use cases such as on-device applications. We provide ablation studies exploring our key design choices. We release EmbeddingGemma to the community to promote further research.

cs.CL↗

arXiv2025

Influence of Network Topology and Vaccination Strategies on HPV Dynamics: A Simulation Study Using the SeCoNet Growth Model

This study examines how contact network topology influences the effectiveness of vaccination programs in the context of human papillomavirus (HPV) transmission. Using the SeCoNet sexual contact network growth model, we evaluate age based, ring based, and several centrality based vaccination strategies across the overall, male, and female cohorts, focusing on peak incidence, timing of peak prevalence, and cumulative incidence. The simulations show that degree, betweenness, and percolation centrality based strategies are generally the most effective, while ring vaccination achieves the greatest reduction in cumulative incidence among females. Network topology also plays a critical role: higher average degree reduces vaccination effectiveness, whereas higher power-law exponent, longer average shortest path length, and stronger clustering improve vaccination outcomes. The results highlight the importance of incorporating network structure into the design of HPV vaccination programs.

cs.SI↗

arXiv2025

Taming Hyperparameter Sensitivity in Data Attribution: Practical Selection Without Costly Retraining

Data attribution methods, which quantify the influence of individual training data points on a machine learning model, have gained increasing popularity in data-centric applications in modern AI. Despite a recent surge of new methods developed in this space, the impact of hyperparameter tuning in these methods remains under-explored. In this work, we present the first large-scale empirical study to understand the hyperparameter sensitivity of common data attribution methods. Our results show that most methods are indeed sensitive to certain key hyperparameters. However, unlike typical machine learning algorithms -- whose hyperparameters can be tuned using computationally-cheap validation metrics -- evaluating data attribution performance often requires retraining models on subsets of training data, making such metrics prohibitively costly for hyperparameter tuning. This poses a critical open challenge for the practical application of data attribution methods. To address this challenge, we advocate for better theoretical understandings of hyperparameter behavior to inform efficient tuning strategies. As a case study, we provide a theoretical analysis of the regularization term that is critical in many variants of influence function methods. Building on this analysis, we propose a lightweight procedure for selecting the regularization value without model retraining, and validate its effectiveness across a range of standard data attribution benchmarks. Overall, our study identifies a fundamental yet overlooked challenge in the practical application of data attribution, and highlights the importance of careful discussion on hyperparameter selection in future method development.

cs.LG↗

arXiv2025

Interleaved Learning and Exploration: A Self-Adaptive Fuzz Testing Framework for MLIR

MLIR (Multi-Level Intermediate Representation) has rapidly become a foundational technology for modern compiler frameworks, enabling extensibility across diverse domains. However, ensuring the correctness and robustness of MLIR itself remains challenging. Existing fuzzing approaches-based on manually crafted templates or rule-based mutations-struggle to generate sufficiently diverse and semantically valid test cases, making it difficult to expose subtle or deep-seated bugs within MLIR's complex and evolving code space. In this paper, we present FLEX, a novel self-adaptive fuzzing framework for MLIR. FLEX leverages neural networks for program generation, a perturbed sampling strategy to encourage diversity, and a feedback-driven augmentation loop that iteratively improves its model using both crashing and non-crashing test cases. Starting from a limited seed corpus, FLEX progressively learns valid syntax and semantics and autonomously produces high-quality test inputs. We evaluate FLEX on the upstream MLIR compiler against four state-of-the-art fuzzers. In a 30-day campaign, FLEX discovers 80 previously unknown bugs-including multiple new root causes and parser bugs-while in 24-hour fixed-revision comparisons, it detects 53 bugs (over 3.5x as many as the best baseline) and achieves 28.2% code coverage, outperforming the next-best tool by 42%. Ablation studies further confirm the critical role of both perturbed generation and diversity augmentation in FLEX's effectiveness.

cs.SE↗

arXiv2025

Boosting Zero-Shot VLN via Abstract Obstacle Map-Based Waypoint Prediction with TopoGraph-and-VisitInfo-Aware Prompting

With the rapid progress of foundation models and robotics, vision-language navigation (VLN) has emerged as a key task for embodied agents with broad practical applications. We address VLN in continuous environments, a particularly challenging setting where an agent must jointly interpret natural language instructions, perceive its surroundings, and plan low-level actions. We propose a zero-shot framework that integrates a simplified yet effective waypoint predictor with a multimodal large language model (MLLM). The predictor operates on an abstract obstacle map, producing linearly reachable waypoints, which are incorporated into a dynamically updated topological graph with explicit visitation records. The graph and visitation information are encoded into the prompt, enabling reasoning over both spatial structure and exploration history to encourage exploration and equip MLLM with local path planning for error correction. Extensive experiments on R2R-CE and RxR-CE show that our method achieves state-of-the-art zero-shot performance, with success rates of 41% and 36%, respectively, outperforming prior state-of-the-art methods.

cs.RO↗

arXiv2024

Assessing the effectiveness of test-trace-isolate interventions using a multi-layered temporal network

In the early stage of an infectious disease outbreak, public health strategies tend to gravitate towards non-pharmaceutical interventions (NPIs) given the time required to develop targeted treatments and vaccines. One of the most common NPIs is Test-Trace-Isolate (TTI). One of the factors determining the effectiveness of TTI is the ability to identify contacts of infected individuals. In this study, we propose a multi-layer temporal contact network to model transmission dynamics and assess the impact of different TTI implementations, using SARS-CoV-2 as a case study. The model was used to evaluate TTI effectiveness both in containing an outbreak and mitigating the impact of an epidemic. We estimated that a TTI strategy based on home isolation and testing of both primary and secondary contacts can contain outbreaks only when the reproduction number is up to 1.3, at which the epidemic prevention potential is 88.2% (95% CI: 87.9%-88.5%). On the other hand, for higher value of the reproduction number, TTI is estimated to noticeably mitigate disease burden but at high social costs (e.g., over a month in isolation/quarantine per person for reproduction numbers of 1.7 or higher). We estimated that strategies considering quarantine of contacts have a larger epidemic prevention potential than strategies that either avoid tracing contacts or require contacts to be tested before isolation. Combining TTI with other social distancing measures can improve the likelihood of successfully containing an outbreak but the estimated epidemic prevention potential remains lower than 50% for reproduction numbers higher than 2.1.

q-bio.QM↗

arXiv2024

Computational Approaches of Modelling Human Papillomavirus Transmission and Prevention Strategies: A Systematic Review

Human papillomavirus (HPV) infection is the most common sexually transmitted infection in the world. Persistent oncogenic Human papillomavirus infection has been a leading threat to global health and can lead to serious complications such as cervical cancer. Prevention interventions including vaccination and screening have been proved effective in reducing the risk of HPV-related diseases. In recent decades, computational epidemiology has been serving as a very useful tool to study HPV transmission dynamics and evaluation of prevention strategies. In this paper, we conduct a comprehensive literature review on state-of-the-art computational epidemic models for HPV disease dynamics, transmission dynamics, as well as prevention efforts. We summarise current research trends, identify gaps in the present literature, and identify future research directions with potential in accelerating the containment and/or elimination of HPV infection.

q-bio.PE↗

arXiv2024

Flow regimes and repose angle in a rotating drum filled with highly concave particles

We present a series of experiments investigating the flow regimes and repose angles of highly concave particle packings in a rotating drum. By varying grain geometry from spherical to highly non-convex shapes, adjusting frictional properties and the particle number of branches, we examine how these parameters and the drum speed influence the flow behavior. Our study identifies two distinct flow regimes: the rolling regime, where granular matter exhibits solid-like behavior near the walls and flows like a liquid near the free surface, and the slumping regime, characterized by cyclic avalanches and solid body rotations. Using quantitative criteria such as the repose angle difference and the area ratio of particle packings, we construct phase diagrams delineating the cross-over between these regimes. Our findings highlight the significant effects of particle concavity, friction, and rotation speed on the flow dynamics of granular materials, providing new insights into the mechanical behaviors of \emph{meta-granular matter}.

cond-mat.soft↗

arXiv2024

Steady Granular Flow in a Rotating Drum: Universal description of stress, velocity and packing fraction profiles covering grain shape effects from convex to very concave

The flow behavior of granular matter is significantly influenced by the shape of constituent particles. This effect is particularly pronounced for very concave particles, which exhibit unique flow characteristics such as higher porosity and sharper phase transitions between jamming and unjamming states. Despite the richness and ubiquitousness of these systems, our understanding of their intricate flow behavior and the local mechanisms driving these behaviors remains incomplete. In this work, we investigate the effect of particle shape, ranging from spherical to highly concave, on steady flows in a rotating drum - a system that facilitates a continuous phase transition from a jamming state at greater depths to an unjamming state at shallower regions. We develop an analytical model to elucidate granular behavior within the rotating drum: (i) Firstly, by decomposing the shear stress, we reconcile the discrepancy between simulation data and theoretical predictions, establishing a relationship with the angle of repose. (ii)Secondly, we extend the generalized Bagnold scaling , coupled with a non-local fluidity relation based on packing fraction, providing a framework for a correlation between shear stress, shear rate, and packing fraction. Additionally, we introduce a characteristic length to quantify the influence of particle shape and drum speed. This analytical model offers explicit functional forms for physical quantity profiles, which are validated experimentally in a thin rotating drum and numerically in a two-dimensional rotating drum. Our results demonstrate that this model accurately describes the change of velocity due to the phase transition of granular flow within a rotating drum. Moreover, for different shapes of particle and drum speeds, the characteristic length captures the interplay between shear stress, shear rate, and the variation of packing fraction.

cond-mat.soft↗

arXiv2024

Dual Auction Mechanism for Transaction Relay and Validation in Complex Wireless Blockchain Network

In traditional public blockchain networks, transaction fees are only allocated to full nodes (i.e., miners). However, the lack of relay rewards reduces the willingness of light nodes to relay transactions, especially in the energy-constrained complex wireless network. This paper proposes a novel dual auction mechanism to allocate transaction fees for relay and validation behaviors in the wireless blockchain network. The dual auction mechanism consists of two sub-auction stages: the relay sub-auction and the validation sub-auction. In the relay sub-auction, relay nodes choose transactions based rewards to forward. Besides, relay nodes adjust the relaying probability through a no-regret algorithm to improve efficiency. In the validation sub-auction, full nodes select transactions using Vickrey-Clarke-Grove (VCG) mechanism to construct the block. We prove that the designed dual auction mechanism is Incentive Compatibility (IC), Individual Rationality (IR), and Computational Efficiency (CE). We also derive the upper bound of the social welfare difference between the social optimal auction and our proposed one. Extensive simulation results demonstrate that the proposed dual auction mechanism decreases energy and bandwidth resource consumption and effectively improves social welfare without sacrificing the throughput and the security of the wireless blockchain network.

cs.NI↗

arXiv2023

SeCoNet: A Heterosexual Contact Network Growth Model for Human Papillomavirus Disease Simulation

Human Papillomavirus infection is the most common sexually transmitted infection, and causes serious complications such as cervical cancer in vulnerable female populations in regions such as East Africa. Due to the scarcity of empirical data about sexual relationships in varying demographics, computationally modelling the underlying sexual contact networks is important to understand Human Papillomavirus infection dynamics and prevention strategies. In this work we present SeCoNet, a heterosexual contact network growth model for Human Papillomavirus disease simulation. The growth model consists of three mechanisms that closely imitate real-world relationship forming and discontinuation processes in sexual contact networks. We demonstrate that the networks grown from this model are scale-free, as are the real world sexual contact networks, and we demonstrate that the model can be calibrated to fit different demographic contexts by using a range of parameters. We also undertake disease dynamics analysis of Human Papillomavirus infection using a compartmental epidemic model on the grown networks. The presented SeCoNet growth model is useful to computational epidemiologists who study sexually transmitted infections in general and Human Papillomavirus infection in particular.

cs.SI↗