Stochastic linear optimization never overfits with quadratically-bounded losses on general data.
Matus TelgarskyPublished in: CoRR (2022)
Keyphrases
- data sets
- training data
- data analysis
- synthetic data
- computer systems
- probability distribution
- experimental data
- high dimensional data
- data collection
- data processing
- data distribution
- small number
- raw data
- end users
- data sources
- original data
- special case
- prior knowledge
- high quality
- linear programming
- noisy data
- complex data
- spatial data
- attribute values
- input data
- image data
- multi objective
- data structure
- website