Metadata Archaeology: Unearthing Data Subsets by Leveraging Training Dynamics.
Shoaib Ahmed SiddiquiNitarshan RajkumarTegan MaharajDavid KruegerSara HookerPublished in: ICLR (2023)
Keyphrases
- data sets
- synthetic data
- statistical analysis
- metadata
- data analysis
- data collection
- training examples
- database
- data processing
- image data
- high quality
- end users
- prior knowledge
- data structure
- experimental data
- supervised learning
- labeled data
- decision trees
- spatial data
- original data
- information resources
- heterogeneous data
- geospatial data
- earth science
- training and testing data