SuperST: Superficial Self-Training for Few-Shot Text Classification.
Ju Hyoung LeeJoonghyuk HahnHyeon-Tae SeoJiho ParkYo-Sub HanPublished in: LREC/COLING (2024)
Keyphrases
- text classification
- co training
- text categorization
- labeled data
- semi supervised learning
- bag of words
- text data
- feature selection
- machine learning
- video shots
- naive bayes
- text mining
- text documents
- visual features
- unlabeled data
- video sequences
- video data
- video content
- n gram
- sentiment analysis
- key frames
- training set
- cost sensitive
- knn
- shot boundary detection
- text classifiers
- data cleaning
- sports video
- automatic text classification
- neural network
- semi supervised
- shot change detection