Federated Recommendation System via Differential Privacy
Authors:
- 1 June 2020
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 2592-2597
Abstract
In this paper we are interested in what we term the federated private bandits framework, that combines differential privacy with multi-agent bandit learning. We explore how differential privacy based Upper Confidence Bound (UCB) methods can be applied to multi-agent environments, and in particular to federated learning environments both in ‘master-worker’ and ‘fully decentralized’ settings. We provide theoretical analysis on the privacy and regret performance of the proposed methods and explore the tradeoffs between these two.Keywords
All Related Versions
This publication has 11 references indexed in Scilit:
- Recommender Systems over Wireless: Challenges and Opportunities2018
- Making recommendations bandwidth aware2017
- Deep Learning with Differential Privacy2016
- Online Context-Aware Recommendation with Time Varying Multi-Armed Bandit2016
- The Algorithmic Foundations of Differential PrivacyFoundations and Trends® in Theoretical Computer Science, 2014
- Decentralized Learning for Multiplayer Multiarmed BanditsIEEE Transactions on Information Theory, 2014
- Private and Continual Release of StatisticsACM Transactions on Information and System Security, 2011
- Distributed Learning in Multi-Armed Bandit With Multiple PlayersIEEE Transactions on Signal Processing, 2010
- A contextual-bandit approach to personalized news article recommendation2010
- Finite-time Analysis of the Multiarmed Bandit ProblemMachine Learning, 2002