MoDS: Model-oriented Data Selection for Instruction Tuning.
Qianlong DuChengqing ZongJiajun ZhangPublished in: CoRR (2023)
Keyphrases
- prior knowledge
- expert knowledge
- experimental data
- input data
- data analysis
- probabilistic model
- data collection
- data sets
- data processing
- simulation data
- statistical model
- database
- image data
- data structure
- measured data
- computational model
- probability distribution
- data quality
- prior information
- network structure
- test data
- database systems
- high dimensional data
- similarity measure
- high quality
- objective function
- databases
- data samples
- learning models
- data sources
- selection mechanism
- data mining
- statistical methods
- high level
- mathematical model
- missing data
- data mining techniques