Distillation from heterogeneous unlabeled collections.
Jean-Michel BegonPierre GeurtsPublished in: CoRR (2022)
Keyphrases
- heterogeneous collections
- information retrieval
- document collections
- semi supervised learning
- digital libraries
- unlabeled data
- unsupervised learning
- training data
- data collections
- metadata
- highly dynamic
- active learning
- supervised learning
- probabilistic model
- semi supervised
- data sets
- training set
- artificial intelligence
- neural network