VDC: Versatile Data Cleanser for Detecting Dirty Samples via Visual-Linguistic Inconsistency.
Zihao ZhuMingda ZhangShaokui WeiBingzhe WuBaoyuan WuPublished in: CoRR (2023)
Keyphrases
- data sets
- raw data
- data processing
- data collection
- data analysis
- data structure
- high quality
- data quality
- original data
- training data
- database
- data distribution
- spectral data
- natural language
- data samples
- complex data
- test data
- synthetic data
- high dimensional data
- image data
- data points
- training samples
- input data
- small number
- data objects
- visual data
- training dataset
- probability distribution
- training set