Sharper bounds for uniformly stable algorithms.
Olivier BousquetYegor KlochkovNikita ZhivotovskiyPublished in: CoRR (2019)
Keyphrases
- computational cost
- error bounds
- worst case
- recently developed
- significant improvement
- theoretical analysis
- machine learning algorithms
- orders of magnitude
- data structure
- randomized algorithms
- data dependent
- computationally expensive
- pac bayesian
- image processing
- upper and lower bounds
- classification algorithm
- upper bound
- lower bound
- computational complexity
- search algorithm
- reinforcement learning