Online Multi-Agent Decentralized Byzantine-robust Gradient Estimation
This addresses the challenge of secure gradient estimation in decentralized multi-agent systems, but appears incremental as it builds on existing techniques like simultaneous perturbation and stochastic approximations.
The paper tackles the problem of distributed Byzantine-robust gradient estimation for black-box models, proposing an iterative scheme based on simultaneous perturbation and secure state estimation, with performance validated through numerical experiments.
In this paper, we propose an iterative scheme for distributed Byzantineresilient estimation of a gradient associated with a black-box model. Our algorithm is based on simultaneous perturbation, secure state estimation and two-timescale stochastic approximations. We also show the performance of our algorithm through numerical experiments.