XMAM: X-raying Models with A Matrix to Reveal Backdoor Attacks for Federated Learning.
Jianyi ZhangFangjiao ZhangQichao JinZhiqiang WangXiaodong LinXiali HeiPublished in: CoRR (2022)
Keyphrases
- learning process
- prior knowledge
- probabilistic model
- accurate models
- hidden variables
- learned models
- learning tasks
- learning systems
- learning algorithm
- supervised learning
- structured prediction
- digital libraries
- reinforcement learning
- learning models
- previously learned
- qualitative models
- inductive inference
- factorization methods
- generative model
- complex systems
- parameter estimation
- online learning
- graphical models
- semi supervised
- data sources
- active learning
- artificial neural networks
- training data