Differentially-Private Federated Linear Bandits
Differentially-Private Federated Linear Bandits
复制标题
DOI：
--
复制
发表时间：
2020-10
期刊：
影响因子：
--
通讯作者：
Abhimanyu Dubey;A. Pentland
中科院分区：
文献类型：
--
作者：
Abhimanyu Dubey;A. Pentland
分散式学习系统的迅速发展要求有区别的私人合作学习。在本文中，我们研究这一背景下的线性土匪：我们认为一个集合的代理合作，以解决一个共同的上下文土匪，同时确保他们的通信保持私密。对于这个问题，我们设计了\textsc{FedUCB}，集中式和分散式（对等）联邦学习的多代理私有算法。我们提供了一个严格的技术分析，其效用的遗憾，改善合作的强盗学习的几个结果，并提供严格的隐私保证。我们的算法提供了竞争力的性能，无论是在伪遗憾界和经验基准性能在各种多智能体设置。
The rapid proliferation of decentralized learning systems mandates the need for differentially-private cooperative learning. In this paper, we study this in context of the contextual linear bandit: we consider a collection of agents cooperating to solve a common contextual bandit, while ensuring that their communication remains private. For this problem, we devise \textsc{FedUCB}, a multiagent private algorithm for both centralized and decentralized (peer-to-peer) federated learning. We provide a rigorous technical analysis of its utility in terms of regret, improving several results in cooperative bandit learning, and provide rigorous privacy guarantees as well. Our algorithms provide competitive performance both in terms of pseudoregret bounds and empirical benchmark performance in various multi-agent settings.