arXiv subjects

Zhihui Zhao

Publications and source records attributed to Zhihui Zhao.

2 recordsLinked to original sources

arXiv2026

Breaking the Secret: Economic Interventions for Combating Collusion in Embodied Multi-Agent Systems

Collusion among autonomous agents poses a critical security threat in embodied multi-agent systems (MAS), where coordinated behaviors can deviate from global objectives and lead to real-world consequences. Existing defenses, primarily based on identity control or post-hoc behavior analysis, are insufficient to address such threats in embodied settings due to delayed feedback and noisy observations in physical environments, which make behavioral deviations difficult to detect accurately and in a timely manner. To address this challenge, we propose a mutagenic incentive intervention approach that mitigates collusion by reshaping agents' payoff structures. By rewarding agents who report collusive behavior and penalizing identified participants, the mechanism induces strategic defection and renders collusion unstable. We further design supporting mechanisms, including reporting deposits, smart contract-based reward enforcement, and encrypted communication, to ensure robustness against misuse of the incentive mechanism and retaliation from penalized agents. We implement the proposed approach in both simulated and real-world embodied environments. Experimental results show that our method effectively suppresses collusion by inducing defection, while preserving system efficiency. It achieves performance comparable to the non-collusion baseline and outperforms representative reactive defenses, thereby fulfilling the desired security objectives. These results demonstrate the effectiveness of proactive incentive design as a practical paradigm for securing embodied multi-agent systems.

cs.CR↗

arXiv2025

SMTFL: Secure Model Training to Untrusted Participants in Federated Learning

Federated learning is an essential distributed model training technique. However, threats such as gradient inversion attacks and poisoning attacks pose significant risks to the privacy of training data and the model correctness. We propose a novel approach called SMTFL to achieve secure model training in federated learning without relying on trusted participants. To safeguard gradients privacy against gradient inversion attacks, clients are dynamically grouped, allowing one client's gradient to be divided to obfuscate the gradients of other clients within the group. This method incorporates checks and balances to reduce the collusion for inferring specific client data. To detect poisoning attacks from malicious clients, we assess the impact of aggregated gradients on the global model's performance, enabling effective identification and exclusion of malicious clients. Each client's gradients are encrypted and stored, with decryption collectively managed by all clients. The detected poisoning gradients are invalidated from the global model through a unlearning method. We present a practical secure aggregation scheme, which does not require trusted participants, avoids the performance degradation associated with traditional noise-injection, and aviods complex cryptographic operations during gradient aggregation. Evaluation results are encouraging based on four datasets and two models: SMTFL is effective against poisoning attacks and gradient inversion attacks, achieving an accuracy rate of over 95% in locating malicious clients, while keeping the false positive rate for honest clients within 5%. The model accuracy is also nearly restored to its pre-attack state when SMTFL is deployed.

cs.CR↗