Table of Content

25 September 2024, Volume 10 Issue 9

Previous Issue   

A Review of Adversarial Attack on Autonomous Driving Perception System

2024, 10(9):  786. 

Asbtract ( )   PDF (1560KB) ( )  

References | Related Articles | Metrics

The autonomous driving perception system collects surrounding environmental information through various sensors and processes this data to detect vehicles, pedestrians and obstacles, providing realtime foundational data for subsequent control and decisionmaking functions. Since sensors are directly connected to the external environment and often lack the ability to discern the credibility of inputs, the perception systems are  potential targets for various attacks. Among these, adversarial example attack is a mainstream attack method characterized by high concealment and harm. Attackers manipulate or forge input data of the perception system to deceive the perception algorithms, leading to incorrect output results by the system. Based on the research of existing relevant literature, this paper systematically summarizes the working methods of the autonomous driving perception system, analyzes the adversarial example attack schemes and defense strategies targeting the perception system. In particular, this paper subdivide the adversarial examples for the autonomous driving perception system into signalbased adversarial example attack scheme and objectbased adversarial example attack scheme. Additionally, the paper comprehensively discusses defense strategy of the adversarial example attack for the perception system, and subdivide it into anomaly detection, model defense, and physical defense. Finally, this paper prospects the future research directions of adversarial example attack targeting autonomous driving perception systems.

A Large Language Model Detection System for Domainspecific Jargon

Ji Xu, Zhang Jianyi, Zhao Zhangchi, Zhou Ziyin, Li Yilong, and Sun Zezheng

2024, 10(9):  795. 

Asbtract ( )   PDF (2610KB) ( )  

References | Related Articles | Metrics

Large language model (LLM) retrieve knowledge from their own structures and reasoning processes to generate responses to user queries, thus many researchers begin to evaluate the reasoning capabilities of large language models. However, while these models have demonstrated strong reasoning and comprehension skills in generic language tasks, there remains a need to evaluate their proficiency in addressing specific domainrelated problems, such as those found in telecommunications fraud. In response to this challenge, this paper presents the first evaluation system for assessing the reasoning abilities of DomainSpecific Jargon and proposes the first domain specific jargon dataset. To address issues related to cross matching problem and complex data calculation problem, we propose the collaborative harmony algorithm and the data aware algorithm based on indicator functions. These algorithms provide a multidimensional assessment of the performance of large language models. Our experimental results demonstrate that our system is adaptable in evaluating the accuracy of questionanswering by large language models within specialized domains. Moreover, our findings reveal, for the first time, variations in recognition accuracy based on question style and contextual cues utilized by the models. In conclusion, our system serves as an objective auditing tool to enhance the reliability and security of large language models, particularly when applied to specialized domains.

A Poisoningresistant Verifiable Secure Federated Learning Scheme #br# in IoT Perception Environments#br#

2024, 10(9):  804. 

Asbtract ( )   PDF (1748KB) ( )  

References | Related Articles | Metrics

To address the issue of model poisoning during predictive model training in the IoT intelligent sensing phase, this study proposes an antipoisoning attack scheme with verification capabilities. The scheme employs a cosine similarity clustering mechanism and a filtering strategy as a trusted thirdparty detection algorithm, integrating homomorphic encryption for authentication. Additionally, lightweight data encryption is used to protect the privacy of local model data. The Shamir Secret Sharing algorithm ensures robustness in model training against users dropout. By introducing a trusted third party, the scheme effectively detects and prevents dishonest users or attackers from compromising the accuracy of federated learning models. Simulation results demonstrate that the scheme can accurately detect model data involved in training while ensuring the security of users’ local model data and handling large volumes of heterogeneous data in IoT intelligent sensing environments.

An Efficient Encrypted Database System Solution Based on Fully  Homomorphic Encryption

2024, 10(9):  811. 

Asbtract ( )   PDF (930KB) ( )  

References | Related Articles | Metrics

In recent years, with the growing focus on privacy protection, fully homomorphic encryption (FHE)based encrypted database management systems (DBMS) have gained significant research attention. FHE allows DBMS to be outsourced to cloud servers without revealing plaintext data, preventing internal leaks and external breaches. However, FHEbased DBMS faces challenges such as high computational latency and low query processing capacity. To address these challenges, an efficient ciphertext database system based on Confusion Modulus Component Fully Homomorphic Encryption (CMPFHE) is proposed. This system designs a ciphertext index method that employs  symbolic functions and modulus operations, which reduces computation overhead and improving query efficiency. Additionally, it employs Nvariable Nequation homogeneous equations to achieve rapid ciphertext index retrieval, significantly decreasing the number of operations. This solution performs keyword queries on 10K rows of ciphertext data in just 54 seconds, demonstrating the practical feasibility of fully homomorphic encrypted databases.

Improved Byzantine Faulttolerant Consensus Algorithm Based on  Node Recognition

2024, 10(9):  818. 

Asbtract ( )   PDF (1380KB) ( )  

References | Related Articles | Metrics

The practical Byzantine fault tolerance (PBFT) algorithm applied to the alliance chain has some problems, such as arbitrary selection of master nodes and high communication overhead. To solve these problems, an improved Byzantine faulttolerant consensus algorithm (NRPBFT) based on node recognition is proposed. Firstly, the consistency process is optimized by introducing BLS aggregate signature. Secondly, CatBoost algorithm is used to identify nodes, select highreputation nodes as the primary nodes, and dynamically process lowreputation nodes. Experimental results show that NRPBFT is superior to PBFT and ABFT in security, throughput and delay.

A Web Vulnerability Detection Solution Integrating LSTM for  Directory Acquisition

2024, 10(9):  824. 

Asbtract ( )   PDF (2828KB) ( )  

References | Related Articles | Metrics

Addressing the limitations of current vulnerability detection methods in directory acquisition capabilities and detection coverage, this paper proposes a Web vulnerability detection scheme that integrates LSTM (Long ShortTerm Memory) for directory acquisition. The proposed solution incorporates Arjun for efficient parameter bruteforcing technique to obtain basic directory paths and introduces an LSTMbased approach to generate fuzzy directory paths, constructing a comprehensive directory path pool that penetrates hidden directories and quickly acquires a larger number of valid directory paths. To overcome the challenge of detecting atypical Web vulnerabilities, the proposed solution has been implemented as an automated, universal vulnerability detection and verification tool. This tool is suitable for both typical and atypical vulnerabilities and is equipped with capabilities for directory acquisition, vulnerability detection, and bypassing techniques for cookies and IP blocking. Experimental results demonstrate that this solution outperforms typical directory bruteforcing tools by acquiring more valid directory paths, exhibiting excellent directory acquisition capabilities, and effectively detecting and covering a wider range of Web vulnerabilities with high efficiency and a low false positive rate.

Research on Risk Analysis and Countermeasures of Software Supply  Chain in Critical Information Infrastructure

2024, 10(9):  833. 

Asbtract ( )   PDF (1295KB) ( )  

References | Related Articles | Metrics

System security protection is crucial to critical information infrastructures (CII), and  software supply chain risk analysis is indispensable. In recent years, the number of supply chain attack incidents has increased rapidly. This paper first analysis the potential problems of “external” software components, personnel, tools, etc., which are the main causes of software supply chain threats, and then summarize the current research of domestic and foreign policies and technologies. Based on these findings, a software supply chain security framework for power industry systems is proposed. It covers 15 groups of security measures across 4 aspects, including external component governance, supplier management, development and operation facilities reinforcement, usage mechanism of the software bill of materials (SBOM), all of which can be  further extended. This framework can provide references on software supply chain security protection in power industry information systems.

Network Traffic Measurement Based on Multilayer Sketch in SDN

2024, 10(9):  840. 

Asbtract ( )   PDF (1536KB) ( )  

References | Related Articles | Metrics

Network traffic measurement for large flow detection, mutation flow detection and base estimation is of great significance for ensuring network security. However, the current related research suffers from the problems of insufficient realtime performance and low measurement accuracy. In response to the above issues, this paper designs a network traffic measurement model based on Multiple Layer Sketch (ML Sketch). First, the model adopts an independently designed ML Sketch structure, which uses a categorized storage structure to improve the accuracy of traffic measurement. Second, we simulate the dynamic occurrence scenarios of traffic in SDN (Software Defined Network) environment using realtime traffic playback technology. Finally, realtime dynamic detection of large, mutating and base estimation classes of traffic is realized in the SDN control plane. The experimental results on UNSWNB15 show that compared with the traditional Sketch structure, the ML Sketch structure designed in this paper improves the F1_Score metric by up to 4.81% and reduces the correlation error by up to 81.12%, verifying the effectiveness of the model in this paper.

Image Processing Model Watermarking Method Based on #br# Attention Mechanism and Passport Layer Embedding#br#

2024, 10(9):  849. 

Asbtract ( )   PDF (2025KB) ( )  

References | Related Articles | Metrics

With the wide application of deep neural networks in the field of artificial intelligence, the copyright protection of deep neural networks has received extensive attention. However, so far, most of the methods for model copyright protection focus on detection or classification tasks, and are difficult to be directly applied to image processing networks. To this end, this paper proposes an image processing model copyright protection framework combining attention mechanism and passport layer embedding. Firstly, the channel and spatial attention network are used in the watermark embedding network to locate the human eye insensitive area in the image, which improves the robustness and imperceptibility of the watermark. Secondly, the passport layer watermark is inserted after the convolution layer of the target model to improve the ability to resist the ambiguity attacks. Finally, the combination loss is designed to guide the convergence direction of the model in combination with structural consistency and passport layer factors. Experimental results on superresolution and semantic segmentation models show that the watermark extraction rate of this method is more than 98%, and it has good robustness to surrogate attack and ambiguity attack.

An Anonymous Roaming Authentication Scheme for Mobile Network

2024, 10(9):  856. 

Asbtract ( )   PDF (1023KB) ( )  

References | Related Articles | Metrics

With the widespread use of mobile devices, issues like roaming authentication and identification privacy become increasingly prominent. Many anonymous authentication protocols have been proposed in recent years. Among them, some schemes depend on a temporary identity instead of real identity and prevent attackers from tracking by updating authentication identities.Other schemes verify the identity of mobile terminals with the help of home server. However, these schemes generally have the problem of low authentication efficiency or increased authentication delay. In view of this, an anonymous roaming authentication scheme is proposed based on the idea of proxy signature. Mobile terminal generates proxy signature information by using the proxy authorization of the home server. Remote authentication servers can directly verify the identity of mobile terminals without the help of home server. Analysis shows that this scheme achieves security features such as anonymity of mobile terminal, nonrepudiation, unlikability and resistance to forgery attacks, while also reducing the computational load of and communication delay for the mobile terminal compared to existing schemes.

Research on Risk Analysis of Opensource Software Supply Chain Security

2024, 10(9):  862. 

Asbtract ( )   PDF (1824KB) ( )  

References | Related Articles | Metrics

Opensource software has become one of the most fundamental elements that support the operation of the digital society. It has also been penetrated to various industries and fields. As the opensource software supply chain becomes increasingly complex and diversified, the risks caused by security attacks on the opensource software supply chain are also intensified. This paper summarizes the current development of the opensource software supply chain ecosystem and the strategic layout of opensource software supply chain security in major countries. From the dimensions of development security, usage security, and operation security, this paper proposes an opensource software supply chain security risk analysis system. It identifies the major security risks currently faced by the opensource software supply chain. Besides, this paper constructs a security assurance model for the opensource software supply chain and offers countermeasures and suggestions for the security and development of China’s opensource software supply chain from the dimensions of supply chain phases, relevant entities, and safeguard measures.


Analysis and Security System Design for the Open Sharing Mode of Public Data

2024, 10(9):  870. 

Asbtract ( )   PDF (4342KB) ( )  

References | Related Articles | Metrics

Data becomes the fifth key factor of production. The overall situation and strategic construction of Digital China relies on the digital economy, which makes the data as the key factor. Public data generated by government departments, enterprises and institutions at all levels in the process of performing their duties or providing public services in accordance with the law becomes an important source of data element supply. By analyzing the public data sharing mode and security risks, the data security sharing protection system is designed, which provides the balance mechanism for the main body of public data to achieve openness and security.

Circulation Policy and Practice of Global Data Elements

2024, 10(9):  877. 

Asbtract ( )   PDF (6066KB) ( )  

References | Related Articles | Metrics