InfoBatch: Lossless Training Speed Up by Unbiased Dynamic Data Pruning.
Ziheng QinKai WangZangwei ZhengJianyang GuXiangyu PengZhaopan XuDaquan ZhouLei ShangBaigui SunXuansong XieYang YouPublished in: ICLR (2024)
Keyphrases
- database
- data sets
- statistical analysis
- image data
- raw data
- experimental data
- data analysis
- synthetic data
- data sources
- knowledge discovery
- original data
- sensor data
- data structure
- training data
- neural network
- data points
- probability distribution
- data model
- small number
- data collection
- prior knowledge
- labeled data
- training set
- high quality
- data distribution
- statistical methods
- data quality