SoftEDA: Rethinking Rule-Based Data Augmentation with Soft Labels.
Juhwan ChoiKyohoon JinJunho LeeSangmin SongYoungbin KimPublished in: CoRR (2024)
Keyphrases
- class labels
- training data
- labeled data
- data sets
- data collection
- database
- synthetic data
- complex data
- knowledge discovery
- text classification
- data mining
- data quality
- experimental data
- missing data
- training samples
- statistical analysis
- data structure
- data analysis
- data driven
- high dimensional data
- small number
- data sources
- active learning
- sensor data
- query processing
- data acquisition
- missing values
- raw data
- prior knowledge
- feature space