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