Gradient Obfuscation Gives a False Sense of Security in Federated Learning.
Kai YueRicheng JinChau-Wai WongDror BaronHuaiyu DaiPublished in: USENIX Security Symposium (2023)
Keyphrases
- learning process
- learning algorithm
- data sets
- reinforcement learning
- prior knowledge
- online learning
- learning systems
- security issues
- inductive inference
- learning problems
- background knowledge
- intrusion detection
- unsupervised learning
- knowledge acquisition
- supervised learning
- software engineering
- machine learning
- neural network