Vulnerabilities in smart contracts may trigger serious security events, and the detection of contract vulnerabilities has become a significant problem. In this paper, to solve limitations current deep learning-based vulnerability methods extracting various code critical features, using multi-scale cascade encoder architecture as backbone, we propose novel Multi-Scale Encoder Vulnerability Detection (MEVD) approach hit well-known high-risk contracts. Firstly, use gating mechanism design...

Though the performance of belief propagation (BP) decoder for polar codes is comparable with successive cancellation (SC) decoder, it has significant gap when compared improved SC decoders, such as list (SCL) decoding. In this paper, we propose an BP decoding good by adapting their parity-check matrices. The process iterative and consists two parts. Firstly, matrix adjusted that one its submatrices corresponding to less reliable bits in a sparse nature. Secondly, applied matrix. Simulation...

Vulnerabilities in smart contracts may trigger serious security events, and the detection of contract vulnerabilities has become a significant problem. In this paper, by using multi-scale cascade encoder architecture as backbone, we propose novel Multi-scale Encoder Vulnerability Detection (MEVD) approach to detect well-known high-risk contracts. Firstly, use gating mechanism design unique Surface Feature (SFE) enrich semantic information code features. Then, combining Base Transformer (BTE)...

The trend prediction of software vulnerability can provide valuable threat intelligence in security event prevention. It is a challenging task for highly accurate prediction. To address this problem, novel method, STL-EEMD-ARIMA, proposed, by incorporating Seasonal and Trend Decomposition using Loess (STL), Ensemble Empirical Mode (EEMD), Autoregressive Integrated Moving Average (ARIMA). Firstly, the random fluctuation components are extracted from samples STL respectively. Secondly, we use...

With the rapid growth of open-source software, code cloning has become increasingly prevalent. If there are security vulnerabilities in a cloned segment, those may spread related software to potentially lead incidents. The existing methods vulnerable detection performed on condition that source is converted into an intermediate representation. However, these do not fully consider rich semantic knowledge and patch information available for codes, which can induce high false positive rate...

As for current datasets of vehicle 3D points cloud, there are some shortcomings like lack large vehicles, sparseness point cloud data and insufficiency parts information. In this paper, a high-density dataset vehicles named ZPVehicles(Points by Zhejiang Institute Metrology) is provided. The original ZPVehicles collected LAMIoVP(Lidar-based Automatic Measuring Instrument Vehicle Profile), which developed us installed in testing station Hangzhou, China. contains the about 800vehicles, consist...

Based on the low rank representation (LRR) and sparse (SR), a composite LRR with SR graph LRRSR for semi‐supervised label propagation is proposed. The aims to capture both global structure of data by constraint local simultaneously. A framework applied fuse two graphs. Then, used transmit labels from labelled samples unlabelled samples. It several face image datasets experimental results demonstrate its good performance classification limited number

Vulnerability detection technology has become a hotspot in the field of software security, and most current methods do not have complete consideration during code characterizing, which leads to problems such as information loss. Therefore, this paper proposes one class Scalable Feature Network (SFN), composite feature extraction method based on Continuous Bag Words Convolutional Neural Network. In addition, characterize source more comprehensively, we construct multiscale metrics terms...

Error-correcting codes on permutations and multi-permutations have recently attracted much attention for their applications in flash memories through rank modulation. We propose a new demodulation method to transform the problem of single burst unstable erasure (BUE) t-balanced multi-permutation into sub-problems permutation-invariant (PIE) t permutations. Based method, we two classes correct BUE length up by interleaving single-PIE-correcting permutation codes. The decoding methods proposed...

Despite the work done by academia and industry in area of vulnerability detection, number vulnerabilities reported each year is still growing rapidly. Nowadays, mainstream detection methods mainly transform source code into an intermediate representation then detect it machine learning or deep algorithms, but problem high false-positive rate negative difficult to solve. This because does not reflect characteristics well. In this paper, we propose a convolutional neural network-based feature...

In this letter, a novel collaborative representation graph based on the local and global consistency label propagation method, denoted as CRLGC, is proposed. The used to reduce cost time in obtaining which evaluates similarity of samples. Considering lacking labeled samples real applications, semi-supervised method utilized transmit labels from unlabeled Experimental results three image data sets have demonstrated that proposed provides best accuracies most times when compared with other...

In this letter, we present a new scheme to find small fundamental instantons (SFIs) of regular low-density parity-check (LDPC) codes for the linear programming (LP) decoding over binary symmetric channel (BSC). Based on fact that each instanton-induced graph (IIG) contains at least one short cycle, determine potential by constructing possible IIGs which contain cycles and additional paths connected cycles. Then identify actual from ones under LP decoding. Simulation results some typical LDPC...

The symbol flipping decoding algorithms based on prediction (SFDP) for non-binary LDPC codes perform well in terms of error performances but converge slowly when compared to other algorithms. In order improve the convergence rate, we design new rules with two phases SFDP first phase, or more symbols are flipped at each iteration allow a quick increase objective function. While second only one is avoid oscillation decoder function close its maximum. Simulation results show that proposed can...

In this letter, a new method is presented to suppress fractional pseudocodewords by eliminating small instantons of irregular low-density parity-check (LDPC) codes under the linear programming (LP) decoding over binary symmetric channel (BSC). By appending several rows found integer formulation original matrix, optimal distribution spectrum BSC-instantons in modified code obtained. Simulation results show that proposed can improve distance matrices and considerably enhance error-correcting...

A single graph cannot comprehensively describe the true relationship among samples with high dimensionality, especially for graph-based methods which directly calculate distances in Euclidean space. To further improve performance of method based on presentation graph, this paper we propose a new composite called as low rank representation and collaborative (LRRCR) graph. The proposed LRRCR can obtain more informative knowledge robust image classification. experimental results three real face...