Vocabulary expansion of compound words for domain adaptation of BERT.
Hirotaka TanakaHiroyuki ShinnouPublished in: PACLIC (2022)
Keyphrases
- domain adaptation
- compound words
- cross domain
- multiple sources
- semi supervised
- labeled data
- transfer learning
- term dependence
- semi supervised learning
- sentiment classification
- test data
- broadcast news
- co training
- keywords
- noun phrases
- automatic speech recognition
- unlabeled data
- training set
- target domain
- feature selection
- text classification