MATES: Model-Aware Data Selection for Efficient Pretraining with Data Influence Models.
Zichun YuSpandan DasChenyan XiongPublished in: CoRR (2024)
Keyphrases
- data sets
- experimental data
- high level
- learning models
- data collection
- prior knowledge
- database
- probabilistic model
- statistical methods
- image data
- data analysis
- probability distribution
- data distribution
- statistical model
- parameter estimation
- data sources
- computational model
- original data
- high quality
- test data
- training data
- data structure
- data processing
- data points
- measured data
- raw data
- simulation data
- statistical models
- sensor data
- synthetic data
- statistical analysis
- end users