Efficiently Computable Safety Bounds for Gaussian Processes in Active Learning.
Jörn TebbeChristoph ZimmerAnsgar StelandMarkus Lange-HegermannFabian MiesPublished in: AISTATS (2024)
Keyphrases
- gaussian processes
- efficiently computable
- active learning
- upper bound
- sufficient conditions
- gaussian process
- lower bound
- computationally hard
- gaussian process regression
- semi supervised
- machine learning
- learning algorithm
- worst case
- covariance function
- multi task
- random sampling
- perceptron algorithm
- gaussian process models
- winnow algorithm
- hyperparameters
- supervised learning
- generalization error
- sample size
- labeled data
- transfer learning
- data sets
- semi supervised learning
- np hard
- learning process