FedASMU: Efficient Asynchronous Federated Learning with Dynamic Staleness-Aware Model Update.
Ji LiuJuncheng JiaTianshi CheChao HuoJiaxiang RenYang ZhouHuaiyu DaiDejing DouPublished in: AAAI (2024)
Keyphrases
- prior knowledge
- dynamically updated
- probabilistic model
- computational model
- experimental data
- statistical model
- online learning
- learning algorithm
- learning scheme
- learning models
- theoretical framework
- parameter estimation
- theoretical analysis
- objective function
- high level
- machine learning
- dynamic environments
- hidden markov models
- learning process
- conditional random fields
- data structure