FIESTA: Fast IdEntification of State-of-The-Art models using adaptive bandit algorithms.
Henry B. MossAndrew MooreDavid S. LesliePaul RaysonPublished in: ACL (1) (2019)
Keyphrases
- computational cost
- adaptive algorithms
- learning algorithm
- computationally efficient
- machine learning algorithms
- free parameters
- bandit problems
- probabilistic model
- optimization problems
- machine learning methods
- regression model
- numerical algorithms
- learned models
- incomplete data
- recently developed
- statistical methods
- times faster
- orders of magnitude
- worst case
- least squares
- computational complexity