ML Security PapersMURIM: Multidimensional Reputation-based Incentive Mechanism for Federated Learning
Sindhuja Madabushi , Dawood Wasif , Jin-Hee Cho
Published on arXiv
2512.13955
Data Poisoning Attack
OWASP ML Top 10 — ML02
Key Finding
Achieves up to 85% robustness improvement against poisoning and noisy-gradient attacks, 5–9% reduction in privacy attack success rates, and 18% fairness improvement versus state-of-the-art baselines.
MURIM (Multidimensional Reputation-based Incentive Mechanism)
Novel technique introduced
Federated Learning (FL) has emerged as a leading privacy-preserving machine learning paradigm, enabling participants to share model updates instead of raw data. However, FL continues to face key challenges, including weak client incentives, privacy risks, and resource constraints. Assessing client reliability is essential for fair incentive allocation and ensuring that each client's data contributes meaningfully to the global model. To this end, we propose MURIM, a MUlti-dimensional Reputation-based Incentive Mechanism that jointly considers client reliability, privacy, resource capacity, and fairness while preventing malicious or unreliable clients from earning undeserved rewards. MURIM allocates incentives based on client contribution, latency, and reputation, supported by a reliability verification module. Extensive experiments on MNIST, FMNIST, and ADULT Income datasets demonstrate that MURIM achieves up to 18% improvement in fairness metrics, reduces privacy attack success rates by 5-9%, and improves robustness against poisoning and noisy-gradient attacks by up to 85% compared to state-of-the-art baselines. Overall, MURIM effectively mitigates adversarial threats, promotes fair and truthful participation, and preserves stable model convergence across heterogeneous and dynamic federated settings.
Key Contributions
- Reliability verification module that detects malicious/unreliable FL clients based on consistency of privacy budgets, resource reports, and training behavior
- Subspace Leverage Equalizer (SLE) — a fairness metric improving geometric representation of underrepresented clients during aggregation
- Unified multidimensional incentive framework integrating privacy, resource constraints, fairness, and reliability without hardware-based TEEs or blockchain
🛡️ Threat Analysis
MURIM explicitly defends against poisoning and noisy-gradient attacks from malicious FL clients — classic Byzantine-fault-tolerant FL defense achieving up to 85% robustness improvement. Byzantine attacks in federated learning where malicious clients corrupt model updates map directly to ML02.