PQLN: Post-Quantum Security for the Bitcoin Lightning Network's Off-Chain Surfaces
本文提出PQLN，一种基于格密码学的后量子安全方案，用于保护比特币闪电网络的离线表面，防止未来量子计算机带来的威胁。
0 citationsRead paper
Selected work
Representative Papers
Empirical Software Engineering TerraProbe: A Layered-Oracle Framework for Detecting Deceptive Fixes in LLM-Assisted Terraform
This work addresses the prevalence of “deceptive fixes” in large language model (LLM)-assisted repair of Terraform security configurations—patches that appear successful under static analysis yet fail to eliminate vulnerabilities or inadvertently compromise infrastructure semantics. To tackle this, the authors propose TerraProbe, the first five-tier oracle framework for evaluating Terraform repairs, integrating static scanning, plan validation, behavioral consistency, security intent alignment, and human adjudication. They also introduce a four-dimensional taxonomy to characterize deceptive fixes. Evaluating 288 samples, they find that while 83.3% pass the target checker, only 10.4% satisfy comprehensive scanning, 39.6% yield valid execution plans, and 71.4% are classified as deceptive. Notably, three leading LLMs show no significant performance differences, underscoring the pervasiveness of the issue. The study provides a reproducible evaluation methodology and an open-source toolkit.
0 citationsRead paper
Specificity- and Calibration-Aware Breast Ultrasound Segmentation via Entropy-Guided Boundary Supervision
This work addresses the challenges posed by speckle noise, low contrast, and acoustic shadowing in breast ultrasound images, which often lead to ambiguous lesion boundaries and false-positive activations. The authors propose an entropy-guided boundary supervision loss that weights gradient updates using the predicted entropy map and ground-truth boundary annotations, thereby focusing learning on regions of high edge uncertainty. Additionally, spatial temperature scaling is introduced during inference to improve probability calibration. Implemented within a U-Net architecture, the method significantly reduces the false-positive rate on lesion-free images in the BUSI dataset from 14/20 and 19/20 to 5/20 (p<0.05), lowers the expected calibration error from 0.0201 to 0.0095, and maintains a stable Dice score (≈0.762), effectively balancing segmentation accuracy and specificity.
0 citationsRead paper
LNTest: A Testbed for Evaluating Bitcoin Lightning Network-Based Botnets
This study addresses the lack of empirical evaluation on Bitcoin Lightning Network (LN)-based botnets regarding their real-world topology formation, command propagation complexity, and resilience to takedown strategies. To bridge this gap, we present LNTest—the first reusable testbed for LN botnets—built upon Dockerized Core Lightning nodes operating on a Bitcoin Core regtest chain, supporting deterministic chain, autonomous discovery, and user-defined overlay topologies. Our experiments reveal, for the first time, that D-LNBot naturally forms clustered chain-like structures in practice. We correct the command propagation complexity from the previously assumed O(m log n) to Θ(n) and systematically quantify the robustness of different topologies against both targeted and random node removal. Notably, the autonomously formed clustered chain exhibits the highest vulnerability among all evaluated configurations.
0 citationsRead paper
A Cost-Effective Framework for Predicting Parking Availability Using Geospatial Data and Machine Learning
To address the scarcity of parking resources and the challenge of dynamic supply-demand matching in university campuses, this paper proposes a hardware-free, lightweight parking availability prediction framework. Methodologically, it integrates heterogeneous spatiotemporal data—including street networks, mobility trajectories, and meteorological records—models vehicle parking behavior via spatial join operations, and evaluates four predictive models: linear regression, support vector regression (SVR), random forest, and long short-term memory (LSTM) networks, optimized via grid search and assessed using RMSE, MAE, and R². Key contributions include: (i) the first sensor-free approach leveraging geospatial operations to uncover implicit parking patterns; and (ii) the incorporation of multi-scale spatiotemporal features to significantly enhance prediction robustness. Experimental results show that the random forest model achieves the best overall performance (RMSE = 0.142, R² = 0.582), while LSTM demonstrates greater potential for improvement in long-horizon forecasting scenarios.
0 citationsRead paper
Recent publications
Latest Papers
PQLN: Post-Quantum Security for the Bitcoin Lightning Network's Off-Chain Surfaces
本文提出PQLN，一种基于格密码学的后量子安全方案，用于保护比特币闪电网络的离线表面，防止未来量子计算机带来的威胁。
0 citationsRead paper
Empirical Software Engineering TerraProbe: A Layered-Oracle Framework for Detecting Deceptive Fixes in LLM-Assisted Terraform
This work addresses the prevalence of “deceptive fixes” in large language model (LLM)-assisted repair of Terraform security configurations—patches that appear successful under static analysis yet fail to eliminate vulnerabilities or inadvertently compromise infrastructure semantics. To tackle this, the authors propose TerraProbe, the first five-tier oracle framework for evaluating Terraform repairs, integrating static scanning, plan validation, behavioral consistency, security intent alignment, and human adjudication. They also introduce a four-dimensional taxonomy to characterize deceptive fixes. Evaluating 288 samples, they find that while 83.3% pass the target checker, only 10.4% satisfy comprehensive scanning, 39.6% yield valid execution plans, and 71.4% are classified as deceptive. Notably, three leading LLMs show no significant performance differences, underscoring the pervasiveness of the issue. The study provides a reproducible evaluation methodology and an open-source toolkit.
0 citationsRead paper
Specificity- and Calibration-Aware Breast Ultrasound Segmentation via Entropy-Guided Boundary Supervision
This work addresses the challenges posed by speckle noise, low contrast, and acoustic shadowing in breast ultrasound images, which often lead to ambiguous lesion boundaries and false-positive activations. The authors propose an entropy-guided boundary supervision loss that weights gradient updates using the predicted entropy map and ground-truth boundary annotations, thereby focusing learning on regions of high edge uncertainty. Additionally, spatial temperature scaling is introduced during inference to improve probability calibration. Implemented within a U-Net architecture, the method significantly reduces the false-positive rate on lesion-free images in the BUSI dataset from 14/20 and 19/20 to 5/20 (p<0.05), lowers the expected calibration error from 0.0201 to 0.0095, and maintains a stable Dice score (≈0.762), effectively balancing segmentation accuracy and specificity.
0 citationsRead paper
LNTest: A Testbed for Evaluating Bitcoin Lightning Network-Based Botnets
This study addresses the lack of empirical evaluation on Bitcoin Lightning Network (LN)-based botnets regarding their real-world topology formation, command propagation complexity, and resilience to takedown strategies. To bridge this gap, we present LNTest—the first reusable testbed for LN botnets—built upon Dockerized Core Lightning nodes operating on a Bitcoin Core regtest chain, supporting deterministic chain, autonomous discovery, and user-defined overlay topologies. Our experiments reveal, for the first time, that D-LNBot naturally forms clustered chain-like structures in practice. We correct the command propagation complexity from the previously assumed O(m log n) to Θ(n) and systematically quantify the robustness of different topologies against both targeted and random node removal. Notably, the autonomously formed clustered chain exhibits the highest vulnerability among all evaluated configurations.
0 citationsRead paper
A Cost-Effective Framework for Predicting Parking Availability Using Geospatial Data and Machine Learning
To address the scarcity of parking resources and the challenge of dynamic supply-demand matching in university campuses, this paper proposes a hardware-free, lightweight parking availability prediction framework. Methodologically, it integrates heterogeneous spatiotemporal data—including street networks, mobility trajectories, and meteorological records—models vehicle parking behavior via spatial join operations, and evaluates four predictive models: linear regression, support vector regression (SVR), random forest, and long short-term memory (LSTM) networks, optimized via grid search and assessed using RMSE, MAE, and R². Key contributions include: (i) the first sensor-free approach leveraging geospatial operations to uncover implicit parking patterns; and (ii) the incorporation of multi-scale spatiotemporal features to significantly enhance prediction robustness. Experimental results show that the random forest model achieves the best overall performance (RMSE = 0.142, R² = 0.582), while LSTM demonstrates greater potential for improvement in long-horizon forecasting scenarios.
0 citationsRead paper