Three Approaches for Personalization with Applications to Federated Learning
arXiv (Cornell University)2020
Citations Over Time
Abstract
The standard objective in machine learning is to train a single model for all users. However, in many learning scenarios, such as cloud computing and federated learning, it is possible to learn a personalized model per user. In this work, we present a systematic learning-theoretic study of personalization. We propose and analyze three approaches: user clustering, data interpolation, and model interpolation. For all three approaches, we provide learning-theoretic guarantees and efficient algorithms for which we also demonstrate the performance empirically. All of our algorithms are model-agnostic and work for any hypothesis class.
Related Papers
- → Edubase Cloud: Cloud platform for cloud education(2012)7 cited
- → Green Cloud Computing(2022)1 cited
- Personalization of Digital Contents(2006)
- Inspirational Personalization: Abstract and Concrete Levels of Personalization(2014)