Combining Deep Generative Models and Multi-lingual Pretraining for Semi-supervised Document Classification.
Yi ZhuEhsan ShareghiYingzhen LiRoi ReichartAnna KorhonenPublished in: CoRR (2021)
Keyphrases
- document classification
- generative model
- semi supervised
- multi lingual
- semi supervised learning
- text categorization
- labeled data
- text classification
- pairwise
- mixture model
- text mining
- language independent
- cross lingual
- unlabeled data
- text documents
- active learning
- classification algorithm
- discriminative learning
- web documents
- probabilistic model
- supervised learning
- unsupervised learning
- topic models
- information retrieval
- prior knowledge
- data mining
- expectation maximization
- em algorithm
- weakly supervised
- pairwise constraints
- machine learning