论文标题
FEDSYSID：一种用于样品效率系统识别的联合方法
FedSysID: A Federated Approach to Sample-Efficient System Identification
论文作者
论文摘要
我们研究了从$ m $客户端的观察结果中学习线性系统模型的问题。捕获：每个客户都在观察来自不同动力学系统的数据。这项工作解决了在异质性存在下如何协作学习动态模型的问题。我们将这个问题作为联合学习问题提出，并表征可实现的性能与系统异质性之间的张力。此外，我们联合样品的复杂性结果比单个代理设置提供了恒定的因素改善。最后，我们描述了一种联合学习算法FedSysID，该算法利用了客户层的现有联合算法。
We study the problem of learning a linear system model from the observations of $M$ clients. The catch: Each client is observing data from a different dynamical system. This work addresses the question of how multiple clients collaboratively learn dynamical models in the presence of heterogeneity. We pose this problem as a federated learning problem and characterize the tension between achievable performance and system heterogeneity. Furthermore, our federated sample complexity result provides a constant factor improvement over the single agent setting. Finally, we describe a meta federated learning algorithm, FedSysID, that leverages existing federated algorithms at the client level.
