Abstract The widespread adoption of deep learning (DL) libraries has raised concerns about their reliability and security. While prior works leveraged large language models (LLMs) to generate test programs for DL library APIs, the hardcoded program behaviors and low code validity rates render them impractical for real-world testing. To address these challenges, we propose FD-FACTORY, a fully automated framework that leverages LLMs to generate fuzz drivers for DL API testing. The fuzz driver programs accept mutated inputs from fuzzing engines to achieve effective code analysis. Inspired by the modular design of industrial production lines, FD-FACTORY decomposes the generation process into eight distinct stages: Preparation, Initial Fuzz Driver Generation, Early Stop Checks, Verification, Issue Diagnosis, Decision Making, Repair Loop, and Deployment . Each stage is handled by dedicated agents or tools to enhance construction efficiency. Experimental results demonstrate that FD-FACTORY achieves 73.67% and 65.33% success rates in generating fuzz drivers for PyTorch and TensorFlow, producing an improvement of 34.66 to $$-$$ - 54.66% than existing approaches. In addition, FD-FACTORY provides more comprehensive coverage tracking by supporting both Python and native C/C code. It achieves a total coverage of 308,351 lines on PyTorch and 528,427 lines on TensorFlow, substantially surpassing the results reported by previous approaches. Unlike prior approaches relying on repeated interactions with the LLM servers throughout the entire testing process, our framework confines the use of LLMs strictly to the fuzz driver generation stages before deployment. Once generated, the fuzz drivers can be reused without further LLM involvement, thereby enhancing the practicality and sustainability of LLM-assisted fuzzing in real-world scenarios.

Automated software vulnerability detection using learning-based approaches has been a focal point in the field of software engineering. However, the training and benchmarking of software vulnerability detection models are significantly influenced by the quality of the training data. Existing solutions have made limited efforts in addressing data quality issues due to limited and challenging data collection. Publicly available datasets have been found to suffer from data quality problems, hindering effective model training and performance evaluation. Although awareness of the potential negative impact of software vulnerability data quality is increasing, to the best of our knowledge, no systematic solution has been proposed to improve data quality during the automated labeling process. In this paper, we propose a data collection and cleansing framework that first collects the latest vulnerabilities and patches from publicly available vulnerability databases. Then, a rule-based filter is applied to classify function-level vulnerability fixing modifications into three categories: high quality, unknown quality, and low quality. Subsequently, a semantic filter trained on high-quality samples is used to filter samples of unknown quality, resulting in a cleansed version of the raw dataset. This is the first framework that distinguishes the quality of function-level modification samples for software vulnerability fixes and performs data cleansing, without solely relying on traditional heuristic label assignment strategies. In our experiments, we evaluate the properties of SRVul, the effectiveness of the framework, and the feasibility of using it for training vulnerability detection models. SRVul outperforms existing works on multiple metrics, demonstrating the best combination of dataset scale and quality, with a well-designed and effective framework and components. Training the advanced LineVul model with SRVul yields improved performance on benchmark datasets, indicating that SRVul is well-suited for the effective training of vulnerability detection models.

Over the past few years, the emergence of backdoor attacks has presented significant challenges to deep learning systems, allowing attackers to insert backdoors into neural networks. When data with a trigger is processed by a backdoor model, it can lead to mispredictions targeted by attackers, whereas normal data yields regular results. The scope of backdoor attacks is expanding beyond computer vision and encroaching into areas such as natural language processing and speech recognition. Nevertheless, existing backdoor defense methods are typically tailored to specific data modalities, restricting their application in multimodal contexts. While multimodal learning proves highly applicable in facial recognition, sentiment analysis, action recognition, visual question answering, the security of these models remains a crucial concern. Specifically, there are no existing backdoor benchmarks targeting multimodal applications or related tasks.

Federated learning (FL) enables collaborative model training across multiple parties without sharing raw data, with semi-asynchronous FL (SAFL) emerging as a balanced approach between synchronous and asynchronous FL. However, SAFL faces significant challenges in optimizing both gradient-based (e.g., FedSGD) and model-based (e.g., FedAvg) aggregation strategies, which exhibit distinct trade-offs in accuracy, convergence speed, and stability. While gradient aggregation achieves faster convergence and higher accuracy, it suffers from pronounced fluctuations, whereas model aggregation offers greater stability but slower convergence and suboptimal accuracy. This paper presents FedQS, the first framework to theoretically analyze and address these disparities in SAFL. FedQS introduces a *divide-and-conquer strategy* to handle client heterogeneity by classifying clients into four distinct types and adaptively optimizing their local training based on data distribution characteristics and available computational resources. Extensive experiments on computer vision, natural language processing, and real-world tasks demonstrate that FedQS achieves the highest accuracy, attains the lowest loss, and ranks among the fastest in convergence speed, outperforming state-of-the-art baselines. Our work bridges the gap between aggregation strategies in SAFL, offering a unified solution for stable, accurate, and efficient federated learning. The code and datasets are available at https://github.com/bkjod/FedQS_.

For deep learning-enabled semantic communication, existing privacy protection methods only take into account the presence of eavesdropper while ignoring malicious receiver aiming to detect confidential information. Only informationtheoretical security can transmitter defend against malicious receiver. However, private information are always entangled with pragmatic information in feature space, which leads global perturbation to degrade communication performance. To handle these difficulties, in this paper a privacy preserving semantic communication system is proposed. Different from traditional paradigm, a novel privacy preserving semantic encoder is designed to realize targeted privacy protection while remaining useful information unaffected. Within proposed privacy preserving semantic encoder, feature decoupling module aims to disentangle semantic information by learning two sets of basis vectors which can express private and pragmatic information of data, respectively. Accordingly differential privacy mechanism is employed to provide information-theoretical security. Experimental results demonstrate that proposed method not only achieves better communication performance in both data recovery and pragmatic task, but also more effectively degrades the accuracy of malicious receiver to infer sensitive information than global perturbation does.

We consider a$\mathrm{K}$user interference network with general connectivity, described by a matrix N, and general message flows, described by a matrix M. Previous studies have demonstrated that the standard interference alignment (IA) scheme might not be optimal for networks with sparse connectivity. In this paper, we formalize a general IA coding scheme and an intuitive number-filling puzzle for given M and N in a way that the score of the solution to the puzzle determines the optimum sum degrees that can be achieved by the IA scheme. A solution to the puzzle is proposed for a general class of symmetric channels, and it is shown that this solution leads to enhanced Sum-DoF compared to the standard IA scheme.

We consider a full-duplex wireless Distributed Computing (DC) system under the MapReduce framework. New upper and lower bounds on the optimal tradeoff between Normalized Delivery Time (NDT) and computation load are presented. The upper bound strictly improves over the previous reported upper bounds and is based on two novel interference alignment (IA) schemes tailored to the interference cancellation capabilities of the nodes. Our second IA scheme additionally applies a zero-forcing strategy that allows to accumulate all interference at any of the nodes on the same (small) subspace, leaving the remaining space for useful signals. The lower bound is proved through information-theoretic converse arguments based on carefully chosen multi-access channel (MAC) type arguments and by finding solutions to the optimization problems resulting from these arguments. The lower bound matches an existing upper bound based on zero-forcing and interference cancellation (but no IA) in the regime where each node can store at least half of the files. While optimal in this regime, zero-forcing and interference cancellation are not sufficient to obtain the optimal NDT in scenarios where each node cannot store half of the files. This follows from the previously established optimal NDT under zero-forcing and interference cancellation and our new IA-schemes.

We consider a full-duplex wireless Distributed Computing (DC) system under the MapReduce framework. New upper and lower bounds on the optimal tradeoff between Normalized Delivery Time (NDT) and computation load are presented. The upper bound strictly improves over the previous reported upper bounds and is based on a novel interference alignment (IA) scheme tailored to the interference cancellation capabilities of MapReduce nodes. The lower bound is proved through information-theoretic converse arguments.

With the development of software vulnerability analysis, the evaluation of different bug-detecting tools has become quite important for selecting better-performed ones and improving existing approaches. To obtain a convincing evaluation result, a well-constructed vulnerability corpus is indispensable. However, the existing corpora are either constructed from real-world bugs or artificially designed, suffering various problems like small volume, lack of ground truth, etc. Thus, generating large-scale bug corpora through an automated way has been widely noticed. In this paper, we propose an automated vulnerability injection system to generate code samples with triggerable vulnerabilities. Specifically, the system analyzes a host program with the symbolic execution tool to generate high-coverage test cases. Meanwhile, it identifies the potential bug injection points and performs static taint analysis to mark tainted variables and their relevance to the bug injection points. Based on the variables, the system modifies the host program to vulnerable code samples that could be verified by the test cases. In conclusion, the system realizes the injection of buffer overflow vulnerabilities in $\mathrm{C}/ \mathrm{C}++$ programs. A study case is shown to demonstrate the injection processes, and the evaluation presents our advantages in the realness and magnitude of generated bug samples as well as solving highcoverage test cases.

The SCADA (Supervisory Control And Data Acquisition) has become ubiquitous in industrial control systems. However, it may be exposed to cyber attack threats when it accesses the Internet. We propose a three-layer IDS (Intrusion Detection System) model, which integrates three main functions: access control, flow detection and password authentication. We use the reliability test system IEEE RTS-79 to evaluate the reliability. The experimental results provide insights into the establishment of the power SCADA system reliability enhancement strategies.

Tor is a relay-based communication network that provides users with anonymous access to the Internet. Tor hidden services protect the identities of the content publishers to achieve anonymity and anti-censorship. However, hidden services are abused to conduct illegal activities, including hosting botnets and trading drugs. This paper proposes a tracing approach to locate illegal hidden services based on the Tor protocol characteristics, which allows a Tor client to embed a signal into a Tor circuit connecting with the illegal hidden service. Once the Tor node nearest to the the hidden service detects the signal, the identity of the hidden service can be revealed. Our approach is simple, powerful, and stable compared with previous methods. We implemented the approach and performed experiments over the Tor network, whose results show that our approach is feasible and effective, with 100% accuracy, 99.25% true positive rate, and zero false positive rate.

Tor is a relay-based communication network that provides users with anonymous access to the Internet. This paper proposes a new attack to associate the anonymous communication relationship between the client and the server in the Tor network, which exploits defects of the Tor protocol, resulting in deanonymizing the Tor network. In this attack, a malicious entry onion router modifies the command field of cell sequences derived from the client to embed a signal sequence. While the exit onion router detects a signal sequence consistent with the embedded signal sequence, the anonymous communication relationships between the client and the server can be confirmed. We have implemented the attack in a private Tor network and our experiments validate its feasibility and effectiveness, with 100% accuracy, 97.5 % true positive rate, and zero false positive rate.

We consider a cooperative X-channel with K transmitters (TXs) and K receivers (Rxs) where Txs and Rxs are gathered into groups of size r respectively. Txs belonging to the same group cooperate to jointly transmit a message to each of the K − r Rxs in all other groups, and each Rx individually decodes all its intended messages. By introducing a new interference alignment (IA) scheme, we prove that when K/r is an integer the Sum Degrees of Freedom (Sum-DoF) of this channel is lower bounded by 2r if K/r ∈ {2, 3} and by $\frac{{K(K - r) - {r^2}}}{{2K - 3r}}$ if K/r ≥ 4. We also prove that the Sum-DoF is upper bounded by $\frac{{{\text{K}}({\text{K}} - {\text{r}})}}{{2{\text{K}} - 3{\text{r}}}}$. The proposed IA scheme finds application in a wireless distributed MapReduce framework, where it improves the normalized data delivery time (NDT) compared to the state of the art.

In vehicular networks, radio-frequency (RF) jamming attacks are considered a major threat to the availability of control channel (CCH). In particular, vehicles may not be able to receive control messages from roadside units (RSUs) due to persistent interference in the CCH, which may claim human lives and result in significant economic losses. In this article, a cooperative anti-jamming beamforming scheme is proposed to address the CCH jamming problems in vehicular networks. This scheme utilizes spatial diversity provided by the multiantenna RSU and relay vehicles to improve the transmission reliability of downlink control messages. In addition, to address the additive effects of the jamming signals and the intergroup interference, the relay selection problem and the beamformer design problem are jointly considered, which is modeled as a mixed-integer nonlinear programming (MINLP) problem. Then, we address this challenging problem by relaxing it into a series of convex subproblems via the semi-definite relaxation (SDR) and convex-concave process (CCP) methods, and then propose to solve these convex subproblems iteratively. The simulation results show that our proposed method convergences rapidly, and compared to the benchmark schemes, significant performance gains can be observed.

As a candidate radio access technique for 5G, Non- Orthogonal Multiple Access (NOMA) has become an important research topic. Radio Frequency (RF) jamming attack can reduce the communication efficiency in NOMA system. Moreover, the jammer equipped Reinforcement Learning (RL) algorithm will be more destructive. On the other hand, the base station (BS) can implement RL to counter the jamming attack. Thus, the whole system evolves to a multi-agent RL system. The interaction between agents results in a highly dynamic environment and the equilibrium state of the system cannot be intuitively predicted. In the past few years, based on Evolutionary Game Theory (EGT), numbers of researchers have developed useful tools to study the multi-agent RL system in detail. The EGT tools give us insight into the equilibrium of the system and make it possible to compare the performance of different RL algorithms. In this paper, we investigate the anti-jamming problem in the NOMA system where both the base station and the jammer equip RL algorithm. We establish the two-player game and demonstrate the existence and uniqueness of equilibrium. Three RL algorithms and their learning dynamics are introduced, which are Q-learning, Lenient Frequency adjusted Q-learning and Regret Minimization. In experiments, the simulation result shows consistency to the theoretical result given by EGT. Regret Minimization outperforms the other two algorithms in term of average reward and converging rate.

The threat of radio frequency jamming attack to cognitive radio network is an issue that has been discussed for a long time. Q-learning is a widely used anti-jamming algorithm due to its model-free characteristic. However, the traditional Q-learning based anti-jamming algorithms suffer from some limitations when dealing with high-dimensional or continuous inputs. The recently proposed double Deep Q-learning Network (DQN) overcomes this weakness by approximating the table based Q function with a deep neural network. In this paper, we apply the double DQN algorithm with frequency hopping strategy against RF jamming attack in a multi-user environment. We test the performances of three types of neural networks which are the fully connected network (FCN), the convolutional neural network (CNN) and the long short term memory (LSTM). The simulation shows the effectiveness of the double DQN algorithm. Meanwhile, the FCN agent gives the best result concerning stability.

Radio Frequency (RF) jamming attacks constitute a major threat to the availability of control channel communications in the vehicular networks. In particular, the victim vehicles may fail to receive the safety related messages from the Road Side Unit (RSU) due to persistent jamming attacks, which can possibly cause tremendous economic loss and claim human lives. In this paper, we propose a cooperative anti-jamming beamforming scheme for the control channel jamming problem in vehicular networks, which takes advantage of the multi-antenna and spatial diversity provided by the RSU and relay vehicles to improve the transmission reliability of the victim vehicles. The anti-jamming beamformer design problem is formulated as a Mixed-integer Nonlinear Programming (MINLP) problem, which is intractable in general. We address this challenging problem by reformulating it as a sequence of convex sub-problems using the semi-definite relaxation (SDR) and convex-concave procedure (CCP) methods. Simulation results are provided to investigate the convergence of the proposed scheme, and significant performance gain can be observed comparing with other benchmark schemes.

Domain-Flux malware is hard to detect because of the variable C&C (Command and Control) domains which were randomly generated by the technique of domain generation algorithm (DGA). In this paper, we propose a Domain-Flux malware detection approach based on DNS failure traffic. The approach fully leverages the behavior of DNS failure traffic to recognize nine features, and then mines the DGA-generated domains by a clustering algorithm and determinable rules. Theoretical analysis and experimental results verify its efficiency with both test dataset and real-world dataset. On the test dataset, our approach can achieve a true positive rate of 99.82% at false positive rate of 0.39%. On the real-world dataset, the approach can also achieve a relatively high precision of 98.3% and find out 197,026 DGA domains by analyzing DNS traffic in campus network for seven days. We found 1213 hosts of Domain-Flux malware existing on campus network, including the known Conficker, Fosniw and several new Domain-Flux malwares that have never been reported before. We classified 197,026 DGA domains and gave the representative generated patterns for a better understanding of the Domain-Flux mechanism.

Radio Frequency (RF) jamming attacks represent a major threat to the availability of services in vehicular networks. In particular, if the control channel is under persistent jamming attacks, the vehicles within the jamming area cannot receive the safety related messages from the road side unit (RSU), which can possibly cause tremendous economic loss and claim human lives. In this paper, we propose to adopt the cooperative relaying technique to address this problem, whereby the neighbouring vehicles outside of the jamming area serve as the relay nodes to forward the received control channel signal to the victim vehicles through the jamming- free service channel. To investigate the performance of this cooperative relaying scheme, we analyse the outage probability at the victims under different jamming scenarios based on Poisson point process (PPP) model. Simulation results are provided to validate the theoretical results and show the effectiveness of the cooperative anti-jamming relay scheme under different conditions.

Abstract As a new networking paradigm, opportunistic networking communications have great vision in animal migration tracking, mobile social networking, network communications in remote areas and intelligent transportation, and so on. Opportunistic networks are one of the evolutionary mobile ad hoc networks, whose communication links often suffer from frequent disruption and long communication delays. Therefore, many opportunistic forwarding protocols present major security issues, and the design of opportunistic networks faces serious challenges such as how to effectively protect data confidentiality and integrity and how to ensure routing security, privacy, cooperation, and trust management. In this paper, we first systematically describe the security threats and requirements in opportunistic networks; then propose a general security architecture of opportunistic networks; and then make an in‐depth analysis on authentication and access control, secure routing, privacy protection, trust management, and incentive cooperation mechanisms; and at the same time, we present a comparison of various security and trust solutions for opportunistic networks. Finally, we conclude and give future research directions. Copyright © 2014 John Wiley & Sons, Ltd.

Advanced Metering Infrastructure (AMI) is the core component in a smart grid that exhibits a highly complex network configuration. AMI shares information about consumption, outages, and electricity rates reliably and efficiently by bidirectional communication between smart meters and utilities. However, the numerous smart meters being connected through mesh networks open new opportunities for attackers to interfere with communications and compromise utilities assets or steal customers private information. In this paper, we present a new DoS attack, called puppet attack, which can result in denial of service in AMI network. The intruder can select any normal node as a puppet node and send attack packets to this puppet node. When the puppet node receives these attack packets, this node will be controlled by the attacker and flood more packets so as to exhaust the network communication bandwidth and node energy. Simulation results show that puppet attack is a serious and packet deliver rate goes down to 20%-10%.

Wireless sensor networks (WSNs) are vulnerable to security attacks due to the broadcast nature of transmission and limited computation capability. After intrusion detection systems (IDSs) identifies an mobile intruder, IDS may broadcast the blacklist to all nodes in network. This method is energy inefficient because all nodes have to receive and forward the alarm packet so as to exhaust communication bandwidth and node energy, especially when there are a large number of sensor nodes in the network. This paper develops an energy efficient intrusion prevention mechanism in WSNs called green firewall. It can isolate an intruder with less overhead, and track the intruder to continually prevent the attack. The paper analyzes the overhead cost of the green firewall and compare it with the flooding broadcast method. Extensive analysis and simulations show that green firewall can prevent the attack and effectively reduce redundant alarm packet transmissions which results in less energy consumption.

Delay tolerant networks (DTNs) attempt to minimize the possible adverse impacts due to limitations and anomalies in intermittently connected networks. Routing in such sparse and dynamic networks is difficult as the source has little information about the destination, rendering a key challenge to find one simple and effective message delivery mechanism. In this paper, we propose PriCost, a protocol based on the cost for efficient routing of messages, and use the node's past interactions with others to determine the cost of potential routing, in the absence of any other information. Our simulations show that PriCost performs better than MaxProp with reduced complexity. The evaluations also show different cost based metrics hardly affect the performance provided they depend on the same feature extraction algorithm.

MIMO (Multiple-Input-Multiple-Output) is a promising structure for wireless communication. While virtual MIMO structure has been proposed for distributed and cooperative wireless networks, this proposed structure has also put additional energy consumption where energy stands in the important position. At the same time, we have observed that this-problem-caused low energy nodes' mibehavior will degrade the whole system efficiency. In this paper, we propose a two-tier correlation matrix based low power nodes detection system which can capture and mitigate relays' malicious behavior before signal combining. This mechanism can effectively find out the low power node in virtual MIMO structure and thus improve the system efficiency. The simulation results show that a better bit error rate performance of this structure in the presence of low power node and our detection mechanism as well is achieved.

Efficient implementation of FIR (finite-impulse-response) equalizers for wireless channels has attracted much attention these days as equalizations can improve the link performance in hostile mobile radio environment by compensating for inter-symbol interference created by multipath within time dispersive channels. Meanwhile, the MIMO (multiple-input-multiple-output) technique becomes a trend in wireless channel design. On the other hand, we have observed a Least-Squares rational function system identification algorithm, which approximates FIR impulse response with IIR (infinite-impulse-response) structures effectively. In this paper, we exploit the approximation algorithm and extend the algorithm for a MIMO response approximation which can generate a hardware-efficient MIMO IIR structure for FIR type time domain equalizers in the wireless system. We demonstrate the efficiency and accuracy of our method with MIMO modeling examples.

Much work has been done to modify the TCP protocol to improve TCP performance - mainly throughput, over wireless link. However, such improvements need to change the TCP implementation on user systems, which is usually difficult to deploy. In this paper, we investigate the effectiveness of enhancing the TCP performance over wireless link by tuning the packet scheduling schemes. Based on an opportunistic packet scheduling algorithm we proposed, we demonstrate that TCP throughput can be significantly improved by using channel-aware packet scheduling algorithms without changing the TCP implementation. Our work suggests that in cases where it is impractical to change the TCP protocols, packet scheduling is an effective way to improve the TCP performance over wireless links.

In this paper, we propose a new group key management scheme for wireless sensor networks in terms of the unreliable wireless channel and unsafe environment. Our proposed scheme implements node revocation through a broadcast polynomial to counteract the node compromise attack and inherits the idea of loss tolerance in LiSP to provide a reliable communication. The analysis shows that the proposed scheme can efficiently revoke the compromised sensor nodes, implicitly authenticate the updated group keys and tolerate the key-update message loss under the unreliable wireless communication channel.