Towards Computationally Feasible Deep Active Learning.
Akim TsvigunArtem ShelmanovGleb KuzminLeonid SanochkinDaniil LarionovGleb GusevManvel AvetisianLeonid ZhukovPublished in: NAACL-HLT (Findings) (2022)
Keyphrases
- computationally feasible
- active learning
- bayesian analysis
- learning algorithm
- semi supervised
- machine learning
- selective sampling
- exhaustive search
- supervised learning
- random sampling
- pool based active learning
- learning strategies
- deep learning
- training set
- sample selection
- annotation effort
- transfer learning
- relevance feedback
- experimental design
- training examples
- database
- active learning strategies
- batch mode
- online learning
- labeled data
- semi supervised learning
- learning process
- multi objective
- belief nets
- batch mode active learning