Federated Learning with Client Subsampling, Data Heterogeneity, and Unbounded Smoothness: A New Algorithm and Lower Bounds.
Michael CrawshawYajie BaoMingrui LiuPublished in: NeurIPS (2023)
Keyphrases
- learning algorithm
- input data
- data sets
- prior information
- cost function
- noisy data
- learned models
- objective function
- prior knowledge
- detection algorithm
- training data
- data sources
- lower bound
- preprocessing
- learning process
- synthetic datasets
- data points
- worst case
- similarity measure
- online learning
- data structure
- statistical queries
- inductive inference
- memory space
- learning tasks
- missing data
- labeled data
- linear programming
- data mining techniques
- upper bound
- probabilistic model
- np hard
- data analysis
- computational complexity
- machine learning