Uncertainty-driven dynamics for active learning of interatomic potentials.
Maksim KulichenkoKipton BarrosNicholas LubbersYing Wai LiRichard A. MesserlySergei TretiakJustin S. SmithBenjamin T. NebgenPublished in: Nat. Comput. Sci. (2023)
Keyphrases
- active learning
- semi supervised
- higher order
- extreme events
- machine learning
- labeled data
- uncertain data
- pool based active learning
- data driven
- supervised learning
- training set
- learning algorithm
- training examples
- high order
- multi agent
- recurrent networks
- selective sampling
- potential functions
- probability theory
- data mining
- imbalanced data classification
- active learning strategies
- sample selection
- robust optimization
- experimental design
- belief functions
- random sampling
- conditional probabilities
- incomplete information
- transfer learning
- learning process