Improving memory banks for unsupervised learning with large mini-batch, consistency and hard negative mining.
Adrian BulatEnrique Sánchez-LozanoGeorgios TzimiropoulosPublished in: CoRR (2021)
Keyphrases
- unsupervised learning
- supervised learning
- data mining
- positive and negative
- text mining
- memory requirements
- memory usage
- web mining
- consistency checking
- mining algorithm
- itemsets
- semi supervised
- knowledge discovery
- limited memory
- batch mode
- machine learning
- random access
- computing power
- deep learning
- memory size
- pattern discovery
- pattern mining
- frequent itemsets
- active learning
- data analysis
- object recognition
- data structure
- feature extraction
- feature selection