Current Issue : January-March Volume : 2026 Issue Number : 1 Articles : 5 Articles
Federated learning (FL) is a type of distributed machine learning that enables multiple participants to collaboratively build machine learning models without transferring data outside their local devices, thereby ensuring data privacy and security. However, free-riding (FR) attacks pose significant threats by sending false, erroneous, or malicious model updates to the central server, attempting to extract private information from other devices during the federated learning process. This results in privacy leakage and reduced model accuracy. Traditional defenses measures against FR attacks typically employ auditing methods to identify malicious clients, but these methods are ineffective when multiple FR clients collude to inflate each other’s scores mutually. This paper proposes a novel defense method against collusion-based FR attacks. We first design a grouping mechanism based on gradient norm to group clients and then update the groups using an inter-client audit system. Finally, the correlation analysis of all groups is carried out to eliminate the attack group to ensure the security of the training process. This method defends against standard FR attacks and effectively detects attackers in collusion scenarios. Experimental results demonstrate that our method significantly improves the detection of malicious clients and enhances model accuracy by 10–20% compared to existing methods. Moreover, the proposed defense mechanism maintains its efficacy even in large-scale client environments, where more than 50% of the clients may be compromised by attackers....
This paper makes two distinct contributions to the security and federated learning communities. First, we identify and empirically demonstrate a critical vulnerability in Krum, a widely deployed Byzantine-resilient aggregation algorithm, showing catastrophic failure (44.7% accuracy degradation) when applied to high-dimensional neural networks. We provide comprehensive analysis of five alternative algorithms and validate FLTrust as a more resilient solution, though requiring trusted infrastructure. This finding has immediate implications for production federated learning systems. Second, we present a rigorous feasibility analysis of quantum-enhanced security operations through simulation-based exploration. We document fundamental deployment barriers including (1) environmental electromagnetic interference exceeding sensor capabilities by 6-9 orders of magnitude, (2) infrastructure costs of USD 3–5M with unproven benefits, (3) an absence of validated correlation mechanisms between quantum measurements and cyber threats, and (4) O(n2) scalability constraints limiting deployments to 20 nodes. This is purely theoretical research using simulated data without physical quantum sensors. Physical validation through empirical noise characterization and sensor deployment in operational environments represents the critical next step, though faces significant challenges from EMI shielding requirements and calibration procedures. Together, these contributions provide actionable insights for current federated learning deployments while preventing premature investment in quantum sensing for cybersecurity....
Currently, deepfake detection has garnered widespread attention as a key defense mechanism against the misuse of deepfake technology. However, existing deepfake detection networks still face challenges such as insufficient robustness, limited generalization capabilities, and a single feature extraction domain (e.g., using only spatial domain features) when confronted with evolving algorithms or diverse datasets, which severely limits their application capabilities. To address these issues, this study proposes a deepfake detection network named EFIMD-Net, which enhances performance by strengthening feature interaction and integrating spatial and frequency domain features. The proposed network integrates a Cross-feature Interaction Enhancement module (CFIE) based on cosine similarity, which achieves adaptive interaction between spatial domain features (RGB stream) and frequency domain features (SRM, Spatial Rich Model stream) through a channel attention mechanism, effectively fusing macro-semantic information with high-frequency artifact information. Additionally, an Enhanced Multi-scale Feature Fusion (EMFF) module is proposed, which effectively integrates multi-scale feature information from various layers of the network through adaptive feature enhancement and reorganization techniques. Experimental results show that compared to the baseline network Xception, EFIMD-Net achieves comparable or even better Area Under the Curve (AUC) on multiple datasets. Ablation experiments also validate the effectiveness of the proposed modules. Furthermore, compared to the baseline traditional two-stream network Locate and Verify, EFIMD-Net significantly improves forgery detection performance, with a 9-percentage-point increase in Area Under the Curve on the CelebDF-v1 dataset and a 7-percentage-point increase on the CelebDF-v2 dataset. These results fully demonstrate the effectiveness and generalization of EFIMD-Net in forgery detection. Potential limitations regarding real-time processing efficiency are acknowledged....
To address the demand for low-cost deployment in quantum key distribution (QKD) networks, this study explores the implementation of unidimensional (UD) modulation continuous-variable quantum key distribution (CV-QKD) protocols within downstream access networks. The UD CV-QKD protocol employs a single modulator for information encoding, offering benefits such as reduced implementation cost and lower random number consumption, which collectively decrease the overall setup expense of QKD systems. Through systematic performance analysis, it is demonstrated that the proposed UD modulation downstream access network scheme exhibits strong scalability and practical applicability. When supporting 32 users, the system maintains secure communication over transmission distances of up to 50 km. As the number of users increases to 64, performance declines slightly; however, the system still achieves a 35 km transmission distance, which remains suitable for many metropolitan access applications. Even in high-density access scenarios involving 128 users, the system sustains a positive key rate within a transmission range of 20 km. Furthermore, this study evaluates the protocol’s practical security under source intensity errors and finite-size effects. These results provide meaningful guidance for deploying low-cost, high-security quantum communication access networks and contribute to advancing QKD technologies toward scalable, real-world implementations....
Secure multi-party extremum, as a significant offshoot of secure multi-party computation, has extensive applications in various domains, including healthcare, financial transactions, market analysis, sports events, etc. Nevertheless, most existing secure multi-party extremum protocols rely on computational hard problems and are thus vulnerable to quantum algorithms. This paper presents a quantum secure multi-party extremum protocol that is built upon the correlations of Greenberger–Horne–Zeilinger (GHZ) states. Within this protocol, multiple participants, with the aid of a semi-honest third party, can obtain the maximum and minimum values of their secret inputs. GHZ states act as the information carriers and are transmitted among the participants and the third party. Their unique correlations ensure the secure transmission of quantum particles. The analysis demonstrates that the proposed protocol is capable of not only warding off common external attacks but also resisting internal attacks launched by dishonest participants and the semi-honest third party. Moreover, the protocol boasts correctness and high scalability....
Loading....