Smart Computing and Communication

With the arrival of big data era and the rapid development of artificial intelligence, deep learning has made breakthroughs in many fields. However, although it has been widely used in many fields, there are still many challenges in itself, such as the slow reference speed of neural networks. In stream processing scenarios that integrate deep learning algorithms, data is often massive and generated with high-speed, requiring the system to respond in seconds or even milliseconds. If the response speed is too slow, the actual application requirements may not be met, and the user experience cannot be guaranteed. How to use stream processing technology to improve the speed and throughput of such systems has become an urgent problem to be solved. This paper used the popular real-time stream processing framework Flink to implement a complete data stream processing program, and integrated three algorithms of face detection, facial key points detection and face mosaic into the processing logic. Setting the operator parallelism realized parallel processing of video data, which improved the system throughput. The user can choose which algorithm to perform on the video, and can also choose the parallelism according to the performance of the machine. The system implemented the Flink framework to process video in parallel, achieved the effect of improving processing efficiency, and completely implemented the front-end interface and back-end program.

Information systems record the current states and the access records in logs, so logs become the data basis for detecting anomalies of system security. To realize log anomaly detection, frameworks based on text, sequence, and graph are applied. However, the existing frameworks could not extract the complex associations in logs, which leads to low accuracy. To meet the requirements of the hyperautomation framework for log analysis, this paper proposes GenGLAD, a generated graph based log anomaly detection framework. The generated graph is used to express the log associations, and the node embedding of the generated graph is obtained based on random walk and word2vec. Finally, we use clustering to realize unsupervised anomaly detection. Experiments verify the detection effect of GenGLAD. Compared with the existing detection frameworks, GenGLAD achieves the highest accuracy and improves the comprehensive detection effect.

This paper proposes a blockchain-based digital-currency model based on certificates, and uses stablecoins as an example to illustrate the model. A stablecoin is anchored on a fiat currency, and often backed by 100% fiat such as USD. Unlike traditional cryptocurrency models, the blockchain system using this model does not store the assets directly, but instead store the certificates of these assets. People can trade digital certificates like trading digital currencies, but the value of these certificates in stored in custodian banks. The transaction process is slightly different from cryptocurrency transactions. The advantage of trading certificates is that the loss of these digital certificate does not mean the loss of value. If a private key is lost, a client can recover her assets from the custodian bank with a replacement certificate. Due to this nature, using this model, the center of digital economy remains at banks, rather than the blockchain network, as suggested by Digital-Currency Areas (DCA).

Analog meter is still widely used due to their mechanical stability and electromagnetic impedance. Relying on humans to read mechanical meters in some industrial scenarios is time-consuming or dangerous, it is difficult for current meter reading robots to operate quickly and maintain high accuracy in edge computing devices. Computer vision-based meter reading systems can solve such dilemmas. We designed an SSD network-based meter image acquisition system that can run in real time in an NVIDIA Jetson NX development board. Moreover, the model can quickly classify meter types and locate meter coordinates in the presence of light changes, complex backgrounds, and camera angle deflection. Tested on NVIDIA Jetson NX using TensorRT acceleration, the inference speed and accuracy reached 9.238 FPS and 53.95 mAP, respectively.

In the field of information interaction, when a project involves a large amount of cross-domain heterogeneous information, it is difficult to transmit and update the required information in a timely, accurate, reliable and safe manner in such a complex environment, thereby maintaining synchronization. The decentralization and immutability of blockchain technology provide new ideas for solving this problem. Therefore, this paper takes this as a starting point and proposes a cross-domain heterogeneous information exchange system based on blockchain, which calls smart contracts to realize data transmission and other functions. This paper states the design idea of the system, introduces the system software framework and network topology in detail, shows the flow chart of the system operation, and conducts simple software testing and verification. Finally, this work is summarized, and the development direction of future work is prospected, hoping to provide inspiration for the solution of such problems.

In a traditional blockchain system, the consensus protocol and the transaction management are coupled, i.e., transactions are performed during the consensus process. When the consensus protocol completes its process, the transaction is considered as completed. Such an integrated approach challenges the system scalability and regulatory issues. And is not in line with the regulatory principles such as PFMI (Principles of Financial Market Infrastructures). This paper proposed a new decoupling mechanism that separates consensus protocols from transaction management, and further separate transactions from settlement with regulation compliance. Based on the new scheme, this paper proposed a new block-building protocol for blockchain systems. This protocol allows consensus steps, transaction management, settlement, and regulation compliance can be done concurrently.

The existing blockchain consensus protocol has reached the level of availability in replicas in small-scale scenarios. However, if the blockchain system is composed of hundreds or even thousands of replicas, the throughput and delay will significantly decrease as the number of replicas increases, which makes it difficult to apply it in large-scale scenarios. This paper proposes an Adaptive Byzantine Fault-Tolerant (AdBFT) consensus protocol, which introduces the optimistic response assumption and proposes an adaptive approach to reach consensus with a latency of 2Δ in steady state, while providing the advantage of tolerating up to half of Byzantine failures, which can ensure security in a weaker synchronization model. Under the optimistic response condition, O(Δ) latency is achieved when the leader is honest and more than three-quarters of the replicas respond, and up to 1/3 of the Byzantine failures can be tolerated under synchronization.

This paper proposes a blockchain-based multi-level corporate digital wallet that meets financial regulatory requirements, including compliance with the Travel Rules, and identification of suspicious financial transactions. Unlike personal digital wallets, a corporate wallet is actually a large financial management system with internal blockchains, databases, and governing rules. Every transaction that the user operates on the wallet client will first be sent to the wallet server system. The wallet server will first check the compliance of the transaction, and then verify that the transaction meets the Travel Rules. At the same time, the digital wallet also checks if the transaction may be a financial fraud. In the Web3 era, many transactions will be performed using digital wallets, rather than bank accounts, a corporate wallet will play a significant role in the new digital world.

Traditional blockchain systems the consensus protocol and transaction management is performed together. In other words, the completion of the consensus protocol means the completion of a transaction. Furthermore, the transaction is considered as settled. This scheme is not compatible with modern financial transaction rules where transaction is distinct from settlement, and thorough regulation compliance should be performed to ensure that no money laundering. This paper proposes using five Merkle trees to maintain data in a blockchain system to separate consensus protocol from transaction management.

Unlike highly purposeful search, a recommender system tends to uncover the user’s potential interests and is a personalized information filtering system. Recently, the performance of hypergraph neural networks in classification tasks has attracted much attention. Compared with traditional recommender systems, hypergraph neural network-based recommender systems have better mining higher-order associations, accurate modeling of multivariate relationships, handling of multimodal and heterogeneous data, and clustering advantages. This fact drives the development of recommendation algorithms based on hypergraph neural networks. To this end, we 1) define generic links of recommender systems, and systematically analyze the challenges of hypergraph neural network-based recommender systems in different research directions. 2) present some new perspectives on existing weaknesses and future developments.

Equipment health assessment is a fundamental task in predictive equipment maintenance practice, which aims to predict the health of equipment based on information about the equipment and its operation, thus avoiding unexpected equipment failures. In the current context, equipment health assessment based on sequential deep learning methods is becoming more and more popular, however, such methods ignore the inter-device correlations, leading to their lack of readiness for health assessment of a large number of devices. To address this problem, this paper proposes a node-embedding-based device health assessment method, which creatively introduces a graph model for device health assessment and effectively improves the performance of health assessment. Firstly, this paper proposes a way to define equipment association graphs. Secondly, we introduce the node embedding technique to extract graph information. Finally, an equipment health assessment method based on the equipment association graph is proposed. Experiments show that the proposed method outperforms the existing prevailing methods.

With the rapid development of computer technology, ERP system arises at the right time as an advanced cost management tool. ERP is mainly implemented in enterprises with the purpose of optimizing resource allocation and giving full play to its functions. ERP system is an enterprise resource planning system with strategic management thought, which tries to reduce the consumption of resources. Cost management is an important part of enterprise management. ERP system provides a tool for the cost management of enterprises. It greatly enhances the comprehensive management ability of enterprises. This paper discusses the improvement of ERP system to enterprise cost. We proposes to combine enterprise cost management theory with big data, use digital technology to adjust the strategic cost plan of enterprises, and improve the cost accounting system of ERP + MES + SCADA industry finance deep integration. We build lean production management and control platform of cost information sharing and use digital technology to deeply integrate performance evaluation index.

With the emergence of more and more “AI + Field + HPC” applications, it is urgent to solve the problem of scheduling and management of High-Performance Computing (HPC) resources, as well as the fast and efficient “cloud service” of HPC applications. This engineering problem is particularly critical because it affects the progress of scientific research, the development period of the research platform, and the learning cost of scientists. To solve the problem, a set of reusable life cycle processes for HPC resources are designed. Based on the life cycle, we propose an open service interface based on HPC, which reduces the startup time under multiple refreshes and abnormal retries by using the mode of contention lock. The active interruption of users is a typical scenario in the startup phase. Furthermore, a read-write strategy with an overlay based on Singularity is implemented to save storage space and improve running speed. In order to evaluate the serviceability and performance of the proposed interface, we deploy the service on the Venus platform and make a startup comparison experiment. In addition, the reduction of storage for 100 users is also tested. The experimental results show that under the HPC environment with SLURM, the proposed open-service interface can effectively shorten 46% startup time of applications and services and reduce 25% storage at least for each user of the Venus platform.

Yao's Millionaires’ problem has led to the emergence of secure multi-party computation. As an important tool for privacy protection in cryptography, secure multi-party computation has attracted more and more scholars to study it. The socialist millionaires’ problem is the basic module of the secure multiparty computing protocol. Designing secure and efficient solutions for the socialist millionaires’ problem can be effectively applied to the secret ballot, electronic auction, and so on. Based on the vector encoding method, the Paillier encryption scheme, and the Goldwasser-Micali encryption scheme, two efficient socialist millionaires’ protocols are proposed and the protocols are analyzed. The correctness analysis, security proof, performance analysis, and experimental simulation show that the efficiency of the two protocols is superior to the existing schemes.

Occupancy information is useful for efficient energy management in the building sector. The massive high-resolution electrical power consumption data collected by smart meters in the advanced metering infrastructure (AMI) network make it possible to infer buildings’ occupancy status in a non-intrusive way. In this paper, we propose a deep leaning model called ABODE-Net which employs a novel Parallel Attention (PA) block for building occupancy detection using smart meter data. The PA block combines the temporal, variable, and channel attention modules in a parallel way to signify important features for occupancy detection. We adopt two smart meter datasets widely used for building occupancy detection in our performance evaluation. A set of state-of-the-art shallow machine learning and deep learning models are included for performance comparison. The results show that ABODE-Net significantly outperforms other models in all experimental cases, which proves its validity as a solution for non-intrusive building occupancy detection.

DNA is one of the most important information in every living thing. The DNA matching experiment is helpful for the study of paternity testing, species identification, gene mutation, suspect determination, and so on. How to study the DNA matching in the case of privacy protection has become the inevitable problems in the research of information security. The Hamming distance can reflect the similarity degree of two DNA sequences. The smaller the Hamming distance is, the more similar the two DNA sequences are. In this paper, the DNA sequence with $$l$$ l length is encoded with a 0–1 string with $$3l$$ 3 l length, and the protocol of confidentially computing Hamming distance is designed, which calculated the matching degree of two DNA under the premise of protecting DNA privacy. In addition, in view of the criminal suspect DNA matching problem, we design a secure computation protocol against malicious adversaries using the zero-knowledge proof and the cut-choose method to prevent or find malicious behaviors, which can resist malicious attacks.

At present, the research on human action recognition has achieved remarkable results and is widely used in various industries. Among them, human action recognition based on deep learning has developed rapidly. With sufficient labeled data, supervised learning methods can achieve satisfactory recognition performance. However, the diversification of motion types and the complexity of the video background make the annotation of human motion videos a lot of labor costs. This severely restricts the application of supervised human action recognition methods in practical scenarios. Since the zero-shot learning method can realize the recognition of unseen action categories without relying on a large amount of labeled data. In recent years, action recognition methods based on zero-shot learning have received great attention from researchers. In this paper, we propose an attention-based zero-shot action recognition model ADZSAR. We design a novel attention-based mechanism feature extraction method that introduces the current state-of-the-art semantic embedding model (Word2Vec). Experiments show that this method performs the best among similar zero-shot action recognition methods based on spatio-temporal features.

IoT data collected in a physical space have intrinsic values, such as the spatial and temporal correlation of people’s activities. As such spatio-temporal data inevitably includes private information like trajectory and waypoints, privacy exposure becomes a serious problem. Local Differential Privacy (LDP) has been gaining attention as a privacy protection procedure on a device collecting spatio-temporal data. However, LDP cannot retain spatial and temporal properties which are essential for cyber-physical systems. The is because LDP makes each data indistinguishable and inevitably removes spatial and temporal properties as well. In this paper, we propose a method enabling LDP to keep spatial and temporal properties on privacy protection process. Our method dynamically changes the strength of privacy protection (called privacy budget) for each of device groups who has resemble spatial and temporal behavior. This makes data of each device in a group indistinguishable within the group but a set of data made by a group distinguishable between groups in terms of spatial and temporal domains. As the whole data merged in a data store will consists of modified data with wide variety of privacy budgets, we arrange every privacy budgets so that merged data keeps particular strength of privacy protection. We call this process as Dynamic Private Spatial Decomposition (DPSD). The experimental results show that our LDP preserves the data utility while maintaining the privacy protection of the entire client because of DPSD.

With the continuous improvement of various high-performance computing systems, various data centers had also been fully expanded. Energy consumption and actual performance measurement were very important indicators, which were also key issues in how to judge parallel calls for some tasks in high-performance computer systems. Modern processors were basically equipped with software control functions such as DVFS (Dynamic Voltage Frequency Scaling), in the actual system operation to ensure that the system could ensure the reasonable operation of the system while reducing energy consumption indicators. This paper considered how the designed scheduling algorithm first divides tasks reasonably to ensure that the maximum completion time and energy consumption of the processor were sufficiently reduced when the directed acyclic graph was executed. Then considered making reasonable adjustments to the processor frequency using DVFS technology to adapt to the task while ensuring the critical path of the task. At the end of the article, make sure that the experimental verification algorithm could ensure that the task was completed and could reduce the energy consumption during task execution as much as possible.

With the rapid development of the software industry, the number of requirements included in the software is also increasing . In particular, in a complex software system such as an avionics software system, dozens of subsystems are usually included, and each subsystem includes hundreds of requirements, so the entire software system will include thousands of requirements. Furthermore, during software development, requirements will undergo frequent changes. When the number of requirements is huge and there are many people from different backgrounds participating in requirements development, it is easy to duplicate requirements. Therefore, in order to prevent redundant development, in the requirements analysis stage, requirements analysts usually need to detect all duplicate requirements in the requirements document and eliminate them, thereby improving the quality of software requirements and saving software development costs. However, when the number of requests is large, manually detecting duplicate requests can be a time-consuming and error-prone task. Aiming at this problem, this paper studies the method of duplicate requirements detection in English requirements.

Mediation analysis was concerned with the decomposition of the total effect of exposure on the outcome into the indirect effects and the remaining indirect effects, through a given mediation. However, when longitudinal data including time varying exposure and mediator variables, the estimated causal effects are affected by time varying confounders. Standard generalized linear equations did not give unbiased estimates. In this paper, we introduced inverse probability weighting technique to adjust such time varying confounders. Considering that the amount of data may be small and the distribution is not uniform, we decide to use Bayesian Inference to estimate the Structural Equation Model (SEM) parameters, and finally estimates the causal effect through counterfactual thought. This paper summarized the relevant theoretical knowledge of this method, verified the feasibility of this method by using the simulated data, and compared the performance of different methods.

Based on the influence of dark obstacles caused by insufficient light in an underground mine on the driving safety of an electric locomotive. This paper proposes an improved YOLOX target detection algorithm to effectively identify and classify the track obstacles of the unmanned electric mine locomotive. On the basis of the YOLOX target detection network, the CBAM attention module is added to the CSPDarket and the FPN part of the feature pyramid, and the loss function of YOLO head part is replaced by SIOU. The collected image data of track obstacles of electric locomotive under different lighting conditions are used as the training set. The Pytorch deep learning framework is used to construct an object detection model for training and verification. Experiments show that the average accuracy and recall rate of the improved YOLOX underground electric locomotive track obstacle detection model can reach 93.05% and 88.29%, and the speed is improved to 45.3 fps. Compared with other target detection models, this model can better realize the accuracy and real-time detection of underground electric locomotive track obstacles. It provides the basis for the intelligence of underground mine transportation equipment.

Cryptocurrency applications with blockchain technology as the underlying architecture have gradually developed into a new means of payment, and are expanding to all walks of life with the support of cryptography and consensus algorithms. Due to the disadvantages of low throughput and high latency, the blockchain has seriously hindered the widespread use of upper-layer applications and cannot meet the growing demand for users and transaction volumes. Drawing on the sharding idea of traditional databases, blockchain sharding, as a representative of on-chain scaling solutions, greatly improves the throughput of the blockchain system. At present, most of the network sharding schemes in the sharded blockchain adopt a strategy based on random sharding. This strategy does not take into account the performance of the node itself, resulting in large performance differences between different shards, further reducing the throughput of the entire system. In addition, the aggregation behavior of malicious nodes may also occur, reducing the security of the system. Aiming at the performance of each node, this paper proposes a sharding strategy based on the approximate ideal solution model (TOPSIS). Through the TOPSIS model, the nodes are scored according to the hardware performance of the node, the response time to the transaction and the results, etc., and the nodes are allocated to the corresponding shards according to the scoring results. The sharding strategy based on this model balances the performance differences among shards and improves the throughput of the entire system.

Spectroscopy devices suffer from the pulse pileup phenomenon, caused by overlapping of the signals. The energy-domain based pileup correction algorithm estimates the pulse energy distribution by measuring the duration and energies of pileups directly and does not need to identify each individual pulses. The correction algorithm can efficiently recovers the energy spectrum even under a very high photon arrival rate. However, the correction algorithm is sequential in nature and is slow when the energy resolution is high. A fast parallel implementation of the original correction algorithm is proposed in this paper. The parallel counterpart leverages state-of-the-art many-core system technology and achieves a nearly linear acceleration when the problem size scales. The speedup ratio exceeds 1,000 when the energy spectrum is split into 2,048 bins.

Machine learning has shown great impact in a lot of applications. Within all types of tools, deep learning should be one of the most important techniques thanks to its ability to capture the correlation between the input features and output results. However, the relatively long training time and high computation complexity remain a big problem in deep learning. In addition, the impossibility to explain the model makes it harder for us to look for alternatives to fix the bad fitting results. Therefore, this paper aims at improving the deep learning model training result by proposing a principal component extraction algorithm. Compared with the previous Principal Component Analysis (PCA) methods, this algorithm creatively consider not only the original input components but also the computed variables in the first hidden layer in neural network so as to capture more representative components. The experiment shows that compared to previous PCA method, this can better capture the principal components from all input variables.

Time series prediction has become an important research direction in data mining because of the time-varying pattern of data in various fields. However, time series prediction suffers from the problem of vulnerability to adversarial example attack, which leads to models making wrong decisions in critical application scenarios and causing great losses to people’s lives and properties. In addition, there is relatively little attacks research on time series prediction, and the existing attack methods simply migrate classical-attack methods in the image to time series prediction. On the one hand, it not only without fully considering the characteristics of temporal data but also without comparing and analyzing the effects of those classical-attack methods on time series prediction models. On the other hand, there is no comparative analysis of the effectiveness of these classical attack methods in different time series prediction methods. To address the above problems, this paper firstly compares the effectiveness of the attack methods on some time series prediction models and analyzes the inner mechanism of these time series prediction models. In addition, this paper finds that the defense ability of those models is related to their ability to portray the overall trend of time series data. Therefore, this paper further propose the new attack method, LowFreqAttack. The experimental results show that LowFreqAttack can attack the three existing time series prediction models better than the existing attach methods.

Machine learning-based models are one of the main methods for detecting reentrant vulnerabilities. However, these models extract smart contract features only from a single form, resulting in incompleteness and inaccuracy of features. To address this problem, we propose a novel machine learning-based model for reentrant vulnerabilities detection. We extract and fuse features from abstract syntax trees, opcodes, control flow graph basic blocks, and combine machine learning algorithms for reentrant vulnerabilities detection. Additionally, to address the time-consuming problem of manual labeling, we also propose an approach for automatically adding dataset labels. We perform experiments on Smartbugs and SolidiFi-benchmark datasets and results show that our model outperforms existing models.

Object detection techniques are a major part of computer vision research, with large-scale applications in industrial, scientific and other scenarios. Technologies such as face detection, medical image detection, autonomous driving, and traffic detection have played a significant role in people’s lives. With the rapid development of deep learning, many application areas, such as image classification, text classification, machine translation, etc., have achieved breakthrough success in combination with deep learning. R-CNN brings object detection into the era of deep learning, and its advantage compared with traditional methods is that the former requires personnel to extract features manually, while the latter uses deep learning to extract features automatically, which greatly improves efficiency, simplifies operation, and opens a new era of object detection research. First, this paper provides an overview of deep learning-based object detection backbone networks, reviews and analyzes milestone object detection algorithms, compares commonly used datasets, summarizes applications, and finally concludes the paper.

For companies using blockchain technology, it is critical to select the most suitable blockchain platform to develop enterprise applications. However, it is still a challenge for enterprises. As an important part of modern decision science, multi-criteria decision-making can well solve the problem of blockchain platform selection. Blockchain platforms integrate various blockchain technologies, and enterprises also need to consider multiple different criteria in the decision-making process. Therefore, this paper will use a heterogeneous multi-criteria decision-making method to solve the blockchain platform selection problem. First, the blockchain platform alternatives and evaluation criteria used for decision-making are identified. Second, blockchain platform alternatives are evaluated with appropriate fuzzy numbers based on defined evaluation criteria. Then, the original evaluations are consistency and normalized to obtain a normalized evaluation. Next, the improved information content formulas of the axiomatic design is proposed to obtain the information content of each normalized evaluation. Then, the weights of all evaluated criteria are obtained using the entropy weight method. Finally, the total weighted information content of each blockchain platform alternative is obtained. With validation, the decision-making model of blockchain platform proposed in this paper has a strong reference value.

As an emerging technology, Software-Defined Industrial Networks (SDIN) appears to be a vital technical approach for powering up new manufacturing modes due to its higher-level scalability and controllability. However, a few threats are still restricting the implementation of SDIN and Distributed Denial of Service (DDoS) is one of the common attacks. In this paper, we focus on the DDoS issue in SDIN, and propose a blockchain-empower SDIN scheme and an Auxiliary Classifier Generative Adversarial Networks (AC-GAN)-based DDoS attack detection model. Our experiment evaluations have demonstrated the effectiveness and performance of our proposed approach.

The federated learning technology provides a new method for data integration, which realizes sharing of a global model and prevent the leakage of user’s original data information. In order to resist data poisoning attack from some participants, ensure reliability and accuracy of the global model, and ensure fairness of the aggregation process in federated learning, we propose a blockchain-based fairness enhanced federated learning scheme. The accuracy of global model and fairness of the aggregation process is guaranteed by an adaptive aggregation algorithm which can defense data poisoning attack. The reliability of federated learning process is ensured by recording the entire process of the model training on the blockchain and using digital signatures. The privacy of each participant of federated learning is protected by public key encryption combined with the use of random numbers. Theoretical analysis and experiments show that the scheme can protect privacy of each participant, mitigate data poisoning attack and ensure the reliability and fairness of the entire federated learning process.

With the development of digital currencies and 5G technology, blockchain has gained widespread attention and is being used in areas such as healthcare, industry and smart vehicles. Many security issues have also been exposed in the course of blockchain applications. Cryptography can ensure the security of data on the blockchain, the integrity and validity of data as well as the ability to authenticate users and anonymize them. This article therefore examines the cryptography underlying blockchain security issues, providing an overview of cryptographic homomorphic encryption, zero-knowledge proofs and secure multi-party computation commonly used in blockchains. At the same time, the development of quantum computing is bound to affect existing cryptographic systems, and blockchains applying these cryptographic systems are bound to be hit hard, so this article discusses four of the most promising post-quantum cryptography techniques available: hash-based public key cryptography, code-based public key cryptography, multivariate public key cryptography, and lattice-based public key cryptography.

We propose a novel smart contract re-entry vulnerability detection model based on BiGAS. The model combines a BiGRU neural network that introduces an attention mechanism with an SVM. We start from the data features of smart contracts, learn the model layer by layer to achieve feature extraction and vulnerability identification, introduce batch normalization, Dropout processing and use improved model classifiers to improve the vulnerability identification accuracy, model convergence speed and generalization capability of smart contracts. We had conducted numerous experiments, and the experimental results showed that BiGAS Detection Model has a strong vulnerability detection ability. The accuracy of vulnerability detection reached 93.24%, and the F1-score was 93.17%. We compared our approach with advanced automated audit tools and other deep learning-based vulnerability detection methods. The conclusion was that our method is significantly better than the existing advanced methods in detecting smart contract reentrancy vulnerabilities.

Blockchain, designed with cryptographic technology, is widely used in the financial area, such as digital billing and cross-border payments. Digital signature is the core technology in it. However, digital signatures in public key cryptosystems face the threat of simple power analysis in Side-Channel Analysis (SCA). The state-of-the-art simple power analysis based on clustering mostly will appear outliers in the process of analysis, which will reduce success rate of key recover. In this paper, we propose a new SCA method with clustering algorithm Density-Based Spatial Clustering of Applications with Noise (DBSCAN) and deep learning technology Convolutional Neural Network (CNN), called DBSCAN-CNN, to analyze public key cryptosystems. We cluster data with DBSCAN firstly. Then we train a CNN model based on the trusted clustering results. Finally, we classify the outliers of clustering results by the trained model. We mount the proposed method to analyze an FPGA-based elliptic curve scalar multiplication power trace which is desynchronized by simulating random delay. The experimental results show that the error rate of the proposed method is at least $$69.23\%$$ 69.23 % lower than that of the classical clustering method in SCA.

With the development of the big data era, the amount of data has entered an explosive growth phase. Limited by the constraints of cost, efficiency, and security of self-built storage systems, enterprises are forced to outsource files to cloud storage systems. However, the lack of file security and auditability in cloud storage systems continues to threaten the security of outsourced files. This paper designs and implements BFSOut, which is a secure file outsourcing method based on consortium blockchain. It uses Hyperledger Fabric and Interplanetary File System (IPFS) as the underlying storage engine, which solves the problem of cloud storage security issues. In BFSOut, in order to ensure the security of outsource files, the client-side offline block encryption is used. Furthermore, a dynamic hybrid encryption scheme is adopted, making the overall encryption effect more Efficient. Experimental performance analysis show that the system has good performance.

Fabric is currently the most popular consortium chain platform with a modular architecture that provides high security, elasticity, flexibility and scalability. Smart contracts realize the automatic execution of transactions and the operation of reconciliation data. The Fabric platform supports general programming languages ​​to write smart contracts. However, in the development process of smart contracts, due to insufficient understanding of the underlying operating logic of smart contracts, developers are prone to introduce some risky operations, resulting in a mismatch between the execution logic of smart contracts and business logic, resulting in a lot of losses. The read-after-write risk is a relatively complex and common security risk in smart contracts. Currently, many detection tools cannot detect this risk. There is an urgent need for a solution that can quickly and accurately detect the read-after-write risk in smart contracts. This paper proposes a static analysis smart contract read-after-write risk detection method based on key methods and call chains. The scheme extracts key method patterns on the abstract syntax tree, identifies and locates key methods with risks, greatly reduces the interference of useless nodes on detection, and realizes rapid detection. By constructing the key method call chain, the real call scene is restored according to the call type and attribute of the key method. After experimental verification, compared with the current popular smart contract risk detection tool Revive^CC, the tool proposed in this paper has higher detection accuracy and can more accurately locate the read-after-write risk in smart contracts.

Smart contracts are one of the most successfully applied technologies on the blockchain, which are decentralized and immutable. Smart contracts cannot be modified once deployed. Therefore, security detection of smart contracts before deployment is essential. Some smart contracts may have tx.origin dependency vulnerabilities. In this paper, we propose a critical path vulnerability detection method for detecting tx.origin dependency vulnerabilities in smart contracts. Then, in order to solve the problem that the traditional search algorithm cannot determine the critical path, we propose a path determination method based on path priority. Our method determines the critical path in the control flow graph, which enables us to detect the vulnerabilities existing in smart contracts more quickly. The experimental results show that our method is more efficient than the existing technology and the false positive rate is lower.

Due to the unique global state and transaction sequence characteristics of smart contracts, the detection method based on a single test case cannot improve the vulnerability detection rate during contract detection. The current contract testing methods based on genetic algorithms have not yet solved the problems caused by these characteristics. Therefore, we propose an adaptive fuzzing method based on dynamic taint analysis and genetic algorithm, SDTGfuzzer. SDTGfuzzer focuses on dynamic taint analysis to collect runtime information as feedback, and focuses on solving the challenges brought by global variables and transaction sequences for contract testing. Genetic Algorithms work well in test case generation for fuzzing. Therefore, SDTGfuzzer optimizes the genetic algorithm based on an efficient and lightweight multi-objective adaptive strategy, focusing on solving the problem that the contract constraints cannot be covered due to the global state. Experimental results show that our method has a higher vulnerability detection rate than other tools for detecting contract vulnerabilities.

It is necessary to accelerate digital development to make digital economy, society and government shine brightly in the future. Digital technology and the real economy will be deeply integrated, and a large number of new industries and new models will emerge. Cloud computing is an important industry to create further advantages in the digital economy. As the key to the development of enterprise business, cloud computing is the decisive factor. The ability to support billing for the entire cloud is a vital link, and it is the core capability of the whole system. It requires high accuracy and performance. Based on cloud billing, this paper analyzes the challenges cloud computing service support faces. It studies the architecture upgrade, builds the cloud service billing support capability based on stream native technology, meets the flexible billing needs of cloud services, and realizes the rapid and efficient operation support of cloud services. This system has been put into production and played a significant role strongly supporting the high-quality development needs of cloud business billing.

Internet of Things (IoT) data from different trust domains is usually shared to assist in providing more services, where privacy sensitive information of shared data will be leaked or accessed without authorization. The traditional centralized access control method is difficult to adapt to the current dynamic and distributed large-scale IoT environment, and there is a risk of the single point of failure. To address these challenges, we propose a fine-grained access control framework for shared data based on cross-blockchain technology and Interplanetary File System (IPFS). In this framework, we firstly introduce a cross-blockchain module to realize cross-domain data sharing and solve the problem of data isolation between different data domains in IoT. Then IPFS is used to store the shared data, avoiding the risk of centralized storage. Combining symmetric encryption algorithm with ciphertext policy attribute based encryption (CP-ABE) algorithm, the fine-grained access control of shared data is guaranteed. In addition, the blockchain is applied to store the decryption key and the storage address of the original data, which records the authorization operation of access transactions and audits the access behavior of users. Experimental results show that the proposed scheme can provide higher performance compared to centralized access control methods.

In today’s world, with the deepening of population aging, chronic diseases have become the main diseases which affecting human health. Diabetes is a common chronic disease. Its incidence rate is high and rising year by year. For patients with diabetes, it is very important to predict the development of the disease and the possible complications for their follow-up treatment and recovery. However, the existing prediction of diabetes is mostly limited to the prediction of the incidence rate of patients, and only uses collaborative filtering or features for prediction, and rarely uses the patient’s condition information to construct sequences and predict the development of the disease. Our aim is to make a accurate prediction on the development of diabetes patients and the possible complications. We need to use the historical development information of patients. Therefore, we propose a sequence based model fusion prediction algorithm, which effectively fuses the sequence and feature information. We use the high-order Markov Chains with attention mechanism as the basic learner for learning sequence information, and we also use XGBoost and CatBoost as the basic learner for learning feature information. Finally, LightGBM is used as a meta learner to fuse the output of the base learner. Experiments on the data of diabetes patients show that our method achieves better results than the original sub learner.

With the rapid development of blockchain technology, smart contracts (SCs) applied in digital currency transactions have been widely used. However, SCs often have vulnerability in their code that allow criminals to exploit them to steal associated digital assets. Benefiting from the development of machine learning technology and the improvement of hardware performance, one can use deep learning techniques to analyze code and detect vulnerabilities. This paper proposes an innovative combination of opcode sequences and abstract syntax trees for source code parsing. And a method based on the combination of self-attention mechanism and bidirectional long-short term memory neural network is proposed to detect the vulnerability of SCs after word embedding. Experimentation results show that the two parsing methods can complement each other and effectively improve the accuracy of vulnerability detection.

Since the launch of Ethereum in 2013, the smart contract has been a momentous part of the blockchain systems due to its character of automatic execution. The generation of smart contracts has also attracted extensive attention from the academic community. However, the preparation and generation of smart contracts are still mainly manual so far, which limits the scalability of the smart contracts. In this paper, we put forward a new method to generate smart contracts automatically based on knowledge extraction and Unified Modeling Language (UML), which can significantly accelerate the generation of smart contracts. We will describe this method in more detail based on the logistics supply chain.

As the underlying implementation technology of the current main-stream digital currency, blockchain can establish a trusted distributed system without relying on third-party trusted institutions or a privacy-protect system. The decentralized characteristics of blockchain have broad application scenarios, such as the Internet of Things, financial technology and other fields. However, the scalability of the current blockchain is seriously insufficient, such as limited throughput in performance, small storage capacity, and difficulty in scaling functions. This paper introduces the definition and technical classification of scalability, analyzes the current problems of scalability and briefly introduces the current main-stream expandable technologies such as sharding, on-chain and off-chain storage, off-chain payment channel and cross-chain technology from two aspects of performance and function, as well as the principles and ideas of these technologies. Finally, the research progress of current blockchain extension technology is summarized, and the problems faced by the current extension scheme are pointed out, which provides a direction for future research work.

Smart cities bring new ideas to solve the social, economic, and environmental problems existing in traditional cities. However, the services of smart cities suffer from centralized data storage and untrustworthiness. As a decentralized distributed database, blockchain provides distributed and trusted infrastructure technology for the construction of smart cities. In this survey, we conduct a brief survey of the literature on blockchain applications in smart cities. We review the application of blockchain in smart cities from the four main application scenarios of energy, transportation, medical care, and manufacturing. Then, we discuss the realization of blockchain-based smart cities from the perspectives of privacy and storage. We believe this survey provides new ideas for the development of smart cities.

This paper introduces an early content propagation popularity prediction model based on graph neural network and variational inference topic dependent dynamic variational autoencoder model (CD-VAE). CD-VAE captures the dynamics in the content propagation process, aggregates the topological information in the information diffusion process using GraphSAGE, approaches the uncertainty in terms of time and node from the perspective of probability by introducing two variational autoencoders, considers the changes in semantic characteristics in the process by integrating natural language processing methods into the model, and therefore significantly improves its prediction performance.

Artificial neural network has achieved great success in many fields. Considering the unique advantages of naturally generated networks, we combine the geometric complex network model with the existing neural network model to build a neural network with geometric space structure characteristics. We proposes a GeoNet neural network model based on a random geometric network structure and finds that the neural network with a natural structure has good classification performance, and the classification accuracy is higher than the widely used neural network structure.

With the continuous development of blockchain technology, smart contract has become an important research object among the achievable technologies on blockchain technology. Based on the characteristics of decentralization, tamper-proof and transparency of blockchain, it provides a reliable technical support for the implementation of smart contract. Based on blockchain smart contract technology, this paper aims to design a smart contract management engine with higher versatility, security, and feasibility to develop a smart contract from the joint participation of multiple users. The smart contract is proliferated through the P2P network and deposited into the blockchain. And the blockchain is designed to automatically execute the smart contract, providing a new solution to the problems of opacity, easy tampering, and low efficiency of the traditional contract. It is safer and more reliable. By specifying the treaty and trigger conditions through the program code, once the conditions are met, the contract will be automatically executed, which greatly reduces the time and space costs. Using Ether and smart contracts to develop distributed applications to realize this technology, the feasibility of applying blockchain technology in this field is explored, and a new technical implementation is provided for traditional contract signing.

Due to the high degree of privacy and sensitivity, it is difficult to share distributed digital identity with multiple parties. Blockchain-based distributed digital identities could address the issues of data sharing. However, the blockchain systems contain a wide variety of heterogeneous autonomous systems, and each blockchain system performs independent identity authentication. To be specific, it is difficult to realize mutual recognition and trust in digital identity. In this survey, we introduce the technical architecture of distributed digital identity and research the technical principle and development application of cross-chain technology. In addition, we expound the research status of cross-chain technology and the cross-chain-based distributed digital identity mechanism to realize unified and trusted cross-chain identity mutual recognition.

At present, the security strategy of the lower boundary protection equipment of the power system can no longer meet the needs of the current business growth. A large number of redundant strategies cause the protection performance of the boundary firewall to decline. At the same time, the large number of business growth causes the network boundary order of the power grid system to be blurred. In order to prevent the paralysis and partial collapse of the network and ensure the reliability and integrity of the power business data and enterprise information, this paper develops a smart border firewall optimization tool. This tool can not only integrate the security device policies of different manufacturers through Simple Policy Specification Description Language (SPSDL), but also prioritize security rules according to the frequency of use through keyword filtering algorithms and rule optimization decision trees, then realize the classification, streamlining, optimization and upgrading of firewall security rules. The research results show that the power system firewall can achieve an accuracy rate of more than 90% when the strategy is imported. The rule optimization part can reduce the unique correlation addition index of this paper to about 0.2, which solves the problem of firewall security strategy import language diversification. It further eases the pressure of firewall policy redundancy under the power system.

Effective vulnerability detection of large-scale smart contracts is critical because smart contract attacks frequently bring about tremendous economic loss. However, code analysis requiring traversal paths and learning methods requiring many features training is too time-consuming to detect large-scale on-chain contracts. This paper focuses on improving detection efficiency by reducing the dimension of the features, combined with expert knowledge. We propose a feature extraction method Block-gram to form low-dimensional knowledgeable features from the bytecode. We first separate the metadata and convert the runtime code to opcode sequence, dividing the opcode sequence into segments according to some instructions (jump, etc.). Then, we mine extensible Block-gram features for learning-based model training, consisting of 4-dimensional block features and 8-dimensional attribute features. We evaluate these knowledge-based features using seven state-of-the-art learning algorithms to show that the average detection latency speeds up 25 to 650 times, compared with the features extracted by N-gram.

This paper explains how large-scale IT systems balance security and enabling under the pressure of supporting complex business models, constantly introducing new technologies, maintaining unified identity management, and meeting the business diversification of digital transformation. The paper introduces the necessity and basic idea of using China Mobile Super SIM technology to realize the enhanced Authentication, Authorization, Accounting, and Audit (4A) identity authentication center. We then describe the system’s design principles, references, and architecture and detail several key processes and capabilities. Finally, the implementation and evolution of the enhanced identity authentication system are described.

With the increasing maturity of Industrial Internet of Things (IIoT) technology, resource-constrained devices are widely applied to segments of the factory, which puts significant pressure on their computational capacity. In order to address this issue securely, we propose a verifiable and privacy-preserving outsourced computation system that employs SRAM PUF to safeguard the hardware security of devices and blockchain to achieve public verifiability and data privacy, thereby greatly guaranteeing the security of outsourced computation in the IIoT environment. Additionally, we protect the rights of calculators using a mechanism that identifies malicious calculators. Finally, compared with other existing schemes, the experimental results demonstrate that our scheme provides more efficient and secure outsourced computation services for IIoT devices.

Deep learning has become a popularly used tool in large amount of applications. Given its ability to explore the input and output relationship, deep learning can perform well in terms of prediction. However, one important drawback in this framework is that the model cannot be easily trained due to huge search space and large number of parameters. In response to such problem, this paper proposes a novel way to build a deep neural network. Specifically, it tries to consider an unbalanced structure of deep neural network by expanding the number of nodes in the beginning to extract as much information as possible and then shrinking quickly to converge to the final result. The experiment results show that our proposed structure can output equally good output with much faster time compared with traditional methods.

In recent years, blockchain research has set off an upsurge in academia, and it is called the next generation of value Internet. Because of its decentralization, anonymity, security, immutability, traceability and other characteristics, blockchain is gradually accepted and developed by people. With the deepening of research and the integration of technologies such as deep learning, blockchain has gradually been applied to various fields such as credit reporting, government, medical care, and industrial Internet of Things, not just the initial virtual currency field. This article mainly discusses the three important stages of blockchain public chain development, namely Bitcoin, Ethereum, and meta-verse, and introduces some basic supporting technologies of blockchain, as well as the research status and future trends of blockchain. Simple Analysis. By vertically introducing the development history of the blockchain, researchers can have a more concrete understanding of the status quo of the blockchain, and provide ideas for blockchain-related research.

Food traceability can be used to quickly pinpoint problematic links and minimise the risks associated with food safety incidents by viewing the trajectory of food circulation after a food safety incident has occurred. Blockchain, as an emerging technology with characteristics such as decentralisation, asymmetric encryption, tamper-proof and traceability, can be used to generate link traceability technology, which provides great convenience for us to effectively regulate food safety. In response to the need for multiple parties to share information in the traditional food safety traceability system there are many problems such as non-uniformity of the traceability chain, user cultivation and high traceability costs. This paper introduces Blockchain technology to improve the food safety traceability system, using the decentralised and fully distributed DNS service provided by Blockchain to achieve domain name query and resolution through peer-to-peer data transfer services between various nodes in the network, which can be used to ensure that the operating system and firmware of the infrastructure of the food production process have not been tampered with, introducing QR code technology, RFID, ZigBee.Web Web server and other key IoT technologies to achieve information identification and coding for identifying production objects, and identify the bottlenecks that limit the performance of the system by analysing and studying the consensus mechanism of the super ledger Fabric, and use smart contracts and consensus mechanisms as support to build food safety data assurance and food supervision methods to improve overall traceability efficiency.

Numerical understanding relies on some content understanding techniques, which can be based on rules, entity extraction, and machine reading comprehension. Traditional methods often require a large number of regular expressions or a large number of data annotations, and often do not have a deep understanding of numerical values, lacking the ability to distinguish similar numerical values. In this paper, we propose a few-shot learning framework for numerical understanding tasks in Chinese medical texts, and through dynamic negative sampling of the training data, the model’s ability to discriminate similar numerical values is enhanced. We use patient text data provided by 13 hospitals in Beijing to conduct experiments. The results show that our newly proposed method is superior to training the baseline pretrained language model directly, the EM increases by 38% and the F1 increases by 27.59%.

In the process of 5G power grid inspection robot moving for a long time, the sensor constantly collects the feature information of the substation. Due to the limited memory capacity, this feature information must be stored on the network, which requires the storage network must have strong security. The storage network based on blockchain can better solve the problem of feature data encryption, and a good feature extraction method can also relieve the pressure of network storage. As the input information of the whole SLAM, feature points play a crucial role in the detection performance and accuracy of the whole SLAM. When the extracted feature points are few or evenly distributed, they cannot express the information of the whole environment, which will make the mapping and localization error of SLAM system larger, and seriously lead to the loss of tracking. In this paper, we first analyze the standard ORB algorithm and the Qtree_ORB algorithm. Aiming at the problems existing in the two algorithms, an improved ORB feature extraction algorithm is developed. For the problem that the feature points extracted by the Qtree_ORB algorithm are too uniform, the maximum division depth of the quadtree is limited according to the number of feature points required for each layer of the image pyramid, which improves the problem that the feature points are too uniform. Finally, we evaluate the performance of the improved algorithm, and analyze the uniformity of feature points to verify the performance and robustness of the improved algorithm.

Blockchain is experiencing the transition from the first generation to the second generation, and smart contract is the symbol of the second generation blockchain. Under the background of the explosive growth of the second-generation blockchain platform and applications represented by smart contracts, frequent smart contract vulnerability events seriously threaten the ecological security of the blockchain, reflecting the importance and urgency of smart contract vulnerability detection. In this paper, we proposed a smart contract vulnerability detection method based on a Siamese network. We combined the Siamese network with Long Short-Term Memory (LSTM) Network neural network to complete the task of smart contract vulnerability detection. The Siamese network used in this paper consists of two subnetworks that share the same parameters onto a low dimension and easily separable feature space. Siamese network is now widely used in the field of image similarity and target tracking. In this paper, we improve the Siamese network so that it can be used for smart contract vulnerability detection. By comparing with previous research results, the model has better vulnerability detection performance and a lower false-positive rate.

This paper proposes a retweeter predction model based on attention model and Tranformer encoding–Forwader Prediction Model based on User Preference (FPM-UP). Considering the impact of release time, content, and current external context information in the forwarding process, FPM-UP integrates user attribute embedding and context-user dependency into a temporal and text attention model for the prediction of the next forwarding user. Compared with the existing methods, FPM-UP significantly improves the prediction accuracy.

In the field of information interaction, when a project involves a large amount of heterogeneous information, it is difficult to transmit and update the required information timely, accurately, reliably and securely in such a complex environment to maintain synchronization. At present, blockchain itself has problems such as high storage pressure of nodes, low access efficiency, and simple query. Therefore, this paper takes this as a starting point and proposes a data mapping method of physical resources based on node attributes and heterogeneous nodes, which provides a general method for the data mapping from actual physical resources to information domain. This method can define the attributes of heterogeneous physical information nodes and support unified expression of various physical information resources on the same platform in the real physical world, then connect heterogeneous physical information nodes and improve the sharing ability of data resources between nodes. When accessing data, the corresponding content can be found in the off-chain database by obtaining the index information of the off-chain location stored on the chain. This method takes advantage of the large space and high access efficiency of the off-chain storage system to share the pressure of on-chain data storage. This paper expounds the design idea of the system, introduces the design objective and method in detail, gives the flow chart of the system operation, and carries out a simple software test and verification. Finally, this paper summarizes the work and prospects the development direction of future work, hoping to provide inspiration for solving such problems.

Over the past several years, due to the progression toward data-driven scientific disciplines, the field of Big Data has gained significant importance. These developments pose certain challenges in the area of efficient, effective, and secure management and transmission of digital information. This paper presents and evaluates a novel Distributed Ledger Technology (DLT) system, Fibereum, in a variety of use-cases, including a DLT-based system for Big Data exchange, as well as the fungible and non-fungible exchange of artwork, goods, commodities, and digital currency. Fibereum’s innovations include the application of non-linear data structures and a new concept of Lazy Verification. We demonstrate the benefits of these novel features for DLT system applications’ cost performance and their added resilience towards cyber-attacks via the consideration of several use cases.

To solve the problem in secure encryption in cryptography, indistinguishability Obfuscation (iO) was born. It is a crypto-complete idea, based on which we can build many cryptographic construction. The implementation of it can hide both the dataset and the program itself. In this paper, we use the idea of translation in the (Natural Language Processing) NLP-like language model to realize the conversion between plaintexts and ciphertexts with the help of hints. We trained a self-attention transformer model, successfully hiding the dataset as well as the encryption and decryption programs. The input of the encryption model is a plaintext prefixed with a hint and the output is the result of encryption using one of the specified algorithms. The input and output of the decryption model are the opposite of the encryption one.

Nowadays, blockchain distributed ledger technology is becoming more and more prominent, and its decentralization, anonymization, and tampering obvious features have been widely recognized. These excellent technical features of blockchain have also made it a hot issue for global research. With the wide application of blockchain technology in various industries, some defects are gradually exposed, and more prominently, the blockchain system is unable to meet the current demand of explosive growth of data volume and frequent data interaction. As one of the key technologies to solve this problem, sharding technology is gaining attention. This article introduces common blockchain scaling schemes and focuses on an overview of blockchain sharding. Sharding technology is introduced from two perspectives of intra-slice consensus and inter-slice consensus. The current mainstream slicing technology is summarized according to three different slicing methods: network sharding, transaction sharding, and state sharding. Finally, the challenges faced by current blockchain sharding technology are analyzed and the full text is summarized.